First find out where AI actually saves you time and money.
I map your processes, check every idea for effort and payoff and do the maths on what it delivers. What you end up with is a roadmap: order, cost and a realistic time saving. Including the cases where the honest answer is that an idea isn't worth it.
We walk through your processes and collect everything that sounds like AI. Every idea gets numbers attached: effort, monthly payoff, risk. Whatever doesn't pay for itself in a foreseeable time goes on the waiting list on purpose. What usually remains is three projects instead of fourteen building sites — and you know which one to start with.
Before budget goes into AI, someone should have done the maths.
The most expensive AI projects are the ones built without anyone asking first what they actually save. An analysis costs a fraction of the build and decides whether the budget goes to the right place. In most companies the biggest lever sits somewhere other than you'd expect.
Plenty of AI ideas are floating around the team, but nobody knows which one comes first.
A vendor has made you an offer and you want to know whether price and payoff add up.
A first attempt with AI fizzled out and you want to know why.
There is budget for digitalisation, but no plan for the order it gets spent in.
What we look at together.
TOPICSWhere does the time go?
We walk through your processes step by step and mark what eats time, gets done twice or regularly sits untouched.
Where do we start?
Every idea gets effort, payoff and risk as a number. Out of that comes an order that adds up instead of just sounding good.
Is it technically possible?
I check whether your systems have interfaces, whether the data is good enough and where a human still has to decide today.
Which data may go where?
We settle early which data is needed at all, which providers qualify and what should run on your own server.
Off-the-shelf tool or custom build?
A comparison with numbers: licence cost over three years against a one-off build, including what each option ties you to.
Who is going to use it?
The best solution is worthless if nobody uses it. We settle ownership and what your team needs to get going.
Your question isn't listed? Usually the first call already settles whether you need an analysis at all. Tell me briefly what it's about and you'll get an honest read.
What an analysis saves before the project.
THE NUMBERSThe analysis doesn't pay for itself on its own, but through the project it picks correctly. The most expensive mistake is a cleanly built system for a process that was never the bottleneck.
The same start, with and without a plan
| STEP | NO ANALYSIS | WITH A ROADMAP |
|---|---|---|
| Selection | The loudest idea wins | The one with the biggest lever |
| Budget | Estimated in conversation | A number per project, upfront |
| Technology | Turns up while building | Checked for feasibility first |
| Data protection | Question comes at the end | Settled before the first line |
| Team | Gets informed | Planned in from day one |
| Result | A project with an open ending | An order backed by numbers |
How an AI analysis works with me.
HOW IT RUNSFirst call
30 minutes, free. You tell me where it hurts, I tell you whether an analysis is needed at all or whether we can start building straight away.
Process mapping
Two to four hours on a video call, ideally with the people who do the work every day. A screen share is enough, no preparation needed.
Review and calculate
I check every idea for feasibility, data and privacy and weigh effort against payoff. That takes me a few days without you.
Hand over the roadmap
You get a document with order, cost and expected time saved, plus an hour in which we go through it together.
An assessment from someone who also builds it.
Consulting isn't a sideline for me: I develop every day with the exact tools we're talking about. So you don't get slides full of buzzwords, you get a read on what works reliably today, what only works in the demo and what it costs to run.
How you can tell it's worth something
IN PRACTICE“Chris was professional, reliable and committed. A good eye for detail and design – I can recommend working with him to anyone.”
The biggest idea usually wasn't the first one
In most conversations a chatbot sits at the top of the list. After the mapping, the bigger lever is almost always upstream of it: the enquiry that gets retyped, or the quote that sits around for three days.
Go through your own list →The price is set before I start working.
From the workshop to the finished roadmap.
The roadmap is yours, even if someone else builds it.
I earn nothing on licences, so I don't push any.
What does consulting cost?
Fixed price, agreed upfront.
ALL PRICES EXCL. VATAI office hour
from €490Two hours for one specific question that's already on the table.
- Two hours on a video call
- An assessment of one concrete idea
- Feasibility and rough cost
- A recommendation on how to proceed
- Short written summary afterwards
AI roadmap
from €1,490The full analysis for one area or a whole company.
- Process mapping with your team
- Every idea reviewed and calculated
- Prioritised order with costs
- Feasibility, data and GDPR checked
- Roadmap as a document plus handover session
Sparring
from €290/monthFor everyone who needs a second opinion on a regular basis.
- A monthly session for open questions
- An assessment of offers and tools
- Roadmap kept up to date
- Direct contact via chat and phone
- Cancel monthly
The range depends on how many areas we look at and how many people are involved in the processes. If you decide to build with me afterwards, the analysis is credited against the project in full.
