Terms & Conditions

1. Who this is between, and how it is accepted

These Terms are between the business that opens an account (“you”, “the Client”) and Softpact (“we”, “us”), at Carrer de Calàbria, 64, Eixample, 08015 Barcelona, Spain.

You accept them by creating an account, by continuing to use the service, or by confirming go-live for an assistant — whichever happens first. If you are accepting on behalf of a company, you confirm you are authorised to bind it.

This is a business-to-business service. It is not offered to consumers, and if you are contracting as an individual outside a trade or profession, you should not accept these Terms.

2. Definitions

• Assistant — the AI agent configured for your business, on the channels you connect.
• End Customer — a person who messages your Assistant. Your customer, not ours.
• Knowledge Base — the business information you provide for the Assistant to answer from.
• Sandbox — the environment where you test the Assistant without it reaching End Customers.
• Go-Live — the point at which you make the Assistant available to End Customers on a live channel.
• Output — anything the Assistant generates: replies, summaries, scores, proposed bookings.
• Your Content — the business information, documents, settings and conversations you provide or that flow through your account.
• Conversation / Credit — the units your plan is metered in, as described on the pricing page and in your account.

3. What the service is — and what it is not

We provide software that answers messages from your customers on channels such as a website widget, Telegram, WhatsApp, Instagram and Facebook; answers questions from the business information you supply; offers and makes appointments in a calendar you connect; captures enquiries; and hands a conversation to a person when it should.

The Assistant is scoped to your business by design. It is instructed to answer from your Knowledge Base and business details, not to invent prices, availability or facts, and to hand over to a person rather than fill a gap with something plausible. That is how it is built and how we intend it to behave. It is not a guarantee that it will never produce a wrong answer, and section 8 sets out what follows from that.

What it is not

• It is not professional advice of any kind — not medical, legal, financial, or otherwise regulated — and must not be presented to your customers as if it were.
• It is not a substitute for a human where a human is required, whether by law, by your professional obligations, or by good judgement.
• It is not a guarantee of any commercial outcome. Nothing on our marketing pages is a promise of a number of leads, bookings, or revenue.
• It is not a system of record. Keep your own records of your bookings and your customers.

4. Accounts, users and access

You are responsible for who has access to your account and for what they do with it. Invite colleagues only where they need access, remove them when they leave, and keep credentials confidential.

Sessions on our dashboard stay signed in for up to 30 days when the person chooses to be remembered. You can end every session on every device immediately from your settings, and you should do that if a device is lost or a colleague departs.

Tell us promptly if you believe an account has been compromised. We may suspend access without notice where we reasonably believe it is necessary to protect the platform, other clients, or End Customers — and we will tell you why as soon as we can.

5. Trials

Where we offer a trial, it is provided as-is for evaluation. We may change or end trial availability at any time. Usage during a trial still counts against the metering described in section 13, and a trial does not remove your obligations under sections 7, 9, 10 and 11 — a test assistant that reaches a real customer is a live assistant.

6. Testing, and your decision to go live

This section describes the most important thing you do on this platform, so it says plainly who decides what.

We build and configure the Assistant’s instructions. You supply the business information, you test it, and you decide when it goes live. We do not switch on an assistant for you.

Before you go live, you confirm that you have:

1. Tested the Assistant in the Sandbox with questions your customers actually ask, including the awkward ones.
2. Reviewed the Knowledge Base and business details it answers from, and confirmed that prices, availability, opening hours and service descriptions are accurate and current.
3. Reviewed the booking rules — durations, staff, locations, buffers — and confirmed that appointments it offers are ones you can honour.
4. Reviewed the handover settings, and satisfied yourself that conversations reaching a person do so where a person needs to see them.
5. Decided, on your own assessment, that the Assistant is fit to speak to your customers on your behalf.

We record that confirmation with the date, the account that gave it, and the version of these Terms in force. Go-live is your decision and your responsibility. Where the product shows a readiness indicator, it is a checklist to help you, not our approval and not our assessment that the Assistant is correct — an indicator at 100% means the fields are filled in, not that the answers are right.

