Heyfixit Agent and Communications Terms
These terms form part of the agreement between Heyfixit, Inc. ("Heyfixit", "we", "us") and you under the Heyfixit Terms of Service. They cover what is specific to AI agents that call, text, message and email people on your behalf, and to Vera, our platform agent. Words defined in the Terms of Service have the same meaning here.
1. How the Agents work
1.1 Your instructions. The Agents act within your instructions and the permissions you give them, and within the limits we set for safety and quality. What an Agent says or does within them is done on your behalf and is treated as authorised by you. Where these terms describe what the Agents do, they describe our standard instructions and your instructions. If an Agent acts otherwise, that is an Agent action, unless a property in clause 1.1A fails. Our standard instructions are offered to you and never applied without your adoption. Adopting or adjusting them is your choice, and they are provided as they are. Clause 4.7 of the Terms of Service applies.
1.1A What the platform does whatever your instructions.
(a) Your Customer Data is kept separate from every other customer's.
(b) The Agents act only within the permissions you give them, and the Services record each action and each change to your instructions and permissions, so we can show who did what, for whom and under which instructions for as long as we hold those records.
(c) Each message the Agents send is sent once, even if a step is retried.
(d) A recorded request to stop contact is checked before every message or call the Agents send.
1.2 Outputs.
(a) The Agents use AI models. Their Outputs can be wrong or incomplete, and the same request can produce different results.
(b) We may change the AI models, voice and speech providers and other providers behind the Agents at any time, including switching between them, to improve quality and performance. Outputs may change as a result.
(c) You decide where a person should check an Agent's Output or action before you rely on it.
(d) You are responsible for Outputs and Agent actions, as clause 12.6 of the Terms of Service describes.
1.3 Decisions that need a person.
(a) You must not rely on an Agent alone for a decision of a kind listed in (b). A person on your side with the training or experience to judge the matter must review it, with the power to change the outcome, and must make or approve the decision.
(b) The kinds of decision are:
- decisions about a tenancy or eligibility for housing;
- enforcement or legal action, including applying to court for access;
- refusing or withdrawing an essential service;
- decisions about an individual's employment, pay or engagement as a worker.
(c) The Services let you require a person's approval before an Agent takes these steps. Where this clause applies, your permissions must require it.
1.4 No professional advice. The Agents are not designed to give legal, financial or other professional advice, or safety advice beyond the instructions you give them, and you must not configure them to. Any safety information they pass on comes from your instructions, and you are responsible for it. They are not a substitute for qualified people.
1.5 Autonomy you set. Through your permissions, you decide how much the Agents may decide and do on their own, including the approval levels they apply to work orders, quotes, purchase orders, risk assessments, method statements, permits and sign-offs. When an Agent makes or approves something within the permissions you give it, the decision is yours, made on your behalf. Your instructions and permissions must still follow clause 1.3 and the Acceptable Use Policy, which apply whatever autonomy you set.
1.6 Health and safety. You remain responsible for health and safety in your work and on every site, including assessing and controlling risks, the competence of the people who do the work and every duty the law gives you as an employer, contractor or duty holder. Giving an Agent permission to make or approve decisions that affect safety does not pass any of these duties or risks to us. We do not assess, manage or supervise risks on any site.
1.7 Vera.
(a) Vera acts only within the permissions of the signed-in User: the records and contracts that User can access. This is part of clause 1.1A(b).
(b) Vera makes a change, sends a message outside your organisation or instructs an Agent after that User confirms it, unless your instructions allow Vera to act first. Our standard instructions have Vera ask first. A change, message or instruction a User confirms is treated as made by that User.
(c) Vera's answers can be wrong or out of date. Users should check facts before relying on them, and the Services say so where Vera's answers appear.
(d) Clauses 3.1 and 3.3 apply to Agents that communicate with Contacts and not to Vera.
1.8 EU AI Act.
(a) Where the EU AI Act (Regulation (EU) 2024/1689) applies, we are the provider of the Agents and Vera, and you are their deployer.
(b) You meet the duties the Act gives a deployer, including telling people when content you publish is AI-generated where the Act requires it, and any duty on AI literacy that applies to you.
(c) You must not use the Agents for a purpose the Act treats as prohibited or high-risk, except as clause 2(v) of the Acceptable Use Policy allows.
1.9 Not an emergency service. The Services are not an emergency service. In an emergency, people should call the local emergency number.
2. You are the sender
2.1 Who is contacted and why. You choose the people the Agents contact, the purpose of each contact and when it may happen, within the limits in clause 4.3. Calls, texts, WhatsApp messages and emails sent by the Agents are sent on your behalf. As between you and us, you are the sender of them.
