Datadog

Summary

The diff does not include the actual added contractual language. It only states:

> “Added approximately 3604 words to the document”

Accordingly, it is not possible to identify:

• The specific legal or commercial changes;
• New customer obligations or provider rights;
• Changes to liability, indemnity, confidentiality, security, or termination provisions;
• Any changes concerning the use of customer data;
• Whether customer data may be used to train, fine-tune, evaluate, or improve AI models;
• Whether such use is permitted by default, requires consent, or can be disabled;
• Whether data is retained, anonymized, shared with affiliates or subprocessors, or used to generate aggregate insights; or
• Whether the provider claims ownership or broad usage rights over inputs, outputs, prompts, or derived data.

Key Limitation

Because the substantive additions are not provided, no reliable legal-risk assessment can be made. The statement that approximately 3,604 words were added indicates a potentially significant amendment, but it does not reveal whether the changes are favorable, neutral, or adverse to the customer.

Information Needed

Please provide either:

1. The full redline showing the additions, deletions, and replacements; or

2. The text of the newly added provisions, particularly sections addressing:

• Data use and licensing;
• Artificial intelligence or machine learning;
• Model training, fine-tuning, testing, or improvement;
• Confidentiality and security;
• Data retention and deletion;
• Subprocessors and third-party providers;
• Ownership of inputs, outputs, and generated content; and
• Service limitations, liability, indemnification, and termination.

Once the substantive text is available, the changes can be analyzed for customer-impacting rights, obligations, and AI-training risks.

Summary of Important Changes

1. Expanded scope of the Terms and Datadog entities

• The contracting party is expanded from Datadog generally to Datadog, Inc. and its affiliates.
• The Terms now apply to accessing or using the Site and its Content, rather than only a more narrowly described website/platform.
• This may broaden the range of Datadog websites, affiliate-operated sites, materials, and functionality covered by the Terms.

Risk: Users may be deemed bound by these Terms when interacting with a wider set of Datadog or affiliate properties, potentially without a separate click-through agreement.

2. Expanded coverage of Datadog products and AI functionality

• The updated text and related product descriptions expressly reference a much broader product ecosystem, including:
• AI Impact
• Bits AI, Bits Chat, Agent Builder, Bits Code, and Bits Security Analyst
• Datadog for AI
• Agent Observability
• MCP Server
• Other monitoring, security, developer, and service-management products.
• The changes appear to clarify that the Terms may apply even where users access Datadog functionality through free trials, consoles, or related product interfaces.

Risk: Product references may create uncertainty about which terms govern a particular AI-enabled feature, especially where the Master Subscription Agreement or another written contract also applies.

3. Master Subscription Agreement clarification

• The Terms state that use of the hosted application, infrastructure monitoring service, and other services is governed by Datadog’s Master Subscription Agreement (MSA) or another written agreement.
• The revised language appears to broaden the services excluded from these Website Terms and covered by the MSA.

Risk: Customers may need to review both the Website Terms and the MSA. Conflicts, differing liability limits, data rights, and confidentiality obligations may depend on which document governs the relevant activity.

4. Customer data and AI-model training

• No express new provision in the supplied diff authorizes Datadog to use customer data to train AI models.
• The diff does not add language expressly stating that customer data, prompts, inputs, outputs, logs, telemetry, or other Service data may be used for:
• training or fine-tuning AI models;
• improving general-purpose models;
• developing products; or
• sharing data with AI providers.
• The existing broad “Your Content” license remains significant: content made available on the Site may be used, reproduced, modified, distributed, sublicensed, and otherwise used without restriction. However, that provision excludes personal information subject to the Privacy Policy and appears directed to Site-posted content, not necessarily customer data submitted through hosted services.

Risk: The absence of a clear AI-training restriction or permission leaves uncertainty. Customers should review the MSA, applicable product terms, privacy documentation, and AI-specific terms for the controlling data-use rules.

Summary of Important Changes

1. Privacy policy date and introductory language

• The policy appears to be updated to March 12, 2026.
• The introduction now states that the policy covers Datadog products, services, events, websites, and applications linking to the policy.
• New language emphasizes that users can control or exercise rights relating to their personal information.
• The policy also adds the sentence: “This Privacy Policy explains how we collect and use your personal information…” and states that the information is intended to be protected.

Risk: The broader description may make the policy applicable across a wider range of Datadog services and applications, including newer AI-related products.

2. Scope of collected data remains broad

The policy expressly identifies “Product information” submitted through Datadog Products, including information contained in:

• Logs
• Traces
• Metrics
• Names, email addresses, IP addresses, online identifiers, and mailing addresses

It also covers usage, device, cookie, session, location, and interaction data.

Risk: Customer telemetry may contain personal, confidential, security-sensitive, or proprietary information. The policy does not appear to limit collection to data intentionally submitted as personal information.

3. Use for product improvement and research

The policy permits Datadog to use collected information—including Product information—for:

• Research
• Monitoring and analyzing trends and usage
• Improving existing products
• Creating new products
• Combining information with data obtained from other sources

It also states that information may be processed for purposes materially different from those listed only after providing an opportunity to opt out.

Risk: This is a broad secondary-use right. It could permit use of customer-submitted telemetry and related personal information for analytics, product development, and potentially automated systems unless restricted by the customer’s contract or a separate data-processing agreement.

4. AI model training

• The diff does not clearly add an express right to train AI models using customer data.
• The prior text appears to contain a reference to “improve model performance,” but that language is replaced or removed in the revised presentation.
• The updated policy includes extensive references to AI products and features, but those references appear primarily to be website navigation or product listings—not an operative data-use provision.
• There is no clear statement addressing whether customer data is used to train, fine-tune, evaluate, or improve generative AI or machine-learning models.

Key concern: The absence of an explicit AI-training restriction or permission creates uncertainty. Customers should review the applicable service agreement, data-processing addendum, product-specific terms, and any AI terms to determine whether customer content or telemetry may be used for model training.

5. Other material privacy rights and sharing provisions

The policy continues to allow:

• Sharing with service providers, affiliates, advertisers, analytics providers, and in business transfers
• International transfers, including to the United States
• Retention for legal or legitimate business purposes
• Use of anonymized information without further notice
• Opt-outs for targeted advertising, profiling, and certain sales or sharing activities

Overall assessment: The principal substantive risk is the broad ability to use Product information for research and product improvement, combined with unclear treatment of AI training.

Summary of Important Changes

1. Privacy Policy expanded and reorganized

• The policy is now identified as the “Privacy Policy” dated March 12, 2026.
• It applies broadly to Datadog products, services, events, websites, and applications that link to the policy.
• The policy adds or more clearly organizes sections covering:
• Personal information collected
• Uses and sharing
• Security and retention
• User choices and rights
• EEA/UK/Switzerland protections
• U.S. state privacy rights
• International transfers

2. Broader description of collected data

The policy expressly identifies as “Product information” personal information submitted through Datadog products, including logs, traces, and metrics. This may include names, email addresses, IP addresses, online identifiers, and mailing addresses.

It also describes collection of:

• Product and website usage data
• Device, IP address, and location data
• Cookies, session information, links clicked, pages visited, and mouse movements
• Information received from linked services, other users, affiliates, third parties, and public sources

Risk

Customer monitoring data may contain personal, confidential, security-sensitive, or proprietary information. The policy does not appear to limit collection to data that Datadog itself controls or can safely use.

3. Use for product improvement and new products

Datadog states that it may use information it collects, including Product information, to:

• Research, monitor, and analyze trends and usage
• Improve existing products
• Create new products
• Combine information about an individual with information obtained from other sources

Risk

This is a broad secondary-use authorization. It may permit use of customer-submitted logs, traces, and metrics beyond merely delivering monitoring services, subject to other agreements or product-specific terms.