When you can save yourself the consulting.
Consulting isn't an end in itself. In these four cases I'll say in the first call that an analysis won't get you anywhere — and what makes sense instead.
You already know what you want
If the process is clear and only needs building, you don't need an analysis, you need a slot for the build. That costs less and is faster.
There is only one process
For a single process the free first call is enough to judge effort and payoff. An analysis only pays off once several areas compete for the budget.
The real problem isn't an AI problem
Missing ownership, messy master data or a process everyone runs differently: technology doesn't fix that. I'll say so plainly before any budget moves.
Nobody is there to implement it
A roadmap without anyone to work through it is an expensive PDF. If nobody internally has time and nothing gets commissioned externally, we'd rather wait for the right moment.
AI consulting for small and mid-sized companies
By Chris Pump, developer since 2021Last updated: August 2026
AI consulting has a questionable reputation, and deservedly so. Too many projects end with a presentation, a maturity matrix and an invoice, but without a single hour of work removed. The job is actually simple: work out which recurring work can be handed over at a reasonable cost, and put that in an order. That is what strategy means here.
What an AI strategy really is for a smaller company
Not a vision for the next five years, but a list with an order. Which processes cost the most time, which of them follow clear rules, which can even be touched with the systems you have? Out of that comes a ranking that tells you what gets built first and what deliberately waits. A horizon of six to twelve months makes sense. Anything beyond that is guesswork at the speed the models are changing.
Where nearly every company should start
In practice the first lever is rarely in customer contact, it's in the back office: capturing enquiries, writing quotes, filing receipts, moving data between systems. Those steps repeat, follow rules and add up to more time than any outward-facing project. They also make a better start, because a mistake shows up internally instead of at the customer. Visible applications like chat or phone come later, once the team trusts the technology.
Data protection belongs at the start
The question of which data may pass through which system decides the architecture, not the finishing touches. That's why it sits near the front of the analysis. Often far less data is needed than expected, and where things get sensitive, European providers, a self-hosted server or locally running models come into play. Processing agreements, deletion periods and ownership go into the roadmap in writing, so nobody has to guess later.
Why Kasoria
At Kasoria you're advised by the person who also builds the implementation. That means no recommendation that falls apart on contact with reality, and no translation loop between consulting and development. I don't sell licences and take no commission from any vendor, so the answer is allowed to be that an off-the-shelf tool for €30 a month will do. And if a project doesn't add up, I say so before you commission it.
Common questions about AI consulting.
How long does an AI analysis take?
From the process mapping to the finished roadmap it's usually five working days. Your own time investment is two to four hours of workshop plus an hour for the handover.
What do I get at the end?
A document with every process we mapped, every idea reviewed, the recommended order, expected costs and estimated time saved. Plus the notes on feasibility and data protection for each item.
Do I have to build with you afterwards?
No. The roadmap is yours and is deliberately written so that someone else can implement it. If you do choose me, the analysis is credited against the project in full.
How do I prepare?
You don't. The people who do the work every day matter more than any documents. If you want to bring something, bring examples: a typical enquiry, a typical quote, a typical week.
Is it worth it with five employees?
If several areas are competing for attention, yes. If it's about a single process, the free first call is enough and we save ourselves the analysis.
Do you also advise on tools you don't build yourself?
Yes, and that's often the outcome. If an off-the-shelf product covers your case, it's almost always cheaper than a custom build. I get no commission for it, which is exactly why I can recommend it.
Can you review someone else's proposal?
Yes. Send me the proposal and the description and you'll get an assessment of price, effort and the points that tend to get expensive later. The AI office hour is usually enough for that.
Might the result be that AI isn't worth it?
It happens, usually for individual ideas and rarely for all of them. If a process occurs too rarely or needs a real decision every time, the roadmap says exactly that.
Roughly what does the implementation cost afterwards?
The roadmap gives a range per item. Single automations usually land in the high three to low four figures, agents and custom software above that. The number becomes binding with the proposal for the specific project.
Do you work with funding programmes?
I don't file funding applications and I'm not a listed consultant. You can, however, use the documents from the analysis — description, effort and expected benefit — for an application.
Does this happen on site?
Normally not, I work remotely. A video call with a screen share is enough for the mapping, and several short sessions are usually more productive than one long day in a meeting room.
What happens to what I tell you?
It stays between us. On request I'll sign an NDA upfront; otherwise it applies tacitly: no numbers, no processes and no name of yours appear anywhere unless you explicitly allow it.
Tell me what's on your list and I'll tell you what actually adds up.
30 minutes, free and without obligation. Afterwards you'll know whether you need an analysis or can start right away.
Usually answered the same working day.