Questions to ask before you hire us

Getting started

What is an AI agent, in plain words?

An AI agent is software that reads your information, decides on a next step within rules you set and then acts in your systems. It might draft a reply for a person to approve, update a record after a call or book a visit.

Is now the right time, or should we wait?

It's a good time when the same kind of request arrives many times a week and the answer depends on information you already hold. It's worth waiting when your records disagree with each other or no one owns the rules yet. That's where we'd start instead, and we'll tell you so.

We're not sure what we need. Where should we start?

Write to us with a few lines about your goal. We'll suggest the smallest step that would tell you something useful. That's often a Five-Day Read if you have a plan in hand, or a Footing Review if you don't.

What does it cost to find out whether you're a good fit?

It's free. Our reply, a first call of up to 45 minutes and the written proposal cost you nothing.

How quickly will you reply?

Within two working days, with questions or a suggested first step, and the earliest date we could start. After a first call, you'll have a written proposal within five working days.

Can we speak to someone you've worked with?

We keep client work confidential, so this site has no client names, logos or quotes. If a conversation with someone we've worked with would help your decision, ask us, and we'll tell you honestly what we can arrange. You can also check our work for yourself, because our method is written out in full and the Recordist gateway's code is public.

How do you bring our staff in?

From the first week. The people who'll work beside the agent help write its rules and the questions it's tested on. They also help decide where a person steps in, and their feedback shapes each release.

Your data and security

Will our data be used to train AI models?

We don't use your data to train any model. Your data goes into an AI service only after you've approved that service in writing, with retention and training switched off wherever the provider allows it. We show you those settings before we start.

How do you handle our data during the work?

We work inside your systems, with the access your team grants, and we ask only for what the work needs. We don't copy your data to our own computers unless you've agreed to it in writing. When a copy can't be avoided, we keep the smallest amount for the shortest time, then delete it and confirm that in writing.

Can our data stay in Canada, or in our own region?

Yes. We build in your environment, so your data stays wherever your systems keep it. If a rule says it must stay in one place, tell us. We'll only use services that can respect it, and we'll say so if that rules out a tool you wanted.

Do you use AI tools to do the work?

Yes, where your policies allow it. We only use an AI tool with your code or data after you've approved that tool and its settings in writing. A person reviews every change before it's merged, and the tests decide whether it's ready to ship.

What happens to our data when the work ends?

We send you a list of every access to remove, and we delete any copy we held and confirm that in writing. Your code, tests and documents are already in your own systems, so there's nothing to hand back.

Can you work with our security review and procurement?

Yes. We'll answer your security questionnaire in writing, follow your policies and use your accounts and devices if you require it. We don't hold a formal certification such as SOC 2 or ISO 27001, so we answer your questions directly and show you how we work. Our trust page sets out what we commit to.

Ownership and confidentiality

Who owns the code and everything else you build?

You do, once it's paid for. The code lives in your repositories from the first day, so you always hold it. If we bring in something we made before, the proposal says how you may keep using it. Open-source parts keep their own licences, and we list them for you.

Will you sign our NDA?

Yes. Send it over, and we'll review it and sign a reasonable one before you share any details.

Will you name us publicly?

We never name a client in public, on this site or anywhere else, unless they've agreed to it in writing. That covers logos, quotes and case studies too.

Independence

Are you tied to any software vendor or platform?

We aren't tied to any vendor or platform. We work on the one you already run whenever it can do the job, and we'll tell you plainly when it can't. If we ever have a commercial relationship with a tool we recommend, we'll tell you in writing before you decide.

Do you resell software or add a margin to AI usage?

We don't. Any software, licences or AI usage you need is bought in your name, directly from the provider. You see the real cost, and you keep the account after we've gone. We estimate those costs in the proposal.

Which platforms do you work with?

We work on the CRM, data and AI platforms you already use, connected through APIs, events and open protocols such as the Model Context Protocol. We've built on CRMs, service desks, customer portals, customer data platforms, integration platforms, identity and sign-in services, and the AI services that run on top of them. Ask us about yours, and we'll tell you plainly which ones we've built on before and where we'd be learning alongside you.

When is a ready-made agent product the better choice?

