FAQ
Frequently Asked Questions | AI for Irish Businesses
Twenty questions we get asked before anyone commits to anything — cost, timelines, GDPR, integrations, and whether this replaces your team. Answered straight, without the jargon.
Twenty questions we get asked before anyone commits to anything — cost, timelines, GDPR, integrations, and whether this replaces your team. Answered straight, without the jargon.
The signals are usually operational, not technical. People re-keying the same data between systems. Information sitting in three places with no single view. A team spending its week on admin instead of the work you actually sell. Headcount going up to handle paperwork rather than delivery. If two or three of those sound familiar, there is almost certainly something worth building.
The honest counter-case: if the process changes every time it runs, if nobody has agreed what good looks like, or if the real problem is pricing or sales rather than operations, AI will amplify the mess instead of fixing it. The free Discovery Call exists to tell you which of those you are — if it's the second one, we'll say so. Our guide on whether your business is ready for AI goes through the same test in more detail.
Book the free 20-minute Discovery Call. That's the only entry point, and it costs nothing. From there the path is a paid Discovery Report, then a fixed-price build — and you can stop at any step. Nothing is committed until you've seen what it costs in writing. How it works walks through all four steps.
No. There's nothing to fill in and nothing to send us in advance. If you want the twenty minutes to go further, it helps to have a rough idea of the two or three tasks that eat the most time in your week, and the main systems you run the business on — accounting, CRM, job management, whatever they are. But turning up with none of that is fine. The questions we ask will get there anyway.
Twenty minutes, video or phone. It's a diagnostic, not a sales call with a different name. We ask specific questions about your workflows, your revenue model, your team, and where the operational pain actually is. You leave with an honest read on whether your business is ready for AI, and where we'd start if it is.
No deck, no proposal in your inbox that evening, no follow-up pressure. If we're not the right fit, we tell you on the call and we tell you why.
At Keystone the ladder is: the Discovery Call is free. The Discovery Report is €1,000, delivered within 48 hours of a deep-dive call, and credited in full against the build if you proceed. The first build is a fixed €1,000–€20,000 depending on scope, agreed in writing before any work starts. Additional workflows after that start from €1,000 each and get cheaper as the hub compounds.
No retainers, no hourly billing, no change orders inside agreed scope. Our average client sees a 3× annual saving against what they spent on implementation. Full pricing has the detail on each step.
Typically four weeks from kickoff to a system live on your real data, with your team using it. Before that, the Discovery Report lands within 48 hours of the deep-dive call, and the deep-dive itself is usually held within a few days of the first call.
The first build takes the longest because it lays the foundation — the data layer, the integrations, the intelligence layer. Everything built on top of it afterwards ships faster.
Yes, and plenty of people do. The Discovery Report is decision-grade material that stands on its own: the top three AI opportunities in your business, what each would cost to build, the projected return, and the architecture behind it.
It's yours. Build it in-house, hand it to another supplier, or use it to decide AI isn't the right move this year. The €1,000 is only credited back if you go on to build with us.
No. There's no monthly fee, no minimum spend and no lock-in. You commission a piece of work, we build it for the agreed price, and that's the end of the obligation. What we build keeps running whether or not we're building anything else for you, and you decide if and when there's another workflow worth commissioning.
Scope, price and delivery date are agreed in writing before any work begins. That number doesn't move. We don't issue change orders against work that was inside the scope we agreed, and if something takes us longer than we estimated, that's our problem rather than your invoice.
If you want to add something genuinely new mid-build, we price it separately and you decide. Nothing gets added to the bill without you agreeing to it first.
We fix it. Every build is scoped against a specific metric agreed with you at the start — hours saved, days sales outstanding, response time, whatever the workflow is meant to move. That metric is measured after go-live against the baseline we captured before it.
If the system isn't delivering what we said it would, we keep working on it. That's the outcome guarantee, and it's the reason we're careful about what we promise at scoping.
Yes, and that's the normal case rather than the exception. We build on top of the systems you already run instead of asking you to move. One of our own builds, AR Guardian, reconciles invoices across Salesforce and Xero. Others sit on job management systems, shared drives and email.
If a system has an API, we can usually work with it — and if it doesn't, there's generally another route to the data. Which of your systems are in scope is one of the things the Discovery Report answers, by name.
