Skip to main content

Reservly AI Use Policy

Last updated: April 26, 2026

The short version

  • We use AI where it makes the product better. Currently: software development (assisting our engineers when writing code) and — on the roadmap — translating business-facing legal templates and notification copy into additional languages.
  • We do not train models on your customer booking data. Not our models, not anyone else's. That line is hard.
  • We do not make automated decisions about you or your customers that have legal or similarly significant effects.
  • If you integrate an AI receptionist or chatbot with Reservly's API (Vapi, Retell, Voiceflow, ChatGPT, Claude, and similar), that integration is yours — this policy describes what Reservly does, not what our customers' AI tools do.
  • If we ever start using AI in a materially new way that affects you, we'll update this policy and let you know before the change takes effect.

Questions: support@reservly.io.

1. Scope and purpose

This page discloses where and how Reservly uses artificial-intelligence tools in operating the platform, and where it doesn't. It exists because regulators (the US Federal Trade Commission, the EU under the AI Act, state-level AI-disclosure rules in California and elsewhere, and Canada's AIDA) increasingly expect platforms to be transparent about AI use even where there is no specific required disclosure — and because you deserve to know anyway.

Three things to keep clear:

  • Reservly's own AI use — what we do in operating the Service. Covered in §§ 3-5.
  • Your business's AI use — when you, as a Business, integrate an AI assistant or automation with Reservly's API or webhooks. Governed by your agreement with the AI provider, not this policy. Covered briefly in § 6 for context.
  • Third-party services Reservly integrates with — Google Meet, Zoom, Microsoft Teams, Stripe, PayPal, Paddle, Sentry, and so on — each have their own AI policies. We don't restate those here.

2. Our principles

  • Transparency over sophistication. We'd rather disclose a boring AI use than hide a clever one.
  • No training on customer data. We do not use customer booking data, end-user personal data, or custom-form responses to train any AI model, ours or anyone else's.
  • Humans own consequential decisions. Account enforcement, refund decisions, privacy-request responses, accessibility triage, and similar customer-affecting decisions are reviewed by a human before being executed. AI may help us draft or analyse; it does not decide.
  • Minimise data exposure. When we use AI tooling internally, we scope the data surface to the minimum necessary and prefer tools that process data in jurisdictions we can verify.
  • Honest failure. AI outputs are sometimes wrong. When we rely on an AI tool in operating the Service, we check the output or label it as AI-generated where that matters.

3. Where Reservly uses AI today

3.1 Software development

Our engineering team uses AI coding assistants (including large language models such as those provided by Anthropic and OpenAI, and IDE-integrated tools such as GitHub Copilot or Cursor) when writing, reviewing, and debugging the software that runs Reservly. These tools process code, inline documentation, and developer-authored prompts.

What this means for your data: Engineers do not paste customer booking data, end-user personal data, or production secrets into AI tooling. The code we write is reviewed before it reaches production through our normal change-management process (pull request review, CI, type checks, tests, pre-launch verification).

3.2 Support assistance

When responding to support tickets, the Reservly team may use AI tools to help triage, summarise, or draft responses. Any response sent to a customer is reviewed by a human before delivery. We do not paste sensitive customer data (payment identifiers, OAuth tokens, security-related details) into third-party AI tooling; we handle those through internal systems.

3.3 Error-report triage

Sentry error reports arriving from production may be summarised by AI tooling for internal triage. The Sentry pipeline scrubs personally identifiable information before ingestion. The AI tool sees scrubbed error traces, not raw data.

4. Where Reservly plans to use AI

These are on the roadmap. We will update this policy to move them from "planned" to "current use" when they ship.

4.1 Translating business-level legal and communication templates

Reservly provides default templates that Businesses can customise (privacy policy, cookie policy, terms of service, cancellation, no-show, accessibility — see Settings → Legal). Businesses operating in non-English-speaking markets need these templates translated into their customers' language. We plan to use AI translation to provide additional language coverage beyond our currently-supported languages.

AI provider for this feature: We plan to use Mistral AI (Mistral AI SARL, Paris, France — EU-based provider). Mistral AI processes translation requests on infrastructure in the EU (Frankfurt). Under Mistral AI's paid API terms, customer inputs are excluded from model training, and data is retained for no more than 30 days. A Data Processing Addendum is available at legal.mistral.ai/terms/data-processing-addendum.

Safeguards that will apply when this feature ships:

  • Legal-translation output will be reviewed by a bilingual human before being offered as a default.
  • Machine translations will be labelled as machine-translated in the Business's dashboard, with a prompt to have a local professional review before publishing.
  • Translation outputs will include a machine-readable marker (x-machine-translated: true) in the template metadata so downstream integrations can identify translated content.
  • The original English master remains authoritative in any dispute; the Business is responsible for deciding whether to adopt a machine translation.
  • We do not send personally identifiable customer data (names, booking details, contact information) to the translation provider. Only template text is translated.

4.2 Admin help assistant (if we build it)

We have scoped a possible admin help bot that would answer operational questions inside the Reservly dashboard (for example, "how do I add a second staff member to my team?"). If we ship it:

  • It will be clearly labelled as AI.
  • It will not have the ability to change settings on your behalf — it will only provide information.
  • Conversations will not be used to train the underlying model.
  • We will disclose which provider backs the bot.
  • EU AI Act Article 50 compliance: Before any customer-facing deployment of an AI conversational interface, Reservly will display a non-dismissable disclosure at the start of every session — visible before the first AI-generated message — informing the user they are interacting with an AI system. This obligation applies from 2 August 2026 (the date EU AI Act Article 50 transparency requirements become enforceable). Reservly will not deploy a customer-facing AI assistant without this disclosure in place.