4. AI model training

• The diff adds extensive references to AI-related products and capabilities, including AI investigation, chat, agent-builder, code, security analyst, agent observability, and AI impact features.
• However, the policy does not expressly state that customer data will be used to train, fine-tune, or improve generative AI or machine-learning models.
• It also does not expressly promise that customer data will be excluded from AI training.

Key uncertainty and risk

The general authorization to use Product information to “improve and create new Datadog Products” could potentially be argued to cover AI development, depending on applicable customer terms and product documentation. The policy does not specify:

• Whether customer content is used for model training
• Whether data is aggregated, anonymized, or de-identified first
• Whether data is used to train shared models or only customer-specific models
• How customers can opt out
• Whether prompts, outputs, logs, or telemetry from AI features are retained or reused

Customers should review the governing service agreement, data-processing addendum, and AI-specific terms for clearer restrictions.

5. Sharing, advertising, and transfers

The policy continues to permit sharing with service providers, affiliates, advertising and analytics providers, legal authorities, and business-transfer counterparties. It also describes cross-context behavioral advertising and international transfers, including to the United States.

6. Customer rights

The policy adds or clarifies rights including access, correction, deletion, restriction, objection, portability, opt-outs from targeted advertising and automated profiling, and limits on use of sensitive information. These rights remain subject to applicable law and contractual limitations.

Summary

The provided diff does not include the text of the approximately 3,603 added words. It only states:

> “Added approximately 3603 words to the document”

AI Training and Data Use

There is not enough information to determine whether the new language:

• Permits the customer’s data to be used to train, fine-tune, or improve AI models;
• Restricts use of customer data to service provision;
• Allows use of prompts, outputs, usage data, or personal information for model development;
• Requires anonymization, de-identification, or aggregation before AI training;
• Gives the provider ownership or a license to use customer data;
• Allows retention of customer data after termination; or
• Provides an opt-out, deletion, or confidentiality protection for AI-related data use.

Other Legal Risks

The substantive legal changes cannot be assessed without the added and deleted contract language. In particular, it is not possible to evaluate changes to:

• Data ownership and intellectual-property rights;
• Confidentiality and privacy obligations;
• Security commitments and breach notification;
• Subprocessors and third-party disclosures;
• Data retention and deletion;
• Warranties, indemnities, and liability caps;
• Termination rights; or
• Governing law and dispute resolution.

Information Needed

Please provide the actual redline text, including the approximately 3,603 added words and any deletions or replacements. Without the text, no reliable legal-risk or AI-training analysis can be performed.

Structured Summary of Important Changes

1. Privacy Policy update and scope

• The Privacy Policy is dated March 12, 2026.
• The policy now expressly covers a broader range of Datadog products, including AI-related offerings such as Bits Investigation, Bits AI, Bits Chat, Agent Builder, Bits Code, Security Analyst, Agent Observability, and MCP Server.
• The policy applies to Datadog products, services, events, websites, and applications that link to it.

Risk: The expanded product scope may subject customer and user information submitted through newer AI products to the same broad collection and use provisions, even where the customer does not expect those products to process production data.

2. Customer/Product data collection

The policy expressly states that Datadog collects “Product information” submitted through Datadog Products, including information contained in:

• Logs;
• Traces; and
• Metrics.

This information may include names, email addresses, IP addresses, online identifiers, and mailing addresses. Datadog may also receive information about an individual through another user’s use of the Datadog Products.

Risk: Customer telemetry may contain personal information or confidential business information. The policy does not appear to distinguish clearly between customer-controlled service data and information Datadog collects for its own purposes.

3. Use for improvement, research, and new products

Datadog states that it uses collected information, including Product information, to:

• Improve existing Datadog Products;
• Monitor and analyze trends, usage, and other activities;
• Create new products; and
• Combine information about users with information obtained from other sources to understand needs and provide new services.

AI-training analysis: The diff does not expressly add language stating that customer data will be used to train AI models, nor does it expressly prohibit such use. However, the broad “research,” “improve,” and “create new products” language could potentially be interpreted as permitting model development or training, particularly because the policy now prominently covers AI products.

Risk: There is no stated limitation concerning:

• Whether customer data is used to train foundation or product-specific models;
• Whether data is de-identified, aggregated, or retained for training;
• Whether customer data is shared with AI subprocessors;
• Whether customers can opt out of AI training; or
• Whether trained models may benefit other customers.

Customers should seek contractual clarification or an explicit no-training commitment.

4. Security and confidentiality

The policy adds or emphasizes that personal information is “protected” and describes technical and organizational safeguards. It does not provide a customer-specific confidentiality commitment, data-processing restriction, or guarantee that telemetry will be excluded from product development.

5. Data sharing and advertising

The policy continues to permit sharing with service providers, affiliates, analytics providers, advertising providers, and search-engine providers. It also describes cross-context behavioral advertising and related opt-out rights.

Risk: The broad service-provider language could include infrastructure, analytics, or AI vendors unless restricted by a separate data-processing agreement.

6. Individual rights and retention

The policy describes access, correction, deletion, objection, restriction, portability, consent withdrawal, and U.S. opt-out rights, including opt-outs for automated profiling and targeted advertising. Datadog may retain archived or cached copies and may continue using anonymized information without further notice.

Executive Summary

The diff is heavily corrupted by what appears to be a website-navigation or content-scraping artifact. It nevertheless indicates a revised Datadog Privacy Policy dated March 12, 2026, with expanded references to Datadog’s AI, security, software-delivery, and developer products.

Important Changes and Risks

1. Expanded scope of covered products

The policy now refers broadly to Datadog products, services, events, websites, and applications, and lists numerous products involving:

• AI investigation, chat, agent-building, coding, and security tools;
• Agent Observability and MCP Server functionality;
• Logs, traces, metrics, session replay, analytics, and monitoring;
• Code, infrastructure, cloud-security, and developer tools.

Risk: Customer information submitted through logs, traces, metrics, code-related tools, session replay, or AI products may fall within the policy’s broad definition of “Product information.”

2. Customer data used for product improvement

The policy states that Datadog uses information collected—including Product information—“for research and to improve and create new Datadog Products.” Product information expressly includes information submitted through use of the products, including logs, traces, and metrics, potentially containing names, email addresses, IP addresses, online identifiers, and mailing addresses.

Risk: This is a broad secondary-use authorization. It may permit Datadog to analyze customer-submitted operational data beyond merely providing monitoring services, subject to the customer’s agreement and any separate data-processing terms.

3. AI-model training

The diff does not expressly state that customer data will be used to train generative-AI or machine-learning models. It does, however:

• add or emphasize many AI-related products;
• authorize use of Product information for research, improvement, and creation of new products; and
• permit combining information with information obtained from other sources.

Risk: Although not an explicit “training” provision, the wording could arguably encompass model development, evaluation, tuning, or improvement. The policy does not clearly state:

• whether customer data is used to train Datadog’s or third-party models;
• whether data is de-identified before such use;
• whether customers can opt out;
• whether customer prompts, outputs, code, logs, or telemetry are retained for training; or
• whether data is shared with AI vendors.

Customers should review the governing services agreement, data-processing addendum, product-specific AI terms, and any “no training” or opt-out commitments.

4. Broader sharing and advertising disclosures

The policy continues to permit sharing with service providers, affiliates, advertising and analytics providers, and parties involved in business transfers. It also describes cross-context behavioral advertising and related opt-out rights.

Risk: Customer or user information may be disclosed to a broader vendor ecosystem, particularly where Datadog products incorporate third-party AI, analytics, advertising, or hosting services.