No. This is the most common reason business owners talk themselves out of starting, and it's usually wrong. Most builds run on what's already there — invoices, emails, job records, timesheets, the spreadsheets people actually use. You don't need a data warehouse or a data team.
If the data is messy or scattered across systems, that's normal, and tidying it into one usable layer is part of the first build rather than a prerequisite for it. What matters more than volume is that the process is consistent enough to describe.
Not in the way it's usually meant, and not in how we build. What we automate is the repetitive, rules-based work — moving data between systems, chasing invoices, updating records, triaging inbound email. What stays with your people is judgement, relationships and the exceptions, which is where the value of an experienced team actually sits.
In practice the effect our clients see is growth without adding headcount to handle paperwork. We're not going to tell you no role in any business ever changes because of this. We will tell you that the builds we do are aimed at the admin, not at the people.
An AI hub is a single intelligence layer that sits across the systems you already run. It holds your data, keeps the connections to your existing tools in one place, and runs the agents and workflows on top. The alternative — a separate point tool for each problem — means paying for the same plumbing over and over.
The hub matters because of what happens after the first build: each new workflow plugs into infrastructure that already exists, so it costs less and ships faster than the one before it. If you run more than one company, the same hub can serve all of them. The AI glossary explains the rest of the terms in plain English.
AI isn't compliant or non-compliant in itself. The deployment is what's compliant or not. Under GDPR and the Irish Data Protection Act 2018 the questions are the same as for any other system: what personal data is being processed, what your lawful basis is, where the data is stored and processed, who can access it, how long it's kept, and whether people have been told.
A workflow that reads your own invoices inside your own tenancy is a very different question from one that sends client files to a third-party model with open-ended training rights. We design for those questions at scoping rather than after the fact, and for sensitive sectors that includes EU data residency. This is general information, not legal advice — your own adviser or DPO signs off the final position.
If a supplier processes personal data on your behalf, GDPR requires a written contract between you. That's the Data Processing Agreement, and Article 28 sets out what it has to cover: what they can do with the data, where it's stored, who can see it, and what happens at the end.
You remain the controller and stay accountable, which is exactly why the paperwork matters. Any serious AI supplier will sign one without a fuss. If a supplier hesitates, that tells you something.
Sometimes. A Data Protection Impact Assessment is required under Article 35 of GDPR where processing is likely to result in a high risk to people's rights — typically special category data such as health records, large-scale monitoring, or automated decisions that have a legal or similarly significant effect on someone. The Irish Data Protection Commission publishes a list of processing operations that require one.
A lot of ordinary SME automation doesn't reach that bar: matching invoices, reconciling ledgers, triaging your own inbox. Clinical or patient-facing work in a healthcare setting usually does. Where a DPIA is needed we provide the technical documentation to support it — data flows, storage locations, access controls — but the assessment itself is yours to complete with your DPO or adviser.
Probably — though not necessarily in your sector, and that's deliberate. Keystone built operational AI inside the partners' own businesses first: waste management, financial services, construction and legal services, before offering it to anyone else. On the client side that includes an operations hub running five workflows for an Irish construction company and call transcription with action-item extraction for an Irish solicitor practice.
What we sell is operational pattern recognition rather than sector specialisation. An invoice reconciliation problem looks much the same in a haulage business as in a law firm once you get underneath the language.
By agreeing the number before the build starts, not after. Every engagement is scoped against a specific metric — hours saved on a named task, days sales outstanding, percentage of invoices matched without a human, time to first response. We capture the baseline before go-live so there's something honest to measure against, then report the result once the system has been running on real work.
Across our clients the average is a 3× annual saving relative to what the implementation cost. If a build isn't moving its metric, that shows up in the measurement rather than in a quarterly review deck.
Ireland is where we're based and where most of our work is, and Irish SMEs turning over €1M–€50M are who the offer is built around. That said, the work itself is done remotely and the operational patterns aren't specific to one jurisdiction — we keep a US line alongside the Irish one.
If you're outside Ireland, book the Discovery Call and we'll give you the same honest read on fit, including whether time zones or local compliance make us the wrong choice.
We'll tell you honestly whether Keystone can help your business, where we'd start, and what each step would cost. If we're not the right fit, we'll say so.
Book Your Discovery Call