Intercom

Summary of Important Changes

1. AI-model training and use of Customer Data

• No express new authorization to train AI models appears in the supplied diff. The revised text does not add language expressly permitting Intercom to use Customer Data to train, fine-tune, or develop artificial-intelligence or machine-learning models.
• The existing data-use language remains important: Intercom may use Usage Data and/or Customer Data that has been de-identified or anonymized to “analyze, improve, support and operate” the Services, including during and after the customer relationship.
• The revision changes quotation marks and formatting around this provision but does not clearly expand or narrow its substantive scope.
• The definition of “Laws” continues to refer expressly to artificial intelligence and the handling of technical or personal data. This is not itself a data-use grant, but it confirms that AI-related legal requirements fall within the agreement’s compliance framework.
• Risk: “Analyze” and “improve” are broad terms. Even without an express model-training right, Intercom might argue that training using properly de-identified or anonymized information is part of improving the Services. Customers seeking to prohibit model training should obtain an explicit contractual restriction covering Customer Data, Usage Data, prompts, outputs, logs, and de-identified information.

2. Agreement formation and customer authority

• The agreement is now framed as being between Intercom and the entity or person ordering, accessing, or using the Services.
• Users acting for a company expressly represent that they are authorized to accept the agreement for that company.
• Acceptance is reinforced through acceptance, access, or use of the Services. The revised language states that the agreement is legally binding and includes mandatory arbitration provisions.
• Risk: Employees or contractors could potentially bind the company through use of the Services, even without a separately signed agreement.

3. Effective date and contracting entity

• The Effective Date is clarified as the earlier of:

1. the customer’s initial access through provisioning, registration, or an order process; or

2. the effective date of the first Order Form referencing the agreement.

• The Intercom contracting entity is determined by the customer’s shipping address:
• Intercom, Inc. for the United States or Brazil; and
• Intercom R&D Unlimited Company for other locations.
• Risk: The revised entity-selection rule may differ from the entity named in an existing Order Form and could affect governing law, dispute procedures, tax treatment, and enforcement.

4. Changes to the agreement

• Intercom may modify the agreement. Unless otherwise specified, changes generally become effective at renewal or when entering a new Order Form.
• Intercom may provide notice through the customer account, email, or other means.
• Continued use after the updated version takes effect constitutes acceptance. Customers may be required to affirmatively accept updated terms before renewal or a new Order Form.
• The prior express right to terminate an affected Subscription Term and receive a prorated refund is retained in substance, although the language is reorganized.
• Risk: The mechanism still permits unilateral changes, including potentially material changes, with continued use functioning as acceptance.

5. Dispute resolution

• Mandatory arbitration and waiver of jury-trial rights are emphasized in the acceptance language.
• Customers should review the arbitration section separately, including limitations on court proceedings and available remedies.

Summary

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

> “Added approximately 83 words to the document”

Accordingly, it is not possible to identify the substantive contractual changes, including any changes concerning customer data or AI-model training.

AI Training and Data Use

• No operative language addressing AI, machine learning, model training, customer data, prompts, inputs, outputs, or data retention is included in the diff.
• It is therefore unclear whether the amendments:
• Permit or restrict using customer data to train AI models;
• Allow data to be used for model improvement or product development;
• Require customer consent or provide an opt-out;
• Apply anonymization, aggregation, or de-identification requirements;
• Permit human review of customer data;
• Allow sharing with affiliates, vendors, or AI service providers;
• Change data-retention or deletion obligations; or
• Grant the provider rights in AI-generated outputs.

Risk Assessment

No specific new legal risks can be assessed from the information provided. The statement that approximately 83 words were added does not reveal the content, scope, or effect of those additions.

Information Needed

Please provide the full redline showing the added, deleted, and replaced language. The substantive wording is necessary to analyze changes to:

• Data ownership and permitted uses;
• AI training and model-improvement rights;
• Confidentiality and security obligations;
• Consent and opt-out mechanisms;
• Intellectual-property rights;
• Data retention and deletion;
• Liability and indemnification; and
• Compliance with applicable privacy and data-protection laws.

Summary

The provided diff states only that “approximately 83 words” were added. It does not include the actual added language or identify where those words appear in the agreement.

Legal and AI-Training Analysis

Because the new text is unavailable, it is not possible to determine whether the amendment:

• Permits the customer’s data, content, prompts, outputs, or personal information to be used to train, fine-tune, or improve AI models;
• Allows use of data in anonymized, aggregated, de-identified, or identifiable form;
• Grants the provider a broad license to reproduce, modify, analyze, retain, or disclose customer data;
• Applies training rights to human review, automated systems, third-party providers, or affiliates;
• Makes AI-training permission optional or subject to an opt-out;
• Changes data-retention, deletion, confidentiality, security, or ownership obligations;
• Expands use of customer data beyond providing the contracted services; or
• Creates regulatory, confidentiality, intellectual-property, privacy, or sector-specific compliance risks.