5. Expanded privacy rights and compliance language

The revised text includes detailed EEA/UK/Swiss GDPR rights, international-transfer mechanisms, Data Privacy Framework language, and U.S. rights concerning targeted advertising, automated profiling, sale/sharing, and sensitive information.

Overall assessment: The most significant practical concern is the broad use of customer-submitted Product information for product improvement in an increasingly AI-enabled product environment, without a clear statement limiting or prohibiting AI-model training.

Key Changes and Risks

1. Expanded definition of Datadog and the Site

• Added: The Site may be owned or operated by Datadog, Inc. or its affiliates, collectively defined as “Datadog.”
• Added: The Terms apply to the Site and “any related websites.”
• Risk: More Datadog entities may benefit from the Terms, including disclaimers, liability limitations, indemnities, and enforcement rights. The specific affiliate responsible for processing data or providing a service may therefore be less clear.

2. Broader acceptance trigger and scope

• Changed: Acceptance is triggered by accessing or using the Site or its content, rather than only using a particular product or platform.
• Risk: A user may become bound merely by browsing or accessing Site content. The Terms also appear to cover a substantially broader range of Datadog products and capabilities, including AI-related offerings.

3. Expanded entity-authority representations

• Added/clarified: If a person enters the Terms for an organization, that person represents that they have authority to bind the organization. “You” includes both the individual and the represented entity.
• Risk: An employee or contractor who lacks authority could expose themselves or the organization to arguments concerning breach, indemnity, or unauthorized acceptance.

4. Hosted services and Master Subscription Agreement

• Changed: The Terms expressly exclude the hosted application, infrastructure monitoring service, and other services from their scope.
• Added: Users accessing the Service—including through a free trial—are subject to Datadog’s Master Subscription Agreement (MSA) or another written contract.
• Added/clarified: Accessing the Datadog Service through a free trial, and potentially through free or trial functionality, may still trigger the MSA.
• Risk: Customers may be subject to a separate agreement without receiving a consolidated contract. The MSA may contain materially different provisions on data use, security, confidentiality, AI, liability, governing law, and dispute resolution. The reference to the MSA being “currently” available online may also create version-control concerns.

5. New AI and AI-enabled product references

• Added/expanded: The product descriptions now include AI-related offerings and capabilities, including AI Impact, Datadog AI, Bits products, Agent Builder, Code, Security Analyst, Agent Observability, and MCP Server.
• Risk: These additions broaden the apparent product scope but do not themselves establish specific rights, restrictions, or safeguards for AI functionality.

Customer Data and AI Training

• No express provision was identified in the supplied diff stating that customer data may be used to train, fine-tune, evaluate, or improve Datadog AI models.
• No express prohibition or opt-out right regarding AI training was identified either.
• The relevant AI-data terms may instead appear in the MSA, product-specific terms, privacy documentation, or an AI service agreement. Customers should review those documents specifically for:
• whether customer content, telemetry, logs, prompts, outputs, or usage data may train models;
• whether data is de-identified or aggregated;
• retention and deletion periods;
• human or third-party access; and
• opt-out, enterprise controls, and confidentiality commitments.

Diff Analysis

Summary

The supplied diff does not include the actual contractual language. It only states:

> “Added approximately 3603 words to the document”

Accordingly, it is not possible to identify the legal changes, new risks, or any provisions concerning the use of customer data to train AI models.

AI Training and Customer Data

No language was provided addressing:

• Whether customer data may be used to train, fine-tune, or improve AI models;
• Whether prompts, inputs, outputs, or usage data are retained;
• Whether data is anonymized, aggregated, or de-identified before use;
• Whether customer data may be shared with affiliates, vendors, or model providers;
• Whether the customer can opt out of AI training or request deletion;
• Ownership or licensing rights in customer data, outputs, or model improvements;
• Security, confidentiality, or cross-border transfer controls related to AI processing.

Risk Assessment

Because the 3,603 added words are not included, the following risks cannot be assessed:

• Expansion of the provider’s rights to use customer data;
• Broad or perpetual data licenses;
• Use of confidential or personal information for model training;
• Reduced confidentiality protections;
• Ambiguous ownership of AI-generated outputs;
• New indemnities, disclaimers, liability limitations, or termination rights;
• Changes to data-retention, deletion, audit, or regulatory-compliance obligations.

Information Needed

Please provide the marked-up contractual text showing the additions, deletions, and replacements. The actual language is necessary to perform a meaningful legal comparison and identify specific AI-training provisions and associated customer risks.

Summary of Available Change

The diff states only that approximately 3,478 words were removed. It does not identify which provisions were deleted or provide replacement language.

AI Training and Customer Data

• No definitive conclusion is possible about changes to AI-model training.
• The removed text may have included provisions addressing:
• Whether customer data may be used to train, fine-tune, or evaluate AI models;
• Whether usage data, prompts, outputs, or personal information may be used for model improvement;
• Opt-out rights or customer consent requirements;
• Restrictions on human review or sharing with service providers;
• Data-retention and deletion obligations;
• Ownership of customer inputs and AI-generated outputs; or
• De-identification, anonymization, or aggregation standards.
• If any such provisions were deleted without replacement, the contract may now provide less transparency or fewer restrictions on the provider’s use of customer data. However, this cannot be confirmed from the information supplied.

Other Potential Legal Risks

The deletion of a substantial portion of the document could have removed important protections relating to:

• Confidentiality and data security;
• Privacy-law compliance;
• Data breach notification;
• Subprocessors and third-party sharing;
• Data storage location and international transfers;
• Intellectual-property ownership and licenses;
• Warranties, indemnities, and limitations of liability;
• Service levels, termination, and data return or deletion; and
• Audit, inspection, or regulatory-cooperation rights.

The commercial impact depends entirely on the content of the deleted provisions and whether other sections cross-reference them.

Recommended Review

The deleted text and the revised full agreement should be compared directly. Particular attention should be given to searching the revised agreement for terms such as:

train, training, fine-tune, machine learning, AI, model improvement, inputs, outputs, usage data, aggregated, de-identified, retain, delete, and subprocessor.

Bottom line: The supplied diff is insufficient to determine whether customer data may now be used to train AI models. The deletion itself creates a material review risk because AI-use restrictions or customer protections may have been removed.

Summary of Changes

Overall Change

• The diff indicates that approximately 4,103 words were removed from the document.
• No replacement text or specific deleted provisions were provided.
• Because the actual deleted language is unavailable, the legal and commercial effect of the changes cannot be determined reliably.

Key Legal Risks

Loss of Customer Protections

The deleted provisions may have contained important protections relating to:

• Confidentiality and data security
• Ownership or licensing of customer data
• Restrictions on data disclosure or sharing
• Warranties, indemnities, and liability limits
• Audit, compliance, and regulatory rights
• Termination and data deletion obligations
• Service levels or remedies for breach

Deleting such provisions could materially weaken the customer’s rights or make the agreement less specific and harder to enforce.

Increased Ambiguity

If the deleted text addressed definitions, exceptions, procedures, or limitations, its removal may create ambiguity about:

• What information is covered by the agreement
• Which party bears responsibility for compliance
• How data may be retained, transferred, or deleted
• What remedies are available if the provider breaches the agreement

AI Model Training and Data Use

• The supplied diff does not identify whether any language concerning AI training was added, removed, or changed.
• It is therefore not possible to determine whether customer data may now be:
• Used to train, fine-tune, or improve AI models
• Combined with other customers’ data
• Reviewed by humans for model development
• Retained after termination for training purposes
• Used for product development or analytics
• Excluded from training by default or only upon opt-out

The deletion of approximately 4,103 words could potentially remove restrictions or consent requirements concerning AI training, but that cannot be confirmed without the deleted text.

