Choosing an AI implementation partner for SMEs
By Joey van Vliet, founderUpdated 6 min read
The short answer
A good AI implementation partner for SMEs first works out what manual work is costing you, then lays a foundation in your own systems and automates one process at a time on top of it. It stays after go-live to make adjustments, and tells you when AI isn't the best solution for a process. Pay particular attention to what happens after the first delivery: that's where the difference lies.
About this guide: written by Your Tech Club, with no ranking and no names of other agencies. How we do it ourselves is set out separately under How Your Tech Club works.
What to look for in an implementation partner
A small or mid-sized organisation has no room for a project that gets stuck in a pilot. These eight points help you find a partner that carries out a plan instead of just writing one.
A substantiated figure before anything is built
Every investment is visible, so you want to know what it will deliver. A partner should first show what your current work costs: the hours that go into it, the errors that get fixed and the revenue left on the table.
With that figure you can make a choice, and explain why.
Ask in the meetingHow do you calculate what our current work costs, and can we check every step?Foundation first, then the first process
Separate AI applications per department quickly become unmanageable. A better order: lay the foundation first, with access to the systems, reliable data, integrations, logging and clear ownership. Then automate one process and expand from there.
Everything that follows goes faster because it sits on the same foundation.
Ask in the meetingWhat would our first process be, and what needs to be in order before that?The people who do the work have a say
The people who do the work every day know where it gets stuck and where the exceptions are. A partner that talks to them before building makes something that actually gets used.
Ask how users are involved and how they learn to work with the new setup.
Ask in the meetingWho in our organisation do you want to talk to, and how do you bring the team along?Oversight and control are part of the delivery
You want to know what an automation, an AI employee or an agent has done, and you want to be able to step in. Monitoring, logging and human approval where it matters belong in the delivery.
If they appear as a separate option in the quote, ask why.
Ask in the meetingHow do we keep track of what the automation does, and where does someone on our side approve?The approach fits the size of your organisation
An approach designed for a large corporate, with lots of meetings and slow decision-making, is a poor fit for an organisation where everyone is already at capacity.
Ask how the partner works with organisations of your size, who needs to decide on your side and how much time it will take from your team.
Ask in the meetingHow much time will this take from our team, and who needs to decide on our side?Careful with your systems and your data
An implementation partner gets access to your systems and your customer data. Ask how that access is granted and revoked, and which data goes to which supplier.
ISO 27001 for information security and ISO 9001 for quality show that a partner has set this up as a management system and has it audited externally.
Ask in the meetingHow do you manage access to our systems, and how do we revoke it?What gets built stays yours
Code, configuration, prompts and data created for you should remain yours. Then you can expand later, hand it over or carry on yourself, without starting over.
Ask in the meetingWhat do we get if we ever want to carry on ourselves?There is a plan for after the first delivery
Processes change, systems get replaced, and an automation has to move with them. An implementation partner makes adjustments after go-live and decides with you what the next process will be.
Ask how that works in practice: who monitors, who chooses the next step, and how you know whether it's working.
Ask in the meetingAfter the first delivery, how do we decide together what comes next?
Questions to ask in the first meeting
Take this list to your first meetings. A partner that suits you gives a concrete answer to every question, even if that answer is that something can't be done.
- How do you calculate what our current work costs?
- What do you think our first process should be, and why that one?
- What needs to be in order before you can automate?
- Who on our team do you want to talk to?
- How much time will this take from us, and from whom?
- How do we see what the automation does, and where do we approve?
- How do you manage access to our systems?
- Who owns the code, the prompts and the configuration?
- What happens after go-live, and who is our point of contact then?
- When would you advise us not to automate?
What is an AI implementation partner?
An AI implementation partner works out where AI, automation or new software adds value in your organisation, builds it into your own systems and stays involved until it works. That's different from a consultant, who makes recommendations and leaves, or a tool vendor, who delivers a product without adapting your processes.
If you don't have your own team to act on advice or build in a tool, the firm doing the implementation decides whether anything really changes.
Signs a partner isn't right for you
- There's a solution on the table before anyone has looked at your processes.
- Nobody can explain what the current work costs.
- Every piece of advice ends in a build project.
- Monitoring and support appear as a separate option in the quote.
- It's unclear who owns the code and the configuration.
- There's no fixed point of contact after go-live.
How Your Tech Club implements AI
This part is about us. The points above apply to any partner, including us: feel free to hold us to them.
For SMEs we usually start with the AI Opportunity Audit. In five working days, with an intake plus three to five interviews, we map where manual work is holding back your growth and what it costs. The report has five fixed sections: what it costs now, the three highest-value opportunities, what we wouldn't do, what implementation looks like, and the first 90 days.
If you then go ahead with implementation, we lay the foundation first, automate the first process and build from there. Monitoring, logging and human approval are part of it, and what we build stays yours.
The outcome of the audit can also be that you take it on yourself or do nothing for now. If so, the report says that.
- ISO 9001 and ISO 27001, certified by Kiwa
- 500 projects delivered
- 4.8 from 127 reviews
- 90% returning clients
- Office on Duivendrechtsekade in Amsterdam
- Founded by Joey van Vliet and Yassine Seghrouchni
Frequently asked questions
What does an AI implementation partner do?
An AI implementation partner works out where AI or automation adds value in your organisation, builds it into your own systems and stays involved until it works. The difference from a consultant is that it handles the delivery and the support after go-live itself.
What does AI implementation cost for an SME?
That depends on the process, the systems and how much needs to be built. A good partner first works out what the current work costs, so you can weigh the investment against it. Ask for a fixed price for the first step and a priced plan for what comes next.
Do we need our own data scientist or IT department?
No. You do need someone who knows the process and can make decisions. The implementation partner handles the technology and makes sure your team knows how to work with the new setup.
Which process do you start with?
With the process where the most value lies and that is feasible with the systems you have. That isn't always the biggest AI idea. Sometimes it's work that happens every day, such as retyping data between systems, sorting email or checks that get done twice.
What if our data isn't in order?
Then that's part of the foundation. You check up front whether the data is usable. A partner tells you what it takes to make the most of an opportunity, and whether it's worth it.
What does an implementation partner need to do about the EU AI Act?
It should be able to explain, for each application, which risk category of the AI Act it falls under and which obligations come with it, such as transparency and human oversight. Include that in the preliminary research, not after the pilot.
What happens if an automation doesn't work as intended?
Agree in advance how you will measure whether an automation does what it should, and what happens if it doesn't. An implementation partner keeps adjusting until it works, or tells you honestly that the process isn't suited to it.