Key Limitation

No new contractual risk can be reliably identified from a word-count description alone. The legal impact depends on the precise wording, placement, defined terms, and interaction with the rest of the agreement.

Information Needed

Please provide the actual 83 added words, preferably in the original marked-up format. The surrounding section and any referenced definitions would also be useful, particularly if the language concerns:

• “Customer Data” or “Content”;
• “Service Improvement”;
• “Training” or “Machine Learning”;
• “Aggregated” or “De-identified Data”;
• Confidentiality;
• Data retention or deletion; or
• Artificial-intelligence tools and outputs.

Summary of Important Changes

1. Major restructuring of the agreement

The revised version substantially reorganizes the Terms and Policies and incorporates or references a broader set of documents, including:

• Service Definitions
• Customer Obligations
• Security
• Third-Party Platforms and Apps
• Ownership
• Subscription, fees, termination, liability, indemnification, confidentiality, publicity, and general terms
• Acceptable Use Policy, Code of Conduct, trademark terms, anti-corruption terms, DMCA policy, and other regional or product terms

Risk: More terms are incorporated by reference. Customers should review all linked policies because they may contain material obligations, restrictions, or data-processing provisions that are not reproduced in the main agreement.

2. Broader formation and acceptance language

The revised agreement states that it is binding when a person:

• Accepts the agreement;
• Places an order; or
• Accesses or uses any Services.

It also confirms that a person acting for a company represents that they are authorized to bind that company.

Risk: Unauthorized or informal use of the Services may create contractual obligations, including acceptance of referenced policies and mandatory dispute-resolution provisions.

3. Effective date and contracting entity

The revised language defines the Effective Date as the earlier of:

1. The Customer’s initial access to Services through provisioning, registration, or an order process; or

2. The effective date of the first Order Form referencing the agreement.

It also clarifies which Intercom entity is the contracting party based on the Customer’s shipping address, including Intercom, Inc. for the United States/Brazil and Intercom R&D Unlimited Company for other locations.

Risk: The earlier Effective Date may cause contractual obligations to begin before a formal Order Form is signed. The applicable Intercom entity may affect governing law, enforcement, taxes, and available remedies.

4. Reduced customer protection against amendments

Previously, customers could object to certain changes—particularly legal, product, or pricing changes—by terminating the applicable Subscription Term and receiving a refund for the unused portion.

The revised language generally provides that changes take effect upon renewal or entry into a new Order Form. For customers outside the United States and Brazil, continued use after the updated version takes effect constitutes acceptance. For certain changes, the customer may be required to click to accept or otherwise agree before renewal or a new Order Form.

Risk: The express termination-and-refund remedy appears to have been removed or narrowed. Customers may have less ability to reject unfavorable amendments without interrupting service.

5. AI model training and use of Customer Data

The diff does not appear to add a new express right to train AI models using Customer Data.

The existing language continues to distinguish:

• Customer Data, which remains subject to the Customer’s rights;
• Usage Data; and
• Anonymized Data, including de-identified Customer Data.

Intercom may use Usage Data and/or Anonymized Data to analyze, improve, support, and operate its Services during and after the agreement.

AI-related risk: Although the revised text does not expressly mention AI training, the broad purposes of analyzing and improving Services, combined with rights to use Usage Data and Anonymized Data, could potentially encompass development or training of AI systems if those activities are treated as service improvement. Customers seeking to prohibit model training should confirm the position in the Data Processing Addendum, Product Privacy Notice, and any AI-specific terms.

6. Arbitration and liability

The agreement continues to include mandatory arbitration and waiver of jury trials. The revised presentation makes acceptance through access or use especially prominent.

Risk: Disputes may need to be resolved through arbitration rather than court litigation, potentially limiting procedural rights and appeal options.

Summary of Important Changes

1. Major restructuring and renumbering

The revised text substantially reorganizes the Terms of Service:

• The introductory “Terms & Policies” list has been reordered and appears to use different headings and groupings.
• Several provisions are moved, consolidated, or presented in a different sequence.
• Definitions and sections are renumbered or reformatted.
• Quotation marks, apostrophes, and punctuation are standardized throughout.

Risk: Because the diff is heavily interleaved, it is difficult to confirm whether every referenced policy remains incorporated on the same terms. Customers should review the complete revised agreement, particularly the linked policies and product-specific terms.

2. Revised contracting language and effective date

The revised agreement:

• More clearly states that Intercom contracts with the entity or person ordering or accessing the Services.
• Adds or clarifies that a person using the Services for a company represents that they are authorized to accept the Agreement.
• Defines the Effective Date as the earlier of:
• the Customer’s initial access to the Services through provisioning, registration, or an order process; or
• the effective date of the first Order Form referencing the Agreement.
• Clarifies which Intercom entity applies based on the Customer’s shipping address, including Intercom, Inc. and Intercom R&D Unlimited Company.