When it does the job within rules you can accept, at a cost you understand at your volume. We'll say so plainly, and we can write the rules and tests around it so you can check that it keeps them.

If you review another firm's proposal, will you try to win the work?

If we read another firm's proposal for you, we won't bid for that same work unless you ask us to. You get our honest read, with no sales pitch attached.

Who does the work

Who will actually do the work?

We're a small company by design, and we take on a few engagements at a time. Your proposal names everyone who'll work with you, and they're the people who do it. We don't bring in anyone else, including subcontractors, without your written agreement.

What if the person working with you is ill or leaves?

We write everything down as we go: the rules, the tests, the runbooks and the decisions. Your code lives in your repositories, so the work doesn't depend on one person's memory. If someone named in your proposal can't carry on, we'll tell you straight away, and you choose what happens next. We can pause the work, and you won't pay for the pause. We can propose a replacement for you to meet and approve. Or your own team, or anyone you choose, can pick it up from where it stands.

Where are you, and what hours do you work?

We're a Canadian company, and we work remotely. Our hours overlap the North American business day. If you're in another time zone, we'll agree the hours we overlap in the proposal.

Will you come to our office?

We work remotely by default. If being in the room matters for your project, tell us, and we'll agree any travel and its cost in writing before anyone books a ticket.

Can you work alongside our IT team or another supplier?

Yes. We fit into your team's ways of working and your release process, and we're glad to work next to another firm. The proposal says who decides what, so everyone knows who owns each piece.

Do we have to use Recordist to work with you?

No. Recordist is our own product, and it's separate from client work. We'll only suggest it if it fits what you need.

Money, risk and results

Why aren't prices on this site?

Every piece of work has a fixed price or a fixed monthly fee, set out in a written proposal before anything starts. We price each piece once we understand it, because a price for a problem we haven't seen would be a guess. Work you buy from us directly is never billed by the hour, so you won't get a surprise invoice.

How do you invoice?

The proposal sets it out. A Five-Day Read is invoiced when you receive it. A Footing Review and each build phase are invoiced half at the start and half at the end. A Proving Run's two phases are invoiced as each one starts, and monthly work at the start of each month. The proposal states the currency, and sales tax is added where it applies.

What if the pilot doesn't work?

You'll find out early, and at a known cost. A Proving Run has a stop point after two weeks, and if either of us stops there, you pay only for the first phase. If it runs to the end and misses the measure, you still keep everything we built. You also get a written account of why it fell short and what would need to change.

How do you measure success?

We agree the measure with you in the first week, in your terms, before any building starts. For an agent, that might be how often it answers correctly from your data and whether it stays inside its rules. It might also be whether it hands over to a person when it should. We report against it every week, in plain words.

Do you tie your fees to results?

Our fees are fixed, and we tie the stop points to results instead. Paying us for each conversation an agent closes would reward it for keeping people away from a person. We'd rather it handed over whenever it should.

What if we need to change the scope?

Tell us, and we'll price the change in writing before we make it. You decide whether it goes ahead, and your invoice changes only after that written yes.

Do you fix what you build if it breaks?

The proposal sets out how we fix any defect in what we delivered. After that, the Long Run can look after it, or we'll price each fix separately.

Rules, regulation and accessibility

Can you help us meet AI rules and regulations?

We build the evidence that auditors and regulators tend to ask for. That means written rules, the tests that check them, evaluation records and clear points where a person reviews or takes over. We're not lawyers, so we work alongside your legal advisers and don't give legal advice.

Will what you build work for people with disabilities?

We aim for WCAG 2.2 at level AA on every screen people use. We check with automated tests on every change, and by hand with a keyboard and a screen reader. We'll tell you plainly what's been tested and what hasn't, and our accessibility statement shows how this site was checked, including what's still to do.

After we finish

What happens after the handover?

Your team runs it. If you'd like us to keep an eye on it, the Long Run is a monthly arrangement. After its first three months, you can end it with 30 days' notice. If you'd rather ask for help now and then, we'll price each piece separately.

Can we bring the work in house, or move to another firm later?

Yes, at any time. Your code, tests, notes and runbooks are already yours and in your systems, and we'll help hand over to whoever takes the work on.