Recommended Follow-Up

Provide either:

1. The full redline showing the deleted provisions; or

2. The prior and revised versions of the agreement.

Particular attention should be given to provisions titled “Data Use,” “Customer Data,” “Confidentiality,” “Artificial Intelligence,” “Machine Learning,” “Product Improvement,” “Aggregated Data,” and “Deletion or Return of Data.”

Summary of Important Changes

1. Scope expanded from website use to Datadog products and platform capabilities

The prior terms governed access to Datadog’s website, related websites, and website content. The revised language appears to cover use of a much broader range of Datadog offerings, including:

• Monitoring, observability, security, infrastructure, application, digital experience, software delivery, and service-management products
• Datadog’s AI offerings, including Bits AI, Chat, Agent Builder, Code, Security Analyst, Agent Observability, MCP Server, and related capabilities
• The Datadog platform, integrations, applications, dashboards, automation, and other product features

Risk: Website terms may now be presented as governing product or platform use, despite the prior terms expressly excluding the hosted Service. This could create uncertainty about which terms apply to customer use, particularly where the Master Subscription Agreement (MSA), product terms, or other contracts also apply.

2. AI-related products and functionality are expressly introduced

The revised text adds numerous AI-related products and references, including:

• “AI”
• “AI Impact”
• “AI-powered” security and observability capabilities
• Bits AI, Chat, Agent Builder, Code, and Security Analyst
• Agent Observability and MCP Server
• Language stating that AI capabilities may “improve model performance” and help identify root causes or anomalies

Risk: These additions expand the contractual scope to AI functionality and may subject users of free trials or individual AI features to additional Datadog agreements or policies.

3. No clear customer-data training authorization is shown

The diff does not clearly add language expressly stating that Datadog may use customer data, prompts, logs, code, telemetry, or other content to train, fine-tune, or improve general-purpose AI models.

The references to improving “model performance” could relate to AI functionality, but the provided diff does not clearly establish:

• Whether customer data is used for model training
• Whether data is used to train Datadog’s models or third-party models
• Whether data is aggregated, de-identified, or retained
• Whether customers can opt out
• Whether customer content is used for product improvement beyond providing the service

These issues should be checked against the MSA, applicable data-processing agreement, privacy policy, AI-specific terms, and product documentation.

4. Contract hierarchy and applicability are less clear

The revised language appears to state that users accessing the Service—even through a free trial—are subject to the MSA or another written contract. It also introduces many product-specific references without clearly identifying which terms control in the event of conflict.

Recommended review: Confirm that the MSA and AI/product terms expressly address customer-data ownership, confidentiality, permitted processing, model training, subprocessors, retention, deletion, and opt-out rights.

Summary

The provided diff does not include the actual contractual language. It only states:

> “Added approximately 4076 words to the document”

Accordingly, it is not possible to reliably identify:

• Changes to customer data rights or permitted uses
• Whether customer data may be used to train, fine-tune, evaluate, or improve AI models
• Any new data-retention, sharing, or disclosure rights
• Changes to confidentiality, security, or privacy obligations
• New intellectual-property ownership or licensing terms
• Customer opt-out, deletion, or data-control rights
• Allocation of liability for AI-generated outputs or misuse of data

Important limitation

The statement that approximately 4,076 words were added indicates a substantial revision, but it does not reveal whether the additions are favorable, unfavorable, or unrelated to data and AI issues. No legal conclusions can be drawn from the word count alone.

Information needed

Please provide the full redline or the actual additions, deletions, and replacements using the stated notation:

• Additions: {new text}
• Deletions: [deleted text]
• Replacements: [old text]{new text}

Once the substantive diff is provided, it can be analyzed for AI-training permissions, customer-data licensing, confidentiality, privacy, security, retention, third-party disclosures, and related legal risks.

Summary of Important Changes

1. Expanded scope and contracting parties

• The Terms now cover the Datadog website and any related websites, collectively defined as the “Site.”
• Datadog is defined to include Datadog, Inc. and its affiliates. This may broaden which entities can enforce the Terms or be treated as the contracting party.
• The Terms now expressly apply to accessing or using the Site and its content, rather than merely referring to use of a particular Datadog platform or product.

2. Broader product and service references

• The revised language appears to incorporate a much broader range of Datadog offerings, including infrastructure monitoring, security, digital experience, software delivery, service management, and AI-related products.
• These product references appear primarily to be website navigation or descriptive content, but they may make it less clear which Datadog offerings are governed by these Website Terms versus a separate commercial agreement.

3. Hosted services and free trials

• The Terms continue to state that use of Datadog’s hosted application, infrastructure-monitoring service, and other services (the “Service”) is not governed by the Website Terms.
• Instead, Service use is subject to Datadog’s Master Subscription Agreement, including when accessed through a free trial.
• The revised language more clearly states that use of the Service—including free, trial, or other access—may trigger the Master Subscription Agreement. Customers should review that agreement because data-use, confidentiality, AI, security, and liability provisions may be located there rather than in these Website Terms.

4. AI-related changes

• The revised website content adds or highlights AI-related offerings, including AI Impact, Bits, Agent Builder, Datadog for AI, and related AI capabilities.
• However, the diff does not identify an express new provision allowing Datadog to use customer data, customer content, logs, telemetry, prompts, outputs, or other Service data to train AI models.
• The existing “Your Content” provision grants Datadog a broad, perpetual, worldwide, royalty-free license to use content posted or made available on the Site, but it excludes personal information governed by the Privacy Policy. This provision appears to concern Site content and is not clearly expanded here to customer data submitted to the hosted Service.
• Because Service use is redirected to the Master Subscription Agreement, any AI-training rights or restrictions may exist there instead. Customers should specifically check that agreement, applicable product terms, privacy documentation, and any AI-specific terms.

5. Key risks and recommended actions

• Affiliates may gain contractual rights or responsibilities.
• The boundary between Site use and Service use may be harder to interpret.
• Free-trial users may be bound by the Master Subscription Agreement even without a paid subscription.
• Confirm which agreement governs customer data and whether it permits:
• model training or model improvement;
• use of data across customers;
• human review or provider access;
• retention of prompts, logs, or outputs; and
• use of data to train third-party or general-purpose models.

Executive Summary

The diff appears heavily contaminated by duplicated website navigation, historical page labels, and form text. The substantive privacy-policy changes are difficult to isolate. Based on the identifiable text, the revised policy:

• Is dated March 12, 2026.
• Adds or clarifies that personal information is “protected.”
• Expands the description of the covered services to include Datadog products, services, events, websites, and applications.
• Adds a statement that the policy explains how individuals may control or exercise rights relating to their personal information.
• Includes expanded descriptions of data collection, use, sharing, retention, international transfers, and U.S./EEA/UK/Swiss privacy rights.

AI Training and Model Use

No clear express authorization to train AI models

The diff does **not clearly add language stating that customer data, logs, traces, metrics, prompts, outputs, or other Product information may be used to train, fine-tune, or improve Datadog’s AI models. The phrase appearing near the end—referencing “improve model performance”—is embedded in apparently corrupted marketing/navigation text and does not appear to be a substantive privacy-policy provision.

The diff also does not clearly add:

• A right for Datadog to use customer content to train models;
• A commitment that customer data will not be used for training;
• A distinction between customer data and consumer personal information;
• Opt-out, deletion, or exclusion controls specifically for AI training;
• Restrictions on using customer data to train third-party or shared models.

Existing broad improvement language remains important

The policy states that Datadog may use information collected—including “Product information”—for research and to improve and create new Datadog Products. Product information includes information submitted through Datadog products, including logs, traces, and metrics, which may contain personal information.

This broad language could potentially support development or improvement of AI-enabled features, depending on how Datadog interprets “improve” and “create new products.” It is not, however, an explicit AI-training grant.