Risk: The agreement may become binding through access or use before a formal signature or Order Form. The authorization representation may expose the individual or company to disputes if the user lacked authority.

3. Changes to amendment rights

The revised language states that, unless Intercom specifies otherwise, changes generally become effective on renewal of the current Subscription Term or when the Customer enters a new Order Form. Intercom may notify customers through their account, email, or other means.

The prior language apparently provided more detail about:

• advance notice;
• changes taking effect before renewal for legal, product, or pricing reasons;
• a Customer’s ability to object and terminate the applicable Subscription Term; and
• a refund of prepaid fees for the terminated portion.

The revised text instead emphasizes that continued use after the updated version takes effect constitutes acceptance, while also stating that Intercom may require click-through acceptance.

Risk: The Customer’s practical ability to reject unfavorable changes may be less clear or more limited. Customers should confirm whether the termination-and-refund remedy remains available for all material amendments, especially pricing or service changes.

4. AI model training and data use

No express change concerning AI training was identified in the supplied diff.

The existing language continues to refer to Intercom’s use of:

• Usage Data; and
• Customer Data that has been de-identified or anonymized,

for analyzing, improving, supporting, and operating the Services during and after the agreement.

The diff does not add an express authorization to train generative AI or machine-learning models, nor does it add a specific opt-out, retention period, or restriction on using Customer Data for model training.

Risk: The existing broad “anonymized/de-identified data” permission could potentially encompass data used for analytics or model-development activities, depending on the complete agreement and Privacy Policy. Customers requiring a strict prohibition on AI training should seek express contractual language.

5. No substantive commercial or liability changes clearly shown

The visible changes largely preserve provisions concerning:

• data processing;
• ownership;
• subscriptions and renewal;
• warranties;
• availability and support;
• liability limits;
• confidentiality;
• indemnification; and
• mandatory arbitration.

However, the formatting makes a full substantive comparison difficult.

Summary

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

> “Added approximately 83 words to the document”

Without the added, deleted, or replaced text, it is not possible to determine:

• What legal obligations or rights changed;
• Whether liability, indemnity, confidentiality, intellectual property, termination, or other provisions were affected;
• Whether the customer’s data may be used to train, fine-tune, evaluate, or improve AI models;
• Whether data may be shared with affiliates, vendors, or third parties;
• Whether the customer has any opt-out, deletion, or audit rights; or
• Whether the new language creates broader licenses or permits secondary use of customer data.

Information Needed

Please provide the full diff using the stated notation:

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

Once the actual language is provided, the analysis can identify the practical and legal effect of each change, including any provisions that authorize or restrict AI-model training using customer data.

Summary

The provided diff does not include the actual text of the 83 added words. It only states that approximately 83 words were added.

AI Training and Data Use

• No determination is possible regarding whether the customer’s data may be used to train, fine-tune, evaluate, or improve AI models.
• The missing language could potentially address:
• Whether customer data is used for model training;
• Whether usage is automatic or requires customer consent;
• Whether data is anonymized, aggregated, or de-identified;
• Whether human reviewers or service providers may access the data;
• Whether customer content is retained after termination;
• Whether the customer can opt out; and
• Whether these rights apply to inputs, outputs, metadata, or usage data.

Legal Risk Assessment

Because the added wording is not provided, it is not possible to identify:

• New permissions granted to the provider;
• Restrictions on the customer’s data rights;
• Changes to confidentiality or privacy obligations;
• New retention, disclosure, or third-party sharing rights;
• Whether the provider may use data for commercial or product-development purposes; or
• Conflicts with existing data-processing, security, or regulatory commitments.

Please provide the actual added language, including any surrounding provisions needed to understand defined terms and cross-references.

Structured Summary of Important Changes

1. Agreement structure and incorporated policies

• The opening provisions have been substantially reorganized and now present a consolidated list of governing documents, including:
• Service Definitions
• Intercom Services and Customer Obligations
• Security
• Third-Party Platforms and Apps
• Ownership
• Subscription, fees, termination, warranty, availability, support, liability, indemnification, confidentiality, publicity and general terms
• Various policies and product-specific terms incorporated by reference
• The agreement continues to incorporate numerous external policies by reference, including the Data Processing Addendum, Privacy Policy, Additional Product Terms, Support Policy, Cookie Policy, Subprocessors List, Acceptable Use Policy and others. This creates a risk that important obligations may exist outside the main agreement and may change separately.

2. Formation and acceptance

• The revised language more clearly states that the agreement binds a customer when it:
• Accepts the agreement;
• Places an order; or
• Accesses or uses the Services.
• A person using the Services for a company expressly represents that they are authorized to accept the agreement on the company’s behalf.
• The revised language emphasizes that continued use constitutes acceptance and that the agreement contains mandatory arbitration provisions and waives jury trial rights.