2.2 Lawful basis and consent. For every Contact, you are responsible for:
(a) having a lawful basis for the contact;
(b) holding any consent the law requires;
(c) before an Agent sends a person a text or WhatsApp message, holding the consent the law and the messaging providers' rules require for that message, including written consent where those rules require it. A message the person sends first lets the Agent reply in that conversation, but not start further contact;
(d) for calls to US numbers, also holding the consents in clause 2.5, which we require for any call made with an AI voice, because US law treats AI voices as artificial voices;
(e) keeping a record of each consent and notice, showing when and how it was given, for at least five years or longer if the law requires;
(f) giving us a copy of a record within five working days of our request, so we can answer a carrier, provider or regulator.
2.3 Privacy notices.
(a) Your privacy notice, or your client's where your client decides how Contacts' information is used, must tell Contacts that:
- AI agents call, text, WhatsApp and email them on your behalf;
- calls are recorded and transcribed, as clause 6.1 describes;
- we and the providers listed at heyfixit.ai/legal/subprocessors process their information for you;
- carriers and messaging platforms also use some of it for their own purposes, such as preventing fraud and abuse;
- we may remove everything that identifies a person, including their voice, from the information we process for you, and use the result for any purpose, including to improve our services and build new ones, as the Data Processing Agreement describes.
(b) Your instructions must include a link to that privacy notice. If you do not have a suitable privacy notice, we can host one for you from our standard template, which you approve and which stays yours. The Agents give the link as clause 3.1(c) describes.
(c) You must keep the notice up to date and tell us if the link changes.
2.4 Data from your clients. Where you contact people for a client, or use contact details a client gave you:
(a) your contract with your client must allow the contact the Agents make;
(b) where you process the data on your client's behalf, your client must have authorised you to use us and our subprocessors;
(c) a consent your client holds counts under clause 2.2(c) only if the message names your client as the sender, or the consent also covers messages from you.
2.5 Calls to US numbers. US law treats AI voices as artificial voices. Before an Agent calls a US number, you must hold:
(a) the person's prior express consent to calls about that subject; and
(b) their prior express written consent, if any part of the call is marketing.
3. What the Agents tell people
3.1 Disclosures. The Agents are built to disclose as follows, and your instructions cannot remove these disclosures.
(a) At the start of each call, and in the first message of each conversation, the Agents say in plain words that they are AI agents and who they act for. On calls, they also say then that the call is recorded and transcribed, and they repeat that they are AI agents during a long call.
(b) They never claim or suggest that they are a person, and never deny being AI when asked. You must not instruct them otherwise.
(c) Where the channel or the law requires it, they also tell people how to stop contact and where to find your privacy notice.
(d) Your instructions must include any other details the law requires the Agents to state where the Contact is, such as a phone number for your business on calls to US numbers. You are responsible for those details.
3.2 Names and voices. You may choose the name and voice an Agent uses from those we offer. A voice that copies or sounds like a real person, or a name that suggests a specific real person, needs that person's explicit written consent and our written agreement. You must keep that consent and give us a copy if we ask.
3.3 Reaching a person.
(a) You must give Contacts a way to reach a person on your team, such as a phone number or email address, and include it in your instructions.
(b) The Agents give these details as your instructions say. Our standard instructions give them when a Contact asks for a person, or says they do not want to speak to an AI agent.
(c) If a Contact objects to the call being recorded or transcribed, the Agent gives your contact details and ends the call.
(d) You must make reasonable adjustments for Contacts who cannot use the Agents, for example because of a disability.
4. Calls, texts and messages
4.1 Opt-outs.
(a) The Services record each request to stop contact and check it before every message or call the Agents send, as clause 1.1A(d) describes.
(b) After an opt-out by text or WhatsApp, the Agent may send one message confirming it, and nothing more.
(c) What counts as a request to stop, what it covers and how a person opts back in follow your instructions. Our standard instructions treat a request to stop, whether by replying STOP or a similar word or by saying so on a call, in a message or in an email, as covering that channel and purpose, or all contact if the person asks, until the person opts back in, for example by replying START, or gives fresh consent. Under them, the Agents may still answer a person who contacts them first. Your instructions must at least meet the law and the messaging providers' rules where the person is.
(d) If you receive a request to stop contact in another way, you must record it in the Services within two working days.
(e) You must not ask us to contact a person who has opted out.
(f) An opt-out stops contact by the Agents only. It does not stop you contacting the person by other lawful means, such as a letter or a call from your team.
4.2 Marketing.
(a) The Services are built for operational contact, such as arranging visits and access. Operational calls and messages must not promote goods or services. A single promotional line can make the whole call or message marketing in law.
(b) You may use the Agents for marketing only if we agree in writing. You are then responsible for:
- holding each person's prior consent to marketing calls or messages from an AI agent. In the UK, checking the Telephone Preference Service is not enough for AI calls;
- in the US, holding prior express written consent and checking the national and state do-not-call registers;
- checking the Corporate Telephone Preference Service and any other register that applies;
- complying with telemarketing law.
