When you hire someone to build an AI system, a well-run build starts with them learning how your business works before they quote. Then you get a written scope and price, regular updates while it's built, testing you can see and sign off, a launch done in stages and a clear plan for after launch. Your data should stay yours, and you should be able to get it out whenever you want. If a builder skips any of that, take it as a warning.
Do you need a builder at all?
Start here, because sometimes the honest answer is no. If an off-the-shelf chatbot or the AI features already in your CRM cover what you need, use them. They're cheaper, and you can switch them off.
A custom build is worth the money when the AI has to work across your own systems. So reading your booking calendar, updating your CRM, following your qualification rules, and handing over to your team in a specific way. You can see what that looks like for us on the platform page.
What each stage should look like
Every builder calls the stages something different, but a proper project goes through the same ones. This is what you should get at each stage, and what should worry you.
| Stage | What you should get | Red flag |
|---|---|---|
| Audit | A map of how enquiries arrive and where work gets stuck, with the cost of each problem | A quote before anyone has asked how your business works |
| Design | A written scope: what is built, what isn't, the systems it connects to, who can do what, and how success is measured | "An AI solution" with no list of what it does |
| Build | Regular short updates you can see | Silence for weeks, then one big demo |
| Test | Real and awkward cases run before any customer sees it, with results you sign off | "We tested it" with nothing to show |
| Launch | A staged rollout, team training and written documentation | Everything switched on at once, with no handover |
| Improve | A review of real conversations and numbers, with agreed next changes | No plan for anything after go-live |
This is the process we follow. Testing is the stage that's easiest to rush when there's a deadline, which is why it gets its own section further down.
What it should cost, and what you pay separately
Expect two kinds of cost, and make sure the proposal keeps them apart.
The build. A clear price for the agreed scope. For us it starts with a 30-minute call, then we demo your own system, then we run a pilot with one team before it goes out to everyone. We set the price per business once you've seen it working, and it runs on a contract. Be careful with any builder who promises a delivery date before there's a signed scope.
Running costs. AI usage, hosting, phone minutes and software subscriptions. These should be written down clearly, so you can see what you're paying for and nobody hides a markup in them. Ask for a monthly estimate at the volume you expect before you sign.
Then there's who looks after it once it's live. With us that's part of the contract, because we run the system for you. We host it, look after it, fix what breaks and keep improving it. Whoever you hire, get the terms in writing.
What you should own
Before you sign, get these in writing. If the answer to any of them is "we keep that", ask why.
- Your data, including every conversation, can be exported whenever you ask
- It's written down what happens to your data if you leave, and how fast you get it
- You can see your own users, settings and records, even if the builder runs the system
- The approved information the AI answers from is written down and editable by you
- Documentation explains every connection, setting and account
This matters most on the day you want to change supplier. If your data comes out cleanly, that's one email. If it doesn't, you're starting again. With us it runs on Agentryx One, your data is yours, and if you leave we export it for you.
Questions to ask before you sign
Good builders answer these clearly and without having to think about it.
| Question | A good answer sounds like |
|---|---|
| What happens when the AI doesn't know the answer? | It says so and hands the conversation to a named person, with the context attached |
| How will you test it? | A written list of cases, including awkward ones, that you review and sign off |
| What can it never do? | A specific list: no advice in regulated areas, no prices outside the approved list, no complaint handling |
| What does it cost to run each month? | An estimate at your volume, written into the proposal |
| Who owns the data, and can I get it out? | You do, and they export it for you if you leave |
| Can I see something similar you've built? | A live example or a case study with specifics |
The first question is the most important one. The builder is making something that talks in your name, and you're still responsible for what it says. In Moffatt v Air Canada (2024), a Canadian tribunal held the airline liable for wrong fare information its website chatbot gave a customer, and threw out the argument that the chatbot was responsible for its own actions.
What you'll need to provide during the build
AI projects slow down when the builder's waiting on access or decisions. Have these ready before the build starts:
- Access to the systems it connects to, like your website, booking system, CRM and inbox.
- The approved information it answers from. Services, published prices, policies, opening times and common questions with the answers you'd want given.
- One person who can answer questions and approve decisions within a day.
- Real examples of enquiries, including the difficult ones, to test against.
Testing you should see and sign off
Before any customer uses the system, you should see it deal with real cases. An angry customer, a question it mustn't answer, someone asking you to stop contacting them, an enquiry at 3am, someone asking whether they're talking to a person. Ask for the test list, the results and what got changed because of them. We lay out a practical test plan in how to test an AI agent before it talks to a customer.
After launch
You can't set up an AI system and forget about it. Your prices change, your services change, the AI models underneath change, and real conversations always show up gaps the tests missed. Decide before launch who reads the conversations each week, who updates the approved information and who fixes problems, whether that's your team or the builder on a contract.
Do this week
Write one page on the problem you want solved. Where enquiries come from, what happens to them now, where they get stuck and what that costs you. Take that page to any builder you're thinking about and see whether their questions go deeper than yours. The ones whose questions do are worth a second conversation.
If you want that conversation with us, book a call. It's 30 minutes, and we'll start with how your business works instead of a pitch.