Risk: An employee or user may bind the company through access or use, even without a separately signed contract. Mandatory arbitration is now especially prominent and may limit court-based remedies.

3. Modifications to the agreement

• The change-management provision has been reorganized. Unless Intercom specifies otherwise, changes generally become effective upon:
• Renewal of the current Subscription Term; or
• Entry into a new Order Form.
• Intercom may notify customers through their account, email or other means.
• Customers may be required to click to accept or otherwise agree to updated terms before renewal or a new order. Continued use after the effective date also constitutes acceptance.
• The prior express right to terminate the affected Subscription Term and receive a refund of prepaid fees for the unused portion appears to have been narrowed or removed from the operative language.

Risk: Intercom may have greater flexibility to impose revised legal, product or pricing terms, while the customer’s express termination/refund remedy is less clear.

4. Customer Data and AI-model training

• The diff does not clearly add or remove an express right to use Customer Data to train artificial-intelligence or machine-learning models.
• The existing data-use language remains materially focused on Intercom’s use of:
• Usage Data; and
• De-identified or anonymized Customer Data (“Anonymized Data”)

to analyze, improve, support and operate the Services during and after the agreement.

• The definition of “Laws” expressly references artificial intelligence, but this appears to be a clarification rather than a grant of training rights.
• The revised text does not expressly state whether Customer Data may be used to train general-purpose models, product-specific models, or models shared across customers.

Risk: “Analyze” and “improve” could potentially be interpreted broadly, but the diff alone does not establish permission to train AI models using identifiable Customer Data. Customers should review the DPA, Privacy Policy and product-specific AI terms for any separate training provisions, opt-out rights, retention limits or restrictions on human review.

5. Other notable changes

• Definitions and formatting have been standardized, including terms such as Customer Data, Customer Properties, Services, Affiliates, Third-Party Apps and Sensitive Personal Information.
• “Affiliate” appears focused on entities controlled by the Customer, which may affect who can use the Services or receive contractual benefits.
• The agreement retains broad disclaimers, liability limitations, indemnification provisions and incorporation of the SLA and Support Policy.

Summary

The provided diff does not include the actual added, deleted, or replaced legal text. It only states:

> “Added approximately 83 words to the document”

Accordingly, it is not possible to identify:

• The substantive contractual changes;
• Any new obligations, rights, limitations, or liabilities;
• Changes to confidentiality, ownership, security, or data-processing terms; or
• Whether customer data may be used to train, fine-tune, evaluate, or improve AI models.

AI-Training Review

No language addressing AI models, machine learning, model training, data retention, de-identification, or use of customer content is included in the diff provided. Therefore, no conclusion can be reached about whether the contract:

• Permits or prohibits training on customer data;
• Requires customer consent for such use;
• Limits training to aggregated or anonymized data;
• Allows human review or data sharing with model providers;
• Grants the provider ownership or broad usage rights; or
• Provides an opt-out, deletion right, or confidentiality protection.

Information Needed

Please provide the actual diff text, including the approximately 83 added words and any surrounding provisions. The analysis should show additions in {braces}, deletions in [brackets], and replacements using the stated format.

Summary of Important Changes

1. Agreement structure and incorporated policies

• The Terms of Service have been substantially reorganized and now expressly incorporate a broader set of documents, including:
• Additional Product Terms
• Privacy Policy
• Data Processing Addendum
• Regional Data Hosting Addendum
• Support Policy
• Cookie Policy
• Subprocessors List
• Acceptable Use Policy
• Service Level Agreement
• Security and other policy documents
• The revised structure may make it harder to identify which policies apply to a particular service. Customers should review all incorporated documents, including documents available only through linked webpages, because they form part of the contractual framework.

2. Expanded and clarified customer acceptance

• The agreement now more clearly states that accessing or using the Services, or indicating acceptance through an order process, binds the Customer to the agreement and referenced terms.
• Users acting for a company expressly represent that they are authorized to accept the agreement on the company’s behalf.
• The language also emphasizes that the agreement contains mandatory arbitration provisions and waives jury-trial rights. This is a significant procedural limitation on dispute resolution.

3. Effective date and contracting entity

• The “Effective Date” is clarified as the earlier of:

1. The Customer’s initial access to the Services; or

2. The effective date of the first Order Form referencing the agreement.

• The Intercom contracting entity is tied to the Customer’s shipping address:
• Intercom, Inc. for customers in the United States or Brazil; and
• Intercom R&D Unlimited Company for customers elsewhere.
• This may affect governing law, tax treatment, contracting-party liability, and available legal remedies.