4.3 Calling hours and frequency.
(a) The Agents call and send texts and WhatsApp messages at the times your instructions say. Our standard instructions set calling hours in the Contact's local time, which your instructions can change. The Agents keep to them unless the Contact asks for another time.
(b) Our standard instructions allow replies to people who contacted the Agents first, and contact your instructions treat as urgent, outside those hours. You are responsible for any hours the law sets.
(c) The Agents follow the limits on attempts your instructions set, within caps we set so that no one is contacted too often. If an Agent cannot speak on an answered call, the Services play a short recorded message saying who called and how to reach you.
4.4 Caller identity.
(a) Calls and messages come from numbers and sender names we provide, or that you ask us to use and are entitled to use. You must show us proof of that right if we ask.
(b) A person who calls back one of the numbers dedicated to you reaches the Agent or your team, as your instructions say, and can ask not to be contacted again.
(c) You must not ask us to display a number or sender name that misleads people about who is contacting them.
(d) You must give us accurate details about your business for registering numbers, sender names and WhatsApp accounts with carriers, Meta and regulators, and keep them up to date.
(e) You must help us answer requests from carriers, regulators and law enforcement about communications made on your behalf.
4.5 WhatsApp.
(a) The Agents send WhatsApp messages from a WhatsApp Business account in your name. Before the Agents use WhatsApp for you, you must accept Meta's WhatsApp business terms, including its data processing terms. Meta processes those messages for you under those terms.
(b) Before an Agent starts a WhatsApp conversation with a person, you must hold that person's opt-in to receive WhatsApp messages from the business the message names.
(c) Meta requires messages sent more than 24 hours after the person last wrote to use templates it has approved. You must not ask us to put promotional content in a template approved for service messages.
(d) If Meta limits or blocks an account, number or template because of how it was used on your behalf, clause 8.2 applies.
5. Phone numbers
5.1 Our numbers. Phone numbers we provide are held by us or our carriers. We let you use them for the term of your Order Form. You do not own them.
5.2 When numbers can change. We may change or withdraw a number if the law, a carrier, a numbering authority or a regulator requires it, for technical reasons, if the number is misused or if it has not been used for 90 days. We tell you as early as we can and provide a replacement where possible.
5.3 Moving numbers. When your Order Form ends, and if you have paid all Fees, we will help you move a number to your new provider where our carrier allows it. You pay any porting charges and must confirm the request in writing. Otherwise, numbers cannot be transferred unless we agree in writing.
5.4 Country rules. Some countries restrict how numbers can be used. Where such rules apply to the numbers you use, you must follow them, and you are responsible for doing so.
5.5 Earlier users of a number. If you receive a call or message meant for an earlier user of a number, you must not use it and must delete it.
6. Recordings and transcripts
6.1 What is recorded. Every call is recorded and transcribed. An Order Form may agree something different for you. The opening line says so, as clause 3.1(a) describes.
6.2 Customer Data. Recordings, transcripts and messages are Customer Data.
6.3 Objections. If a person objects to the call being recorded or transcribed, clause 3.3 applies.
6.4 Local law. You are responsible for any further notice or consent the law requires where Contacts are. If the law in a place requires a person to say yes before a call is recorded, you must tell us before the Agents call people there, so that the Agents ask at the start of each call, record only once the person says yes, and end the call if the person says no.
7. Carriers and networks
7.1 Our dependence on carriers. Calls and messages travel over telecoms networks we do not control. Carriers may filter, delay or block calls and messages, and may require changes to how numbers and messages are used.
7.2 What we are not responsible for. We are not responsible for carrier or network failures, and they do not count against our service levels.
8. Our providers' rules
8.1 Rules that also bind you. The Services rely on telecoms, messaging, voice and AI providers, and their usage rules also apply to how the Services are used, so far as they cover the parts of the Services you use. You must follow the provider rules described in clause 4 of the Acceptable Use Policy. Where a provider's rule is stricter than ours, the provider's rule applies.
8.2 Providers acting against us. If a provider suspends or limits us because of how the Services were used on your behalf, clause 7 of the Terms of Service applies.
9. Charges caused by your use
So far as the law allows, you pay any fine, penalty or charge that a carrier, regulator or other authority imposes because of communications made on your behalf. This includes charges for complaints, unlawful contact or misuse of numbers. It does not include a charge to the extent our breach of the Agreement caused it, as clause 5.7 of the Terms of Service also says.
10. Software provider only
10.1 What we are. We provide software. We are not a party to your contracts with your clients, Contacts, suppliers or contractors.
10.2 What we are not responsible for. We are not responsible for:
- the work carried out in any property;
- the condition or safety of any property;
- the conduct of anyone the Agents arrange a visit with.
10.3 Not insurance. The Services are not an insurance product or a maintenance guarantee.
11. Your own systems
Systems, spreadsheets, accounts and integrations that you provide or control are your responsibility. Problems with them do not count against our service levels.