Other Important Risks

• Customer content may contain personal information: Customers are responsible for the data placed in logs, traces, and metrics, potentially creating compliance obligations.
• Broad sharing rights: Personal information may be shared with service providers, affiliates, advertising/analytics providers, and in business transfers.
• International transfers: Data may be transferred to or accessed from the United States and other countries.
• Retention flexibility: Datadog may retain data for legal, contractual, legitimate-business, archival, or dispute-resolution purposes.
• Marketing and profiling: The policy permits personalized advertising, behavioral/targeted advertising, and related analytics, subject to stated opt-out rights.
• Policy changes: Changes become effective when posted, without an express individualized notice requirement.

Bottom Line

The diff does not establish a clear new right to train AI models using customer data. Nevertheless, the policy’s broad product-improvement language and inclusion of logs, traces, and metrics create ambiguity that should be addressed in a separate AI/data-use provision or customer agreement.

Executive Summary

The diff appears heavily corrupted by website-navigation content being inserted into the legal text. As a result, many apparent “changes” are not coherent contractual edits and should be verified against a clean, rendered version of both documents before relying on them.

Important Apparent Changes

1. Expanded scope of covered products and services

The revised text appears to replace the narrow reference to the “Site” and Datadog’s hosted monitoring service with a much broader collection of Datadog offerings, including:

• Infrastructure, application, database, log, security, and digital-experience products;
• Software delivery and developer tools;
• Service management and incident-management tools;
• Datadog AI products, including Bits, Agent Builder, Bits Code, Security Analyst, Agent Observability, and MCP Server;
• “AI Impact” and other AI-enabled capabilities.

Risk: If effective, the terms may apply to a substantially broader range of products than the prior website-only terms. However, the insertion is not grammatically reliable and may simply reflect scraped website menus rather than operative contract language.

2. Potential change to governing agreement for the hosted service

The prior language clearly stated that the Website Terms did not govern Datadog’s hosted application and infrastructure-monitoring service, which was instead governed by the Master Subscription Agreement (“MSA”) or another written contract.

The revised text appears to add or expand references to Datadog’s platform, AI capabilities, and other services, and may state that use of those services is governed by the MSA or another written agreement.

Risk: Customers should confirm whether AI products are expressly excluded from these Website Terms and governed by the MSA, or whether the revised terms attempt to apply directly to them. Conflicting terms could create uncertainty regarding data rights, confidentiality, security, liability, and dispute procedures.

3. AI-related provisions and model training

The diff introduces numerous references to AI products and capabilities, but it does not clearly show a substantive clause stating that customer data may—or may not—be used to train AI models.

In particular, the diff does not clearly add language addressing:

• Whether customer content, telemetry, prompts, code, logs, or outputs may be used for model training;
• Whether data is used to train Datadog’s models, third-party models, or both;
• Opt-out rights;
• De-identification or aggregation;
• Retention and deletion of training data;
• Ownership of inputs, outputs, or model improvements;
• Restrictions on using confidential or personal information for training.

Risk: The absence of a clear training authorization is not necessarily protective. AI data-use terms may exist in the MSA, product-specific terms, privacy policy, or documentation incorporated by reference. Those documents should be reviewed together.

4. Drafting and enforceability concern

The revised material contains extensive navigation labels, product names, URLs, and marketing text embedded within legal sentences. This creates ambiguity about what is actually contractual text and what is webpage content.

Recommended action: Obtain the clean October 25, 2024 version and compare it with the prior version using a conventional redline. Do not treat the apparent AI-related additions as operative until confirmed in the signed or published terms.

Summary

The provided diff only states:

> “Added approximately 4076 words to the document”

It does not include the actual added, deleted, or replaced legal language. As a result, it is not possible to determine:

• What contractual terms changed;
• Whether customer data may be used to train, fine-tune, test, or improve AI models;
• Whether the customer granted a license or other rights in its data;
• Whether data may be shared with affiliates, service providers, or third parties;
• Whether confidential information, personal data, or uploaded content is covered;
• Whether the provider may retain data after termination;
• Whether the customer can opt out of model training;
• Whether the provider disclaims responsibility for model outputs or data security; or
• Whether liability, indemnity, audit, deletion, or regulatory obligations changed.

Key limitation

The statement that approximately 4,076 words were added is not itself a legally reviewable change. The actual text of the additions, deletions, and replacements is required to assess their legal effect.

Information needed

Please provide the full redline or the specific added and deleted clauses, particularly any provisions containing terms such as:

• “train,” “training,” “fine-tune,” “improve,” or “develop”;
• “artificial intelligence,” “machine learning,” or “models”;
• “customer data,” “content,” “inputs,” or “outputs”;
• “license,” “ownership,” or “intellectual property”;
• “retain,” “store,” “delete,” or “de-identify”; and
• “service providers,” “affiliates,” or “third parties.”

Once the actual language is provided, the changes can be analyzed for new permissions, restrictions, customer rights, and associated legal risks.

Executive Summary

The diff appears to update Datadog’s Privacy Policy to a March 12, 2026 version and substantially changes the surrounding website/navigation content. Much of the diff is corrupted by webpage formatting and appears to include duplicated or unrelated site text. The clearest substantive policy changes are summarized below.

Important Changes

1. Broader description of covered services

The policy now expressly applies to Datadog products, services, events, websites, and applications that link to the policy, collectively defined as “Datadog Products.” It also references a broad range of newer offerings, including AI, security, developer, analytics, and digital-experience products.

Risk: The policy’s scope may extend to data processed through newly added products without requiring a separate notice.

2. Customer product data expressly included

The policy identifies “Product information” as personal information submitted through Datadog Products, including information in:

• Logs
• Traces
• Metrics
• Names, email addresses, IP addresses, online identifiers, and mailing addresses

Risk: Customer telemetry may contain personal, confidential, regulated, or commercially sensitive information. The policy does not distinguish clearly between data provided by an individual customer and data submitted by the customer’s end users or other third parties.

3. Use for research, improvement, and new products

Datadog states that it may use information collected—including Product information—to:

• Monitor and analyze trends, usage, and other activities;
• Improve Datadog Products; and
• Create new products and services.

It may also link or combine information with data obtained from other sources.

Risk: This is a broad secondary-use right. It could permit use of customer data beyond merely delivering the contracted service, subject to other contractual terms that are not included in this diff.

AI Model-Training Analysis

The diff does not clearly add an express statement that customer data will be used to train, fine-tune, or evaluate AI models. It also does not clearly state that customer data will not be used for those purposes.

However, the combination of:

• Broad rights to use Product information for “research,” improvement, and creating new products;
• References to AI products and AI-related functionality; and
• No specific customer-data AI restriction or training opt-out in the visible text

creates ambiguity and a potential risk that customer data could be used in connection with AI development or model improvement.

Recommended clarification: Customers should confirm in the governing agreement or data-processing terms whether customer content, logs, traces, and metrics may be used to train or improve Datadog’s or third-party AI models, whether de-identification is required, and whether an opt-out is available.

4. Additional privacy-rights language

The policy describes access, correction, deletion, restriction, objection, portability, consent withdrawal, and certain U.S. opt-out rights, including rights concerning targeted advertising, automated profiling, sale/sharing, and sensitive information.

Risk: These rights appear directed primarily to individuals and may not give business customers control over operational data unless contract terms provide it.

5. Broad sharing and retention provisions

The policy permits sharing with service providers, affiliates, advertising and analytics providers, legal authorities, and transaction counterparties. It also permits retention for legal or legitimate business purposes and continued use of anonymized information.

Risk: The policy does not specify precise retention periods or identify all providers and downstream uses.