4. Changes to the agreement

• The revised terms retain a unilateral modification right, but provide a more specific process:
• Changes generally become effective upon renewal of the current Subscription Term or entry into a new Order Form.
• Intercom may notify customers through account communications, email, or other means.
• Customers may be required to click to accept or otherwise agree to updated terms.
• Continued use after the effective date constitutes acceptance.
• If a Customer objects, its apparent remedy is termination of the applicable Subscription Term and a refund of prepaid fees for the unused terminated portion. The revised wording appears less focused on allowing termination specifically for pricing-related changes and may give Intercom greater flexibility to modify terms.

5. Customer Data and AI-model training

• The shown changes do not add an express authorization to use Customer Data to train artificial-intelligence or machine-learning models.
• The agreement continues to distinguish Customer Data from Usage Data and Anonymized Data. Intercom may use Usage Data and/or de-identified or anonymized Customer Data to analyze, improve, support, and operate the Services during and after the agreement.
• Although “training AI models” is not expressly stated, the broad “analyze” and “improve” language could potentially support development or optimization of automated systems, depending on the DPA, Privacy Policy, and Product Privacy Notice.
• Customers should confirm whether those incorporated documents contain separate AI-training provisions, opt-out rights, retention rules, or restrictions on using data from AI-enabled features.

6. Other observations

• Most apparent changes to defined terms are formatting, punctuation, or editorial corrections.
• No clear change is shown to Customer ownership of Customer Data or to the general confidentiality framework.

Summary

The provided diff states only that approximately 83 words were added; it does not include the actual added language or identify where those words appear.

Legal and Commercial Impact

Because the text of the additions is unavailable, it is not possible to determine whether the changes:

• Expand or restrict either party’s rights or obligations;
• Change liability, indemnity, confidentiality, security, or termination provisions;
• Introduce new fees, warranties, disclaimers, or compliance requirements;
• Permit broader use, disclosure, retention, or transfer of customer data; or
• Alter governing law, dispute resolution, or other material contractual terms.

AI Training and Customer Data

No conclusion can be reached about whether the contract now permits or restricts using customer data to train, fine-tune, evaluate, or improve AI or machine-learning models.

The added language should be reviewed specifically for terms such as:

• “Train,” “training,” “fine-tune,” “improve,” “develop,” or “enhance” models;
• “Inputs,” “outputs,” prompts, usage data, telemetry, or customer content;
• Rights to use data in aggregated, de-identified, anonymized, or pseudonymized form;
• Whether use is limited to providing the services or extends to product development;
• Whether customer consent is required or may be withdrawn;
• Data retention, deletion, and audit rights;
• Restrictions on using confidential, personal, regulated, or sensitive data; and
• Whether the provider may permit affiliates, subcontractors, or third-party model providers to access or use the data.

Required Information

Please provide the actual 83-word addition, including any surrounding text and markup showing replacements or deletions. Without the substantive diff, a reliable legal-risk analysis is not possible.

Executive Summary

The diff appears to be a substantial restructuring and modernization of Intercom’s Terms of Service, but much of the visible change is reordering, renumbering, punctuation, terminology, and consolidation of linked policies. The most important substantive changes concern contract formation, amendment procedures, entity identification, and dispute resolution. No express new permission to train AI models appears in the provided diff.

Important Changes and Risks

1. Broader and clearer contract formation

• The agreement now applies to an entity or person “placing an order for, or accessing” Services.
• Use of the Services, including accessing or using them on behalf of a company, constitutes acceptance.
• Users acting for a company expressly represent that they are authorized to bind it.
• The prior language describing the Effective Date as the earlier of initial access or an order process is retained in a reorganized form.

Risk: Employees or contractors may bind the customer through use of the Services, even if they did not complete a formal signature process.

2. Changes to amendment rights

The revised language states that Intercom may modify the Agreement from time to time. Unless otherwise specified:

• Changes generally become effective at renewal or when the customer enters into a new Order Form.
• Intercom may provide notice through the customer account, email, or other means.
• The customer may be required to click to accept or otherwise agree to the updated Agreement.
• Continued use after the updated version takes effect constitutes acceptance.
• If the customer objects, its remedy is generally limited to terminating the applicable Subscription Term and receiving a refund of prepaid fees for the unused terminated period.

Risk: Intercom retains significant unilateral amendment flexibility. The revised wording may also make it easier for changes to take effect through continued use or renewal. Customers should monitor notices and renewal dates.

3. Intercom contracting entity

The definition of “Intercom” is reorganized to identify:

• Intercom, Inc. for customers whose shipping address is in the United States or Brazil; and
• Intercom R&D Unlimited Company for customers elsewhere.

Risk: The applicable contracting entity, governing legal relationship, invoicing, and potentially dispute or enforcement considerations may depend on the shipping address in the applicable Order Form.

4. Mandatory arbitration

The revised acceptance language prominently states that the Agreement contains mandatory arbitration provisions and that disputes will be resolved by arbitration rather than jury trial.