4.3 Communication improvements

We may use AI internally to analyse aggregate support-ticket patterns to improve Reservly's documentation, in-product help text, and onboarding flow. Such analysis operates on aggregated, de-identified ticket content; no AI sees your customer booking data.

4.4 Customer-facing content translation (planned — activates when R4.2 ships)

A future release (R4.2 on our roadmap) will allow Businesses to offer their booking pages, service descriptions, and notification copy in languages beyond the 66 already supported through manual translation. When this ships, Reservly may use AI translation to generate the initial translated copy for Business review.

This section will be moved from § 4 to § 3 when the feature launches. Until then, this is a forward-looking disclosure only. The same Mistral AI provider and safeguards described in § 4.1 will apply.

5. Where Reservly does not use AI

To set expectations explicitly:

  • We do not train AI models on your data. We do not train our own models, and we do not contribute to any third party's training corpus with your booking data, end-user personal data, staff information, or custom form responses.
  • We do not use AI to make automated decisions about you or your end users that produce legal or similarly significant effects (in the sense of GDPR Article 22 and equivalent laws). Subscription cancellation, refund processing, account suspension, and similar customer-affecting actions are reviewed by a human.
  • We do not use AI-powered facial recognition, voice biometrics, or similar biometric-analysis features. If we ever enable a feature that touches biometrics, we will ship a full policy update and explicit opt-in.
  • We do not generate AI "deepfake" content from customer data.
  • We do not use end-user booking data to target advertising. End-user booking data on Reservly-hosted booking pages is never shared with advertising platforms, even after Reservly activates paid advertising campaigns on its own marketing site. Ad pixels (when active) live on reservly.io marketing pages only. See Cookie Policy and Do Not Sell or Share.

6. When you connect an AI tool to Reservly

Reservly's API and MCP server are explicitly designed to support AI integrations, including voice receptionists (Vapi, Retell AI, Voiceflow, Synthflow, Bland AI), chat assistants (ChatGPT Custom GPTs, Claude Projects, custom LLM agents), and automation platforms (Zapier, Make, n8n). We welcome these integrations — see the API Terms of Service.

When you connect one of these tools:

  • The tool is yours, not ours. You chose it, you configured it, you're the customer of its provider.
  • Data flows outside Reservly. Booking data, customer messages, call transcripts, and related content leave Reservly and enter the AI tool's environment when you use one. We cannot control what happens there.
  • Disclose to your end users. Local law (including EU AI Act transparency rules for certain AI systems, CCPA/CPRA disclosures for automated decision-making, and various state laws) may require you to disclose to your end users that they are interacting with an AI system when they talk to your AI receptionist or chatbot. The default business-level templates include placeholder language; consult your own counsel for the specifics.
  • Reservly's responsibility ends at our API. We warrant the behaviour of the Reservly API as published. We do not warrant the accuracy, reliability, bias, or outputs of third-party AI tools you connect.

7. Your rights regarding AI use

You have the same rights under this policy as under the Privacy Policy § 9. Specifically:

  • Right to know. Ask us how we used AI in connection with your account and we will explain to the extent we can.
  • Right to human review. If a consequential decision about your account has been made with the aid of AI tooling, you may request a human review by emailing support@reservly.io with subject line "Human review".
  • Right to object to automated processing (where applicable under GDPR Article 22, LGPD, or similar laws) — see Privacy Policy § 9.2-9.3.

We will not retaliate for exercising these rights.

8. AI and accessibility

Accessibility at Reservly is developed and tested by humans, not AI. We use AI coding assistants as part of engineering work (§ 3.1), which incidentally helps us implement ARIA patterns, focus management, and other accessibility affordances — but the accessibility requirements, success criteria, and user-journey testing come from our internal accessibility posture (see Accessibility Statement). We do not outsource accessibility decisions to AI.

9. AI and security

Reservly uses AI tooling for some internal security operations — anomaly detection in error traces, log-pattern analysis, and similar. These tools process scrubbed operational data, not raw customer data. Security researchers reporting vulnerabilities should use the Responsible Disclosure channel; our response is human-led, not AI-triaged (though an AI tool may help us prepare the response, it is reviewed by a human before sending).

10. Changes to this policy

We will update this policy when:

  • We move a planned AI use from § 4 to § 3 (i.e., ship it).
  • We start using AI in a materially new way that affects customers.
  • A material change in our AI-provider stack occurs (for example, a significant shift in which models power which internal workflow).
  • Applicable law evolves (the EU AI Act, state-level AI statutes, FTC enforcement guidance).
  • EU AI Act Article 50 enforcement date (2 August 2026): On this date, the EU AI Act's Article 50 transparency obligations for AI systems interacting with natural persons become enforceable. This policy will be updated before that date to reflect any customer-facing AI deployments and the non-dismissable AI-disclosure requirement. If Reservly has not deployed any customer-facing AI interface by that date, this policy will be updated to confirm that status.

We will:

  • Update the Last updated date at the top.
  • Email active Business subscribers at least 30 days before material changes take effect.
  • Keep prior versions on request.

11. Contact

Email: support@reservly.io Subject line for AI-related questions: AI Use Subject line for human review of an AI-assisted decision: Human review Subject line for right-to-know about AI use: AI right to know

Reservly c/o Northwestern Registered Agent Services 30 N Gould St Ste R Sheridan, WY 82801 United States

This AI Use Policy works together with our Privacy Policy, Terms of Service, API Terms of Service, Sub-Processor List, and Accessibility Statement.