Summary

The supplied diff does not include the actual added contract language. It only states:

> “Added approximately 4103 words to the document”

Accordingly, it is not possible to identify the legal changes, new obligations, allocation of risk, or modifications to customer-data rights.

Customer Data and AI Training

No substantive language addressing any of the following is provided:

• Whether customer data may be used to train, fine-tune, or improve AI or machine-learning models
• Whether prompts, inputs, outputs, personal data, or confidential information may be retained
• Whether data is used for the provider’s general product development or shared across customers
• Whether customer data is anonymized, de-identified, aggregated, or otherwise protected
• Whether the provider may use subcontractors or third-party AI services
• Whether customers can opt out of AI training or data-use practices
• Ownership and permitted use of model outputs or derivatives
• Security, deletion, retention, audit, or breach obligations relating to AI systems
• Restrictions on using customer data to train models that benefit competitors

Risk Assessment

Because the added text is missing, the following potentially important risks cannot be evaluated:

1. Expanded data-use rights: New language could authorize broader use of customer content beyond providing the contracted services.

2. AI-training rights: The provider may have added rights to use customer data for model training, testing, evaluation, or improvement.

3. Confidentiality and privacy exposure: New exceptions may permit disclosure to affiliates, vendors, or AI providers.

4. Ownership changes: The additions may alter ownership or licensing rights in customer content, inputs, outputs, or trained-model materials.

5. Reduced control: The contract may limit deletion rights, opt-out rights, or the customer’s ability to restrict secondary uses.

6. Compliance impact: Broader processing could create additional obligations under privacy, sector-specific, or data-localization laws.

Information Needed

Please provide the full redline or the actual 4,103 words added to the document. Without the added clauses, no reliable comparison or legal-risk analysis can be performed.

Summary of Important Changes

1. Broader scope of the Terms

• “Site” is expanded to include the Datadog website and related websites.
• Datadog, Inc. and its affiliates are collectively defined as “Datadog.” This may allow affiliates to assert rights or enforce obligations under the Terms.
• The Terms apply to anyone accessing or using the Site or its Content, rather than only users of a particular Datadog platform or product.
• The product and navigation references indicate coverage of Datadog’s expanding product portfolio, including AI-related products and services. However, these references do not necessarily create contractual rights to use every listed product.

2. New or clarified authority and responsibility provisions

• A person entering the Terms for an organization represents that they have authority to bind that organization.
• The definition of “you” includes both the individual and the represented entity.
• Datadog and the customer are defined as the “Parties,” and each as a “Party.”
• This creates potential organizational liability if an employee or contractor accepts the Terms without proper authority.

3. Hosted services moved under separate agreement

• The Terms state that the hosted application, infrastructure-monitoring service, and other Datadog services (collectively, the “Service”) are governed by Datadog’s Master Subscription Agreement (MSA) or another written contract.
• This expressly applies even to free trials.
• Users may therefore be subject to different terms depending on whether they are merely browsing the Site or using Datadog’s products. Customers should confirm which agreement governs data handling, confidentiality, security, intellectual property, and AI functionality.

4. Customer content and data-use risks

• The “Your Content” provision grants Datadog a very broad, perpetual, worldwide, irrevocable, royalty-free, sublicensable license to use posted or uploaded Site content, including to reproduce, modify, distribute, sell, and otherwise use it without restriction.
• The provision excludes personal information subject to the Privacy Policy, but it does not clearly address confidential business information, telemetry, logs, source code, prompts, outputs, or other data submitted through Datadog products.
• The indemnity provision also covers content the customer posts, uploads, interfaces with, transfers, processes, uses, or stores in connection with the Site.

5. AI-model training

• The diff adds or references numerous AI products and capabilities, including AI Impact, Bits AI, Agent Builder, Bits Code, Bits Security Analyst, Datadog for AI, and Agent Observability.
• No express provision was identified stating whether customer data, prompts, logs, telemetry, source code, outputs, or other content may or may not be used to train Datadog’s or third-party AI models.
• Accordingly, the revised Terms do not provide a clear customer-data training commitment or opt-out. This issue should be checked in the MSA, Data Processing Addendum, Privacy Policy, and product-specific AI terms.

6. Other continuing risks

• Datadog retains broad suspension, termination, monitoring, and disclosure rights, including disclosure of customer data to authorities where appropriate.
• Liability remains highly limited, including an aggregate cap of $200 under these Site Terms.
• Datadog may modify the Terms by posting a revised version, with continued access constituting acceptance.

Summary

The provided diff does not include the actual added or modified legal language. It only states:

> “Added approximately 3475 words to the document”

Accordingly, it is not possible to identify:

• New contractual rights or obligations
• Changes to liability, indemnification, warranties, or termination
• Changes to confidentiality or data-security obligations
• Changes to intellectual-property ownership or licensing
• Changes concerning use of customer data to train, fine-tune, evaluate, or improve AI models
• Whether customer data may be shared with affiliates, vendors, or other third parties
• Any new retention, deletion, or data-location provisions

AI Training and Customer Data

No conclusions can be drawn regarding AI-model training because the relevant added text was not provided. The phrase “added approximately 3475 words” does not itself establish whether customer data may be:

• Used to train or improve models
• Used only in de-identified or aggregated form
• Used for service delivery or product development
• Reviewed by human personnel
• Shared with model providers or subprocessors
• Retained after termination
• Excluded from training upon customer request or opt-out

Required Information

Please provide the full redline or the actual text of the additions, deletions, and replacements. Once provided, the analysis can identify the material legal changes, explain their practical effect, and specifically flag any provisions authorizing or restricting AI training on customer data.

Summary of Available Changes

The supplied diff states only that approximately 4,104 words were removed. It does not identify which provisions were deleted or provide the replacement language.

AI Training and Data Use

• No conclusions can be drawn about whether the agreement now permits, restricts, or prohibits using customer data to train AI models.
• The deleted material may have included provisions addressing:
• Whether customer content may be used to train, fine-tune, or improve AI models;
• Whether usage data, prompts, outputs, or metadata may be used for those purposes;
• Opt-out rights or customer consent requirements;
• De-identification, aggregation, or anonymization standards;
• Human review or model-training access;
• Restrictions on using confidential, personal, or regulated information;
• Ownership of customer data and AI-generated outputs; or
• Security, retention, and deletion obligations.

Potential Legal and Commercial Risks

Because the content of the deletion is unavailable, the principal risk is loss of contractual protections or clarity. In particular, deleting data-use provisions could:

• Remove an express prohibition on using customer data for AI training;
• Eliminate a customer’s right to opt out or require prior consent;
• Broaden ambiguity regarding whether prompts, outputs, or usage data qualify as “customer data”;
• Remove confidentiality, privacy, security, or deletion commitments;
• Affect compliance with data-protection, sector-specific, or confidentiality obligations;
• Make it harder to determine ownership or permitted use of AI-generated outputs; or
• Create interpretive uncertainty if remaining provisions refer to terms or rights that were deleted.

Alternatively, the deletion may have removed permissive language and therefore reduced the provider’s rights. The direction of the change cannot be determined from the available information.

Required for Definitive Analysis

Please provide either:

1. The full redline showing the deleted and added language; or

2. The prior and revised versions of the agreement.

Without that text, no reliable determination can be made regarding the legal effect of the changes, including any changes to AI-model training or customer-data use.

Summary

The supplied diff does not include the actual contractual language. It only states:

> “Added approximately 4104 words to the document”

As a result, it is not possible to identify:

• The specific legal changes;
• New customer obligations or provider rights;
• Changes to liability, indemnities, confidentiality, intellectual property, or termination;
• Any changes to how customer data is collected, accessed, retained, disclosed, or used; or
• Whether customer data, prompts, outputs, or usage information may be used to train, fine-tune, evaluate, or improve AI models.

