Before you buy AI

    Thirteen questions to ask before you buy AI.

    These are the questions owners, CFOs and operators ask most before they put AI into a business. Here is how we answer them. Ask every one of them to anyone you're considering, us included.

    1. 01

      Where do we start?

      Where your customer would feel it first: the call nobody answers at night, the quote that waits until Monday, the question your team answers forty times a week.

      Pick one workflow, put it live, and watch it. Decide what comes next once it's running. A long list of ideas usually ends with nothing shipped.

      Ask them“Which single workflow would you build first, and why that one?”

    2. 02

      Can't we just build this ourselves with Claude or ChatGPT?

      For a lot of it, yes. Reports, research, drafting, content: get your people using AI every day. You don't need a firm for that, and we'll tell you so.

      You need a builder when the work touches a system of record, customer, patient or payment data, or when it has to run every day without someone babysitting it. Our AI audit is yours to keep: build from it with us, or take it and build it yourself.

      Ask them“What part of this could we do ourselves, and what genuinely needs you?”

    3. 03

      How do we measure ROI?

      Name the number before anything gets built: calls answered after hours, days to close, hours per onboarding, quotes sent the same day. Write down where it stands today.

      If nobody can name the number, you're buying a demo. We ask for it in the first conversation, and the build is judged against it.

      Ask them“What number will this move, and where does it stand today?”

    4. 04

      How do we know the AI is right?

      You test it. We build a set of real questions with answers your team agrees are correct, and score the system against that set before it goes live, and again after major changes. Anything below the bar goes to a person.

      And the AI never does the math. Prices, costs, balances, anything auditable runs on ordinary code that can't make things up. The AI handles the language, the conversation and the judgment calls, inside the knowledge you gave it.

      Ask them“Show me how you test it, and what happens when it isn't sure.”

    5. 05

      Who owns it, and who keeps it running?

      Your data, your content, your domain and your accounts are yours outright and exportable anytime. The engine is licensed to you and maintained by us, and if you walk away your system keeps running. Owning the source outright is a separate arrangement, said up front.

      Before kickoff, name the person on your side who will run it day to day. The best first AI hire is someone who can operate the system and keep it healthy, not someone to write prompts.

      Ask them“What exactly do we own, and what happens to the system if we stop paying you?”

    6. 06

      Is our data safe?

      It should live in your accounts. You run on your own AI, voice and hosting keys, so your data sits where you can see it, and it is never used to train public models.

      Ask three things of any vendor: where the data goes, who at the vendor can see it, and what the model provider keeps. You should get a plain answer to all three. You'll get one from us on the first call.

      Ask them“Where does our data go, who on your side can see it, and what does the model provider keep?”

    7. 07

      Do we have to fix our data first?

      Only the data your first workflow touches. Then build in order: a simple pipeline that runs every day, a layer that defines what your numbers mean, and then the agent on top.

      One clever prompt can't do all three. That's why our systems keep the numbers in code and the language in the AI.

      Ask them“Which of our data does the first workflow need, and in what shape?”

    8. 08

      What happens to our people?

      Start with them. Agents don't work on top of a team that doesn't use them, and the hardest part of adoption is never the technology. It's fear, overload and the quiet workaround.

      Gabriela Cruz designs adoption at Anabasis, with a background in psychotherapy and years leading teams. Your team is part of the build from the first conversation, not handed a tool at the end.

      Ask them“How will our team learn this, and who on your side helps them?”

    9. 09

      Which model should we use?

      It matters less than you think, and you'll change it. Pick a setup that lets you switch providers without starting over.

      We already move live agents between models as better ones arrive, on your own keys. What you keep is the part that matters: your workflows, your tests, your data.

      Ask them“If a better model comes out next month, what does switching cost us?”

    10. 10

      What will it cost to run?

      Get a written estimate of the monthly running cost before the build: model usage, hosting and licences.

      Because you run on your own accounts, those costs bill to you directly and nothing meters back to us. You can see every line.

      Ask them“Give us a written estimate of the monthly running cost.”

    11. 11

      Do we pay for time or for outcomes?

      Either can work, as long as you can walk away. We measure the work by what it does for your business, not by hours billed: clear scopes, no hidden fees and no surprise invoices.

      Larger builds start with a paid AI audit. The report is yours to keep, and its fee is credited to the build if you build with us. Smaller builds are a simple flat fee. Either way, there is a checkpoint once the first workflow is live: if the number hasn't moved, you've lost nothing you can't take with you.

      Ask them“What do we pay for, what do we get to keep, and where is the first checkpoint?”

    12. 12

      How do we tell a real AI partner from one that started last year?

      Before you sign, ask to see something they built that's running in production for another business today, and use it yourself.

      Everything we sell is already running for a real company, including the voice agent on this page. Talk to it. Then ask us for the live links.

      Ask them“Show me something you built that's running in production today.”

    13

    And the one most people forget: what happens if it works?

    Who runs it, who improves it, and what does your team do with the time it gives back? Ask it before you start. We do: it's one of the questions our agent asks every business on the way in.

    Have these answered before kickoff

    • The one workflow you start with
    • The number it moves, and where that number stands today
    • The person on your side who runs it
    • The data that first workflow touches
    • A written estimate of the monthly running cost
    • What your team does with the time it gives back

    Ask Aleksandr any of these right now, by voice or text. Or describe your business and get a page built for it.