Risk: Customers and users may lose access to court proceedings and jury trials, subject to the detailed arbitration terms elsewhere in the Agreement.

Customer Data and AI Training

• The definition of Customer Data appears unchanged in substance.
• The existing permission remains for Intercom to use Usage Data and/or Customer Data that has been de-identified or anonymized to analyze, improve, support, and operate the Services during and after the Agreement.
• The diff does not expressly add the words “train AI models,” “machine learning,” or equivalent permission.
• However, “analyze” and “improve” are potentially broad and could encompass model development if the data is genuinely anonymized or de-identified.

Risk: The text does not clearly state whether customer data may be used to train generalized or customer-specific AI models. Customers requiring a strict no-training commitment should seek an express contractual restriction and clarify whether de-identified data may be used for AI development.

Summary

The provided diff does not include the actual contractual language that was added, deleted, or replaced. It only states:

> “Added approximately 83 words to the document”

Accordingly, the legal and commercial impact cannot be assessed from the information provided.

AI Training and Customer Data

No text is 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 customer data is shared with affiliates, vendors, or third parties for model training
• Whether data is anonymized, aggregated, or de-identified before use
• Whether the customer can opt out of AI training or revoke consent
• Whether the provider owns or receives rights in customer data or model outputs
• Security, confidentiality, deletion, or retention obligations relating to AI systems

Therefore, no conclusion can be drawn about whether the amendment creates or expands rights to use customer data for AI training.

Key Limitation

The actual added, deleted, and replacement language is necessary to identify:

• New customer obligations or provider rights
• Changes to confidentiality, privacy, intellectual-property, or data-use terms
• Expanded purposes for processing customer data
• New liability exclusions, indemnities, or limitations of liability
• Changes to termination, deletion, audit, or compliance rights

Please provide the full redline or the 83 added words, together with any surrounding provisions needed to interpret them.

Executive Summary

The diff appears to replace and substantially reorganize the opening portions of the Terms of Service, while leaving many operational provisions substantively unchanged. The most important customer risks concern contract formation, incorporation of policies, unilateral amendments, dispute resolution, and the absence of any clear restriction on AI-model training.

Important Changes and Risks

1. Broader and more consolidated contract structure

• The revised terms reorganize the agreement into a single framework covering:
• Service definitions and customer obligations
• Security
• Third-party platforms and apps
• Ownership
• Fees and payment
• Termination
• Availability and support
• Liability, indemnification, confidentiality, publicity, and general terms
• Numerous policies and documents remain incorporated by reference, including the DPA, Additional Product Terms, Privacy Policy, Support Policy, Subprocessors List, Acceptable Use Policy, and others.

Risk: Customers may be bound by external web-based policies that can change independently or are difficult to review as a complete contract. The hierarchy among the various incorporated documents is not clear from this diff.

2. Contract formation is broader

The revised language states that the agreement is accepted by:

• Clicking acceptance;
• Signing or accepting an Order Form; or
• Accessing or using any Services.

It also expressly binds a person using the Services on behalf of a company and requires that person to represent they have authority to accept for the company.

Risk: Unauthorized or informal use by an employee, contractor, or other person could create contractual obligations for the customer company.

3. Effective date and contracting entity clarified

The Effective Date is the earlier of:

• Initial access through online provisioning, registration, or ordering; or
• The effective date of the first Order Form referencing the agreement.

The applicable Intercom entity is determined by the customer’s shipping address, including Intercom, Inc. for the United States/Brazil and Intercom R&D Unlimited Company for other locations.

Risk: Obligations may begin before a formal Order Form is signed. The applicable legal entity may affect governing law, enforcement, tax treatment, and regulatory rights.

4. Amendments remain unilateral

The revised clause preserves Intercom’s ability to modify the agreement. Ordinarily, changes become effective on renewal or entry into a new Order Form. Continued use after the updated version takes effect constitutes acceptance. Intercom may also make changes effective before renewal, including for legal, product, or pricing reasons.

The customer’s remedy for objecting is termination of the applicable Subscription Term and a refund of prepaid fees for the unused terminated period.

Risk: Intercom retains significant flexibility to change contractual terms, potentially including commercially important provisions. Continued use may waive objections.

5. Mandatory arbitration emphasized

The revised acceptance notice expressly states that the agreement contains mandatory arbitration provisions requiring disputes to be resolved by arbitration rather than jury trial.

Risk: Customers may lose access to court proceedings and jury trials, subject to the detailed arbitration section and applicable law.

AI Training and Customer Data

• The diff does not show a new express authorization for Intercom to use Customer Data to train AI models.
• The existing language permits use of Usage Data and de-identified or anonymized Customer Data to analyze, improve, support, and operate the Services during and after the agreement.
• “Laws” expressly includes laws concerning artificial intelligence.
• No clear prohibition or affirmative permission regarding training generative AI or machine-learning models appears in the supplied diff.