7. Your responsibilities

You are responsible for the business you are running. Specifically, you agree that:

• The information you give the Assistant is accurate. It will repeat your prices, hours and service descriptions to your customers. If they are wrong, it will be wrong, confidently and at scale.
• You have a lawful basis to communicate with the people you communicate with, and where the channel or the law requires opt-in — as WhatsApp does — you can evidence it.
• You comply with the rules of the channels you connect, including the WhatsApp Business and Meta platform policies, and you accept that a breach of those rules can get your number or page restricted by the platform, which is outside our control.
• You tell your customers what you are doing — see section 10.
• You do not use the Assistant to give regulated advice without your own professional controls and supervision, and you do not configure it to make representations you would not stand behind in writing.
• You keep a human available to take over conversations that need one, at a standard appropriate to your business.
• You have the right to provide Your Content, and providing it to us does not breach anyone else’s rights.

8. AI output — what you can and cannot rely on

The Assistant is built on large language models. These produce text that is statistically likely given their input. They do not know things, and they can be confidently wrong.

We reduce that risk in the ways the product describes: the Assistant is scoped to your business, instructed to answer only from your Knowledge Base, instructed not to invent prices or availability, instructed to hand over rather than guess, and phrases you block by name are checked before a message is sent.

None of those measures is a guarantee, and we do not offer one. They are mitigations, and mitigations fail. In particular: an instruction not to invent an answer is an instruction, not a mechanism; a handover happens when the system recognises it should, and recognition is imperfect.

Accordingly: Output is not verified, is not advice, and you are responsible for what your Assistant says to your customers. You should review Outputs, monitor live conversations, and act on what you find. Do not put the Assistant in a position where a single wrong answer causes harm you could not accept.

9. Data protection — and the processor terms

Our roles are set out in the Privacy Policy and repeated here because they are contractual. For your account data we are the controller. For End Customer conversations you are the controller and we are your processor, processing only on your documented instructions.

What you warrant

• You have a lawful basis for the processing you instruct us to carry out.
• You have given your End Customers the privacy information the law requires, including that an AI assistant handles their messages, that those messages are recorded and stored, and who to contact about it.
• You will not instruct us to process data in a way that would put us in breach of applicable data protection law.
• You will not deliberately route special-category data through the platform without telling us first and agreeing what additional measures apply.

What we undertake, as processor

• To process End Customer data only on your documented instructions, including for transfers, unless legally required otherwise — in which case we will tell you unless prohibited.
• To ensure our personnel are bound by confidentiality.
• To apply appropriate technical and organisational measures (Art. 32) — described in the Privacy Policy.
• To use sub-processors only under written terms no less protective than these, and to give you 30 days’ notice before adding or replacing one so that you may object on reasonable data-protection grounds.
• To assist you, so far as we reasonably can, with data subject requests, breach notification, DPIAs and prior consultation.
• To delete End Customer data on your instruction, and to delete or return it at the end of the contract, except where we must retain it by law.
• To make available the information you reasonably need to demonstrate our compliance, and to submit to audits on reasonable notice, no more than once a year unless a regulator or a breach requires otherwise.

These terms are our Data Processing Agreement and form part of this contract. If your organisation requires a separately signed DPA, ask and we will provide one; where it conflicts with this section, the signed document prevails.

10. Recording and storing conversations

Conversations with your Assistant are recorded and stored. This is not incidental — it is how the service works. The Assistant needs the conversation so far in order to answer coherently, you need it in your inbox in order to take over, and you need the history in order to know what was said to whom.

Stored conversations include the messages in both directions, any attachments and voice notes, the channel identifier, and anything the person volunteered. They are visible to the users you have given access to, and to a limited number of our staff where necessary for support or investigation.

You must tell your End Customers that this happens. The website widget displays a notice for this purpose and links to your privacy policy; you must keep that notice in place and make sure the policy it points to is yours and is accurate. On messaging channels, the disclosure has to come from you. Removing, suppressing or contradicting those notices is a breach of these Terms.