AI Training and Data-Use Risk

No conclusions can be drawn about AI-model training from the information provided. The added text may contain important provisions concerning:

• Use of customer content for model training or service improvement;
• Whether training is automatic, optional, or subject to an opt-out;
• Use of data in de-identified, aggregated, or identifiable form;
• Retention of prompts, outputs, and related metadata;
• Provider ownership or licensing of customer inputs and outputs;
• Sharing data with affiliates, subprocessors, or third-party AI providers; and
• Security, deletion, and confidentiality protections applicable to training data.

These provisions should be reviewed carefully because broad rights to use customer data for AI training could conflict with confidentiality obligations, privacy laws, sector-specific requirements, or the customer’s commitments to its own users.

Information Needed

Please provide the actual redline or the full text of the approximately 4,104 added words, using the stated notation:

• {additions}
• [deletions]
• []{} replacements

Without the substantive text, a reliable legal-risk analysis is not possible.

Summary of Changes

Overall Change

• Approximately 3,475 words have been removed from the document.
• No replacement language or specific deleted provisions were provided.
• Because the deleted text is not shown, the legal and commercial impact cannot be determined reliably.

AI Training and Customer Data

• The diff does not identify whether any provisions concerning AI training, machine learning, or use of customer data were deleted.
• If the removed text included an AI or data-use clause, the deletion could potentially:
• Remove the provider’s permission to use customer data, prompts, outputs, or usage data to train or improve AI models.
• Remove restrictions on using customer data for model training.
• Eliminate customer consent or opt-out rights.
• Remove confidentiality, anonymization, retention, or deletion requirements applicable to training data.
• Change whether the provider may use aggregated, de-identified, or derived data.
• Affect ownership or licensing rights relating to customer inputs, outputs, or model improvements.
• Conversely, the deletion could have removed a provision that previously authorized broad AI training use, which might improve customer privacy—but this cannot be confirmed without the deleted wording.

Other Potential Legal Risks

The removed material may also have contained important provisions relating to:

• Confidentiality and data security
• Data processing and privacy-law compliance
• Intellectual-property ownership and licenses
• Warranties, indemnities, and liability limitations
• Service levels and termination rights
• Data retention, return, and deletion
• Audit, compliance, or regulatory obligations
• Customer restrictions and provider rights

Deleting these provisions could create ambiguity, eliminate protections, or cause default contractual rules to apply. It could also create inconsistencies with provisions that remain in the document.

Recommended Review

The actual deleted text—or a redline showing the specific deletions—is required for a meaningful legal analysis. In particular, review any removed language containing terms such as “AI,” “artificial intelligence,” “machine learning,” “train,” “improve,” “model,” “inputs,” “outputs,” “customer data,” “usage data,” “aggregated,” “de-identified,” or “service improvement.”

Structured Summary of Important Changes

1. Major expansion of the Terms’ scope

• The terms are no longer limited to Datadog’s website and website content.
• They now appear to cover Datadog’s broader products, platform, monitoring, security, digital experience, software delivery, service-management, and AI capabilities.
• The scope also appears to include products such as Bits AI, Agent Builder, Code, Security Analyst, Agent Observability, MCP Server, and other AI-related functionality.

Risk: Users may become bound by website terms when using a much broader range of Datadog products, including services that may previously have been governed only by a subscription agreement or product-specific terms.

2. AI services and incorporation of other agreements

• The revised language expressly refers to Datadog’s AI services and capabilities, including use through free trials.
• Use of the Service is stated to be subject to Datadog’s Master Subscription Agreement or another written contract, rather than being governed solely by these website terms.
• The terms appear to distinguish between the general platform and hosted applications or infrastructure services, but the revised drafting is difficult to parse and may create uncertainty about which agreement controls.

Risk: Customers may be subject to multiple overlapping documents. The priority rules between these terms, the MSA, product terms, AI-specific terms, and other written contracts should be confirmed.

3. Customer data and AI-model training

• The diff introduces or references broad categories of customer-related data, including:
• Observability data
• Logs and monitoring data
• Security and sensitive data
• Code and software
• Data processed by AI-powered security and observability tools
• However, the diff does not clearly state whether Datadog may use customer data to train, fine-tune, evaluate, or improve general-purpose or Datadog AI models.
• It also does not clearly provide an opt-out, deletion right, retention period, or restriction against using customer data for model training.

Important risk: The expanded AI scope could increase the practical possibility that customer data is processed by AI features, but the diff alone does not establish clear contractual limits on AI training. Customers should review the MSA, privacy/data-processing terms, AI-specific terms, and product documentation for:

• Whether customer data is used for model training or service improvement;
• Whether data is isolated between customers;
• Whether prompts, outputs, code, logs, or telemetry are retained;
• Whether human review or third-party model providers are involved; and
• Available opt-out or data-deletion mechanisms.

4. Broader customer authority representations

• The customer must represent that it is legally able to enter into the terms and, when acting for an organization, has authority to bind it.
• “You” expressly includes both the individual and the represented entity.
• The revised scope may apply to organizations using Datadog products, including AI and security tools.

Risk: An employee’s acceptance or use could create obligations for the company, including responsibility for data submitted to AI or monitoring services.

5. Free trials and product access

• The revised terms expressly cover access through free trials and possibly free or limited AI functionality.

Risk: Trial use may trigger contractual obligations, data-processing consequences, and incorporation of the MSA or other terms even before a paid subscription begins.

6. Drafting and version-control concern

• The diff contains substantial concatenated product-menu text and appears technically corrupted or improperly rendered.
• It is unclear whether all inserted product names are operative legal language or website navigation content.

Recommendation: Obtain the clean, final version and confirm the effective date, controlling agreement, and exact AI-data-use provisions before acceptance.

Summary

The diff indicates that approximately 3,475 words were removed, but it does not identify which provisions were deleted. Because no replacement language or section headings are provided, the substantive legal impact cannot be determined reliably.

Key Risks

• Loss of protections: Deleted language may have contained important customer rights, warranties, indemnities, confidentiality obligations, security requirements, audit rights, service commitments, or remedies.
• Changed allocation of liability: Removing limitation-of-liability exceptions, indemnities, or responsibility provisions could materially shift risk to the customer.
• Reduced data protections: Deletions may affect restrictions on collection, use, retention, disclosure, security, deletion, or cross-border transfer of customer data.
• Ambiguity and interpretation risk: If deleted provisions leave cross-references, definitions, or exceptions incomplete, the agreement may become internally inconsistent or subject to broader interpretations.

AI Training and Model Use

The supplied diff does not include the deleted text, so it is impossible to determine whether the changes:

• Permit or prohibit using customer data to train, fine-tune, improve, or evaluate AI models;
• Restrict the use of prompts, inputs, outputs, personal information, or confidential information for training;
• Apply only to de-identified, aggregated, or anonymized data;
• Allow human review or use by service providers for model development;
• Provide an opt-out, consent requirement, or customer control over training;
• Require deletion of customer data from training datasets or model weights; or
• Address ownership, confidentiality, security, or intellectual-property rights in AI-generated outputs.

If the removed provisions previously restricted AI training or secondary use, their deletion could materially expand the provider’s ability to use customer data. Conversely, deletion of a permission may limit that use. The direction of the change cannot be established without the actual text.

Recommended Review

Obtain the redlined deletions or the prior and revised versions, including section headings and surrounding language. Priority should be given to provisions concerning:

1. Customer data and confidential information;

2. AI, machine learning, model training, or service improvement;

3. Privacy and security;

4. Data retention and deletion;

5. Ownership and intellectual property;

6. Liability, indemnities, and warranties; and

7. Subprocessors and third-party service providers.

Summary of Available Change

The diff only states that approximately 4,076 words were removed. It does not identify which clauses, sections, or defined terms were deleted. As a result, the specific legal and commercial effects cannot be determined from the information provided.