Key concern: The phrase “analyze, improve… and operate” could potentially be argued to cover model development if not limited elsewhere. Customers should review the DPA, Privacy Policy, Product Terms, and any AI-specific terms for a direct training restriction, opt-out, retention rule, or statement that Customer Data is excluded from model training.

Summary

The provided diff only states:

> “Added approximately 83 words to the document”

It does not include the actual added, deleted, or replaced legal language. As a result, the substantive changes and associated risks cannot be determined.

AI Training and Customer Data

• No language addressing customer data, use of data, artificial-intelligence model training, machine learning, model improvement, or related rights is included in the diff.
• It is therefore not possible to determine whether the revised terms:
• Permit the provider to use customer data to train or improve AI models;
• Limit such use to de-identified, aggregated, or anonymized data;
• Require customer consent or provide an opt-out;
• Give the provider rights to retain prompts, outputs, inputs, or usage data;
• Allow disclosure of customer data to third-party model providers; or
• Allocate responsibility for confidentiality, security, intellectual property, or regulatory compliance.

Risk Assessment

No specific legal risks can be identified without the text of the 83 added words. The change could be immaterial, or it could materially expand the provider’s rights to use customer content and associated metadata.

Information Needed

Please provide the actual redlined language, using the stated notation:

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

The surrounding original language may also be necessary to assess how the amendment changes the agreement.

Key Changes and Risks

1. Customer Data may now be used for AI model training

• The data-use provision expressly changes Intercom’s permitted use of Customer Data from providing and improving the Services to providing and improving the Services, “including for Intercom AI model training.”
• This is a material expansion and removes ambiguity about whether Customer Data may be used to train Intercom’s AI models.
• The provision does not appear to specify:
• Whether training applies by default or requires opt-in;
• Which AI models or subprocessors may receive the data;
• Whether Customer Data may be used to train models serving other customers;
• Whether trained models or derived outputs can be deleted or separated from Customer Data;
• Any exclusion for sensitive, confidential, regulated, or proprietary information; or
• Customer’s ability to opt out or impose data-use restrictions.

Risk: Customer content, including potentially personal, confidential, or commercially sensitive information, may contribute to Intercom’s AI-model development. The agreement should be reviewed alongside the DPA, privacy policy, AI terms, and any order-form commitments to determine whether an opt-out or stricter limitation exists.

2. Anonymization/de-identification language is broadened

• Usage Data and/or Customer Data may be used for analytics and related purposes if it is “de-identified or anonymized” to remove personal data.
• Intercom is now expressly permitted to “de-identify or” anonymize Customer Data for those purposes.

Risk: “De-identified” and “anonymized” are not necessarily equivalent. De-identified information may still carry re-identification risk, particularly when combined with other data. The clause also does not define the required technical standard, prohibit re-identification, or clearly limit use of de-identified data to aggregate statistics.

3. New API-key and credential restrictions

• Customers may not share or permit others to use Customer-assigned API keys or passwords, including in violation of the Intercom Developer Terms.
• Customers must require Permitted Users to keep credentials confidential.

Risk: The Customer may bear responsibility for credential sharing or misuse by employees, contractors, or third parties. The incorporated Developer Terms should be reviewed because they may impose additional obligations or change separately.

4. Usage-based pricing is substantially restructured

• The agreement introduces “service-specific credits,” which apply only to the applicable Service.
• Included credits reset monthly and generally do not carry forward.
• After credits are exhausted, additional usage may be charged on a pay-as-you-go basis, using pricing in the applicable Order Form.
• Usage quantities and credit consumption may vary according to Documentation.
• Contracted-usage changes are prorated from the effective date; overages are billed in arrears.

Risk: Costs may be less predictable, particularly where credit consumption varies by activity and pricing or calculation rules are located in Documentation rather than the agreement.

5. Other notable changes

• Customer access to Customer Data ends 30 days after termination.
• Liability exclusions now expressly preserve liability that cannot legally be limited, including certain gross negligence, fraud, or intentional misconduct.
• Tax language is reorganized and may broaden Customer’s responsibility for taxes and legally required charges.

Summary

The supplied diff does not include the actual contractual language that was added, deleted, or replaced. It only states:

> “Added approximately 83 words to the document”

Accordingly, no legally meaningful comparison can be performed.

AI Training and Customer Data

• The diff does not identify whether customer data may be used to train, fine-tune, evaluate, or improve AI models.
• It does not reveal whether any such use is:
• Permitted without the customer’s consent;
• Limited to de-identified or aggregated data;
• Conducted by the provider or third-party service providers;
• Optional or subject to an opt-out;
• Limited after contract termination; or
• Subject to confidentiality, security, or data-retention restrictions.