11. The Assistant must not pretend to be human

Article 50 of the EU AI Act requires that people interacting with an AI system are informed that they are. It has applied since 2 August 2026 and the penalties are substantial.

You therefore may not configure, instruct or name your Assistant so that it denies being an AI, claims to be a specific real person, or is otherwise presented in a way designed to make an End Customer believe they are talking to a human being. Giving the Assistant a friendly name is fine; telling someone who asks "are you a real person?" that it is, is not.

We may enforce this in the product, and we may suspend an Assistant that breaches it. This is one of the few places where we will act without waiting for you.

12. Things we depend on and do not control

The service depends on third parties: messaging platforms (WhatsApp, Telegram, Instagram, Facebook), scheduling and calendars (Cal.com, Google), payment processing (Stripe), AI model providers, and infrastructure providers.

Their outages, rate limits, policy changes, account restrictions and pricing decisions are outside our control and are not a breach of these Terms by us. Where a platform restricts or removes your account on its own network — which it may do for reasons of its own — we can help you understand what happened, but we cannot reverse it.

Your use of those services is also subject to their own terms, and you are responsible for complying with them.

13. Fees, plans and metering

Fees, plan limits and billing periods are as shown when you subscribe and in your account. Unless stated otherwise, fees exclude VAT and other taxes, which are added where applicable.

Usage is metered in the units your plan describes. Figures shown in the dashboard are estimates for your guidance; the invoice is the record.

Upgrades, downgrades and cancellation

• An upgrade takes effect immediately and is reflected on your next invoice.
• A downgrade takes effect as described at the point you make it. Downgrading may reduce the number of user seats, channels or conversations available to you, which can remove access for colleagues — check what a change does before you confirm it.
• You may cancel at any time. Cancellation stops future renewals; it does not refund the current period unless we say otherwise or the law requires it.
• We may change prices with at least 30 days’ notice before your next renewal. If you do not accept a change, cancel before it takes effect.

If payment fails we may suspend the service after notifying you and allowing a reasonable period to fix it. Suspension does not delete your data — see section 18.

14. Acceptable use

You must not use the service to:

• Send unlawful, deceptive, harassing, defamatory or abusive content, or content that infringes anyone’s rights.
• Send unsolicited bulk messages, or message people who have not given the consent the channel or the law requires.
• Impersonate any person or organisation, or misrepresent your affiliation with one.
• Attempt to make the Assistant act outside the business it is scoped to, or to extract its instructions or other clients’ data.
• Probe, scan, overload or interfere with the platform, or circumvent metering, rate limits or access controls.
• Resell or provide the service to third parties as your own without a written agreement with us.
• Process data you have no right to process.

We may suspend or terminate for breach of this section, and for serious or repeated breaches we will do so without notice.

15. Intellectual property

We own the platform — the software, the interface, the agent architecture, the prompt design, the templates, and everything we build to make it work. Nothing here transfers any of it to you. You get a non-exclusive, non-transferable right to use the service while this contract is in force.

You own Your Content. You grant us the licence we need to host, process, transmit and display it in order to run the service for you, and to send it to the sub-processors identified in the Privacy Policy for that same purpose. That licence exists only to operate the service and ends when the content is deleted.

We do not use Your Content to train AI models, and we contract with our model providers on terms that exclude it from theirs.

If you send us feedback or suggestions, we may use them without obligation to you. We are not asking for your ideas; we are declining to create a dispute if you volunteer one.

16. Confidentiality

Each of us may learn confidential information about the other. Each agrees to use it only for the purposes of this contract, to protect it with at least reasonable care, and not to disclose it except to people who need it and are bound to keep it confidential.

This does not apply to information that is public through no fault of the recipient, was already known, is independently developed, or must be disclosed by law — in which case the recipient will give notice where it lawfully can.

17. Warranties and disclaimers

We warrant that we will provide the service with reasonable skill and care, and in accordance with these Terms.