Key Risks Created by the Deletions

• Loss of contractual protections: Deleted language may have contained warranties, indemnities, confidentiality obligations, security requirements, service commitments, limitations of liability, audit rights, or termination rights.
• Unclear governing obligations: Removing definitions or operative provisions may create ambiguity about what each party must do and when.
• Potential imbalance: Deletions may remove protections for the customer without adding replacement obligations for the provider.
• Reduced enforcement ability: Deleted notice, dispute-resolution, governing-law, or remedies provisions could make breaches harder or more costly to address.
• Data and security exposure: Removed restrictions concerning data handling, retention, subprocessors, security controls, breach notification, or deletion could materially increase privacy and cybersecurity risk.

AI Training and Customer Data

The provided diff does not contain enough information to determine whether the contract’s treatment of customer data for AI training has changed.

In particular, it is impossible to establish whether the deletion:

• Removes a prohibition or restriction on using customer data to train, fine-tune, evaluate, or improve AI models;
• Removes a requirement to obtain the customer’s consent before such use;
• Eliminates limits on using customer content, prompts, outputs, metadata, or usage data;
• Removes obligations to de-identify, aggregate, or pseudonymize data;
• Weakens confidentiality, ownership, or intellectual-property protections;
• Deletes opt-out, deletion, retention, or model-unlearning rights; or
• Removes restrictions on sharing customer data with affiliates, vendors, or model providers.

Recommended Review

The full redline or the text of the deleted provisions should be reviewed, with particular attention to sections titled:

• Data Use, Privacy, or Customer Data
• Artificial Intelligence or Machine Learning
• Confidentiality
• Security and Data Processing
• Intellectual Property
• Retention and Deletion
• Subprocessors or Third-Party Services

Until the deleted text is available, the amendment should be treated as potentially material, especially if the agreement previously contained express limits on using customer data to train or improve AI models.

Key Changes and Risks

1. Scope expanded from website use to Datadog products and platform capabilities

• The original terms governed access to Datadog’s website and website content.
• The revised language appears to cover access to or use of Datadog’s broader integrated platform, monitoring, security, software delivery, service management, AI, and related capabilities, including free trials.
• The revision adds or references numerous products, including Bits AI, AI Impact, Agent Builder, Code, Security Analyst, Agent Observability, MCP Server, and other AI-related features.

Risk: Customers may unintentionally accept these terms while using product functionality, rather than merely browsing the website. The boundary between these Website Terms and the Master Subscription Agreement (“MSA”) is less clear.

2. New express treatment of AI-related services

• The revised text appears to state that users of Datadog’s AI and other platform services are subject to the MSA or another written agreement, even where access is through a free trial.
• It also introduces AI-related product descriptions and references to AI-powered analysis, investigation, monitoring, and model performance.

Risk: AI features may become contractually incorporated into the service framework without a detailed explanation in these terms of data handling, model governance, output ownership, confidentiality, or prohibited uses.

3. Customer authority and responsibility language remains broad

The revised provisions continue to require users entering terms for an organization to represent that they have authority to bind it. “You” may include both the individual and the represented entity.

Risk: An employee’s use of a free trial or AI feature could potentially create obligations for the company, depending on how the terms are presented and accepted.

4. Customer data and AI-model training

• The diff does not clearly add an express right for Datadog to use customer data to train AI models.
• The visible “Your Content” license in the supplied material grants Datadog broad rights to use content posted on the Site, but excludes personal information governed by the Privacy Policy. It does not expressly mention AI training.
• However, the revised scope appears to bring AI products and platform functionality into closer proximity with these terms, while the supplied text does not clarify whether telemetry, prompts, logs, code, security data, outputs, or other customer data may be used for:
• training or fine-tuning models;
• improving Datadog or third-party models;
• human review or product development; or
• creation of aggregated or de-identified datasets.

Risk: The absence of a clear AI-data provision creates uncertainty rather than confirming that training is prohibited. Customers should review the MSA, product-specific terms, privacy documentation, and any AI terms or data-processing addendum for the controlling position.

5. Practical review points

Customers should seek confirmation of:

• whether customer data is used for AI training or tuning;
• whether use is opt-in, opt-out, or automatic;
• whether data is shared with model providers;
• retention, deletion, and de-identification standards; and
• whether contractual confidentiality and data-processing protections apply to AI inputs and outputs.

Summary

The provided diff only states:

> “Added approximately 4076 words to the document”

It does not include the actual added, deleted, or replaced contract language. As a result, it is not possible to reliably identify:

• Changes to the parties’ rights or obligations
• New fees, liability provisions, warranties, or termination rights
• Changes to confidentiality, privacy, or security obligations
• Whether customer data may be used to train, fine-tune, evaluate, or improve AI models
• Whether such use is optional, subject to consent, or automatic
• Whether customer data is anonymized, aggregated, retained, or shared with affiliates or vendors
• Any restrictions on using outputs or customer content for model development

Information Needed

Please provide the full redlined text, including:

• Additions in {curly braces}
• Deletions in [square brackets]
• Replacements in []{replacement} format

The statement that approximately 4,076 words were added is not sufficient to assess the legal impact or identify AI-training provisions.

Summary of Important Changes

1. Broader scope of the Terms

• The Terms now apply to Datadog’s Site, defined to include the website and related websites operated by Datadog or its affiliates.
• The prior language primarily addressed the website and stated that the Terms did not govern Datadog’s hosted application and infrastructure-monitoring services.
• The revised language appears to cover access to or use of a much broader range of Datadog offerings, including:
• Infrastructure, application, security, software-delivery, digital-experience, and service-management products;
• Platform capabilities and integrations; and
• AI products and features, including Bits, Agent Observability, AI Integrations, AI Agents, and related tools.

Risk: Users may now be contractually bound by these Website Terms when interacting with product pages, embedded tools, trials, or AI-related functionality, even where the user may believe they are only browsing the website.

2. Service use is redirected to other agreements

• The revised text states that use of the Service is governed by Datadog’s Master Subscription Agreement (MSA) or another written agreement, including use during a free trial.
• AI features accessed through the Service are stated to be subject to the MSA or another separately agreed contract.

Risk: Important rights and obligations may now be located outside these Terms. Customers should review the MSA and any product-specific agreements to determine which terms control, especially for data rights, confidentiality, security, retention, liability, and AI usage.

3. AI-related changes

• The revision expressly adds AI-related products and services to the Datadog offering covered or referenced by the Terms.
• It also appears to distinguish AI features from the general Site and to subject them to the MSA or another contract.
• The revised language may allow access to AI functionality through free trials or platform features, potentially bringing users within additional contractual terms.

AI training and customer-data use

• The diff does not expressly add language stating that customer data, Customer Content, telemetry, logs, prompts, outputs, or other inputs may be used to train Datadog’s AI models.
• It also does not expressly prohibit such training or provide a new opt-out.
• The existing “Your Content” license grants Datadog broad rights to use content submitted to the Site, but excludes personal information subject to the Privacy Policy. It is unclear from the diff whether that license applies to customer data submitted through the Service or AI features.
• The controlling MSA, privacy documentation, AI terms, and product-specific agreements should therefore be reviewed for any training, model-improvement, de-identification, or service-development rights.

4. Customer authority and responsibility

• Users entering into the Terms for an organization expressly represent that they have authority to bind it.
• “You” includes both the individual and represented entity.
• Customers remain responsible for authorized and unauthorized use of their accounts and for violations by end users.

Risk: Individual users may unintentionally create obligations for their employer, particularly when activating AI features or free trials.