Potential Risks

Because the 83 added words are not provided, it is not possible to assess whether the changes:

• Expand the provider’s rights to use customer data;
• Permit AI training on identifiable, confidential, or personal information;
• Allow disclosure of data to affiliates or third-party AI vendors;
• Reduce restrictions on data retention or deletion;
• Change ownership or licensing rights in customer data or generated outputs; or
• Shift liability for unauthorized AI use, privacy violations, or security incidents.

Information Needed

Please provide the actual redlined text, using the stated notation—{additions}, [deletions], and []{replacements}—to permit a substantive legal analysis.

Summary of Important Changes and Risks

1. Agreement structure and incorporated policies

• The revised text reorganizes the agreement and expressly incorporates a much broader set of documents, including:
• Additional Product Terms
• Privacy Policy
• Data Processing Addendum
• Regional Data Hosting Addendum
• Security terms
• Third-Party Platform and App terms
• Ownership, payment, termination, liability, indemnification, confidentiality, publicity, and general terms
• Acceptable Use Policy, Code of Conduct, Subprocessor List, SLA, and other policies
• Risk: Important obligations may exist outside the main agreement and may be changed separately. Customers should review all incorporated documents and confirm which version controls if documents conflict.

2. Broader acceptance mechanics

• The agreement now applies to an entity or person ordering or accessing Services.
• Use by an employee or representative constitutes a representation that the person is authorized to bind the company.
• Acceptance may occur by clicking, ordering, accessing, or using the Services.
• The revised language prominently states that the agreement is legally binding and contains mandatory arbitration provisions requiring disputes to be resolved by arbitration rather than jury trial.
• Risk: Users may bind their organization without a separately signed contract. Unauthorized users could create disputes regarding authority and enforceability.

3. Intercom contracting entity

• The agreement clarifies that the contracting entity will be:
• Intercom, Inc. for customers whose shipping address is in the United States; or
• Intercom R&D Unlimited Company for customers elsewhere.
• Risk: The applicable entity affects governing law, tax treatment, venue, regulatory obligations, and potentially enforcement. The use of shipping address as the determining factor should be verified.

4. Unilateral amendments

• Intercom may modify the agreement. Ordinarily, changes take effect at renewal or when the customer enters a new Order Form.
• Intercom may require the customer to click acceptance before renewal or a new Order Form. Continued use after the updated version takes effect constitutes acceptance.
• For certain changes, including legal, product, or pricing reasons, the customer’s remedy is termination of the affected Subscription Term and a refund of prepaid fees for the unused period.
• Risk: Continued use can create acceptance by conduct. The termination-and-refund remedy may be the customer’s exclusive remedy for objectionable amendments.

5. Customer data and AI-model training

• The data-use language states that Intercom may use technical data about customer and People usage, and/or Customer Data anonymized to remove personal data, to analyze, improve, support, and operate the Services during and after the agreement.
• No express change or authorization to train AI models is visible in this diff. The terms “train,” “training,” or “machine learning” do not appear in the supplied text.
• Nevertheless, “analyze” and “improve” are broad and could potentially support machine-learning development, particularly if data is treated as anonymized.
• Risk: The wording is unclear about whether all Customer Data is anonymized before use, what anonymization standard applies, whether anonymized data can be retained indefinitely, and whether it may be used to train generalized models serving other customers. Review the DPA, Privacy Policy, and AI-specific product terms for clarification.

Summary

Scope of the Change

• The diff states only that approximately 21 words were removed from the document.
• The actual deleted language is not provided.

Legal and Commercial Impact

• It is not possible to determine which obligations, rights, limitations, or protections were changed without seeing the specific removed text.
• The removal could affect provisions such as:
• Confidentiality and data-protection obligations
• Customer data ownership and permitted uses
• Service provider access to customer data
• Liability, indemnification, or warranties
• Termination and data-deletion requirements
• Audit, security, or regulatory compliance rights

AI Training and Model Use

• The provided diff does not identify whether any language concerning AI, machine learning, model training, data analysis, or use of customer data was added or removed.
• Accordingly, no conclusion can be reached about whether:
• Customer data may be used to train or improve AI models;
• Customer data may be aggregated, anonymized, or de-identified for training;
• customer prompts, inputs, outputs, or usage metadata may be retained;
• the provider must obtain consent before using customer data for AI training; or
• the provider must exclude customer data from model training.

Risk Assessment

• The principal risk is lack of visibility: removal of even a short phrase could materially broaden the provider’s rights or weaken customer protections.
• The deleted language should be reviewed specifically for terms such as “train,” “improve,” “develop,” “machine learning,” “artificial intelligence,” “aggregated,” “de-identified,” “service data,” “customer data,” and “model.”

Recommended Next Step

• Provide the actual 21 deleted words, or a redline showing their location in the agreement. Without that text, a reliable legal or AI-data-use analysis is not possible.