Beyond that, and to the fullest extent permitted by law, the service is provided “as is”. We do not warrant that it will be uninterrupted or error-free, that the Assistant’s Output will be accurate or complete, or that it will achieve any particular result for your business. Any implied warranties are excluded to the extent the law allows.

Nothing in this section limits rights you have under mandatory law that cannot be excluded.

18. Limits on liability

Read this section. It is the one that decides what happens when something goes badly wrong.

Subject to the exclusions below, our total aggregate liability arising out of or in connection with this contract, in any twelve-month period, is limited to the fees you paid us in the twelve months before the event giving rise to the claim.

Neither party is liable for indirect or consequential loss, loss of profit, loss of revenue, loss of anticipated savings, loss of business or goodwill, or loss or corruption of data, however arising.

What is never limited

Nothing in this contract limits or excludes liability for:

• death or personal injury caused by negligence;
• fraud or fraudulent misrepresentation;
• gross negligence or wilful misconduct;
• anything else that cannot lawfully be limited or excluded under Spanish or applicable EU law.

19. Indemnity

You will indemnify us against claims, losses and reasonable costs arising from: Your Content; your use of the service in breach of these Terms or applicable law; your failure to give End Customers the information required by section 10; and claims by End Customers arising from the instructions you gave the Assistant or the accuracy of the information you supplied it.

We will tell you promptly about any such claim, let you control the defence of it where you confirm you will indemnify us, and cooperate reasonably at your cost. You may not settle in a way that admits fault on our part or imposes an obligation on us without our consent.

20. Term, suspension and termination

This contract runs until either of us ends it. You may cancel at any time. Either of us may terminate for material breach that is not remedied within 30 days of written notice, or immediately if the other becomes insolvent.

We may suspend the service where necessary to protect the platform, other clients, or End Customers — including for a breach of section 11 or section 14. We will tell you why, and restore access when the cause is resolved.

What happens to your data

• Your Assistant stops answering when the contract ends.
• You have 30 days from termination to export your data. Ask us and we will help.
• After that we delete or irreversibly anonymise it within a further 60 days, except records we are required to keep by law — principally invoices and accounting records.
• You may instruct us to delete End Customer data sooner, and we will.

Sections that by their nature should survive termination do so: confidentiality, intellectual property, liability, indemnity, and governing law.

21. General

• Force majeure. Neither party is liable for failure caused by events beyond its reasonable control.
• Assignment. You may not assign this contract without our consent, not to be unreasonably withheld. We may assign it to a group company or in connection with a merger or sale of the business.
• Sub-contracting. We may use sub-processors as described in section 9.
• Entire agreement. These Terms, the Privacy Policy and the Cookie Policy are the whole agreement between us on this subject, and replace anything said beforehand. Neither party relies on any statement not set out here — except that nothing excludes liability for fraudulent misrepresentation.
• Severability. If a provision is unenforceable, the rest stands and the provision is read down to the minimum extent necessary.
• No waiver. Not enforcing a term once does not waive it.
• Notices. To us: hello@opusdone.com, and by post to Carrer de Calàbria, 64, Eixample, 08015 Barcelona, Spain. To you: the email address on your account — keep it current.

22. Governing law and disputes

These Terms are governed by the laws of Spain. The courts of Barcelona have exclusive jurisdiction over any dispute arising out of or in connection with them.

If you are established in the United States, that choice of law and forum still applies as between us as businesses. It does not displace any mandatory consumer-protection or state law that applies to you regardless of contract, and nothing in these Terms is intended to waive a right that cannot lawfully be waived.

Before starting proceedings, please raise the issue with us in writing. Most disputes are a misunderstanding about what the product did, and those are quicker to resolve by looking at the conversation together.

23. Changes to these Terms

We may update these Terms. The version number and effective date at the top of this page change when we do, and the version history at the bottom says what changed.

For material changes we will email account owners at least 30 days before they take effect. Continuing to use the service after that means you accept them; if you do not, cancel before the effective date.

24. Contact

Questions about these Terms: hello@opusdone.com. Softpact, Carrer de Calàbria, 64, Eixample, 08015 Barcelona, Spain.