.svg)
AI consulting that adds up
Before anyone builds anything, it should be on the table where artificial intelligence pays off in your company and where it does not. You get both in writing.
- AI consulting
- AI consulting establishes, before anything is built, which processes pay off with AI and which do not. What you get is an assessment with an order of work, effort and risks, not a finished system.
The question is rarely whether AI works.
Technically almost anything can be built today. The expensive question is which of your processes earns the effort back.
In the conversations we have, three to five ideas are usually on the table. Two of them carry, the rest sound good and later fail on the data situation or on the fact that nobody in the building wants to own the running system.
Our assessment separates those beforehand. It costs nothing, because we have no interest in selling a project that nobody uses in its second year.
Would this pay off for you?
Two questions decide it almost every time: how often the process runs, and how good the data behind it is. That gives four fields.
Start here
The standard case for a first agent. A process that runs daily, with data already sitting cleanly in a system. The benefit shows within the first quarter.
Clean up first
Worth doing, but not yet. The data has to hold up first, otherwise you automate an error at scale. That groundwork goes into the plan.
Keep it small
A process that runs ten times a year rarely carries a system of its own. Often a surface automation or a better form is enough.
Leave it
This is where we say no. It happens more often than a provider on a day rate would tell you, and it saves both sides a year.
What the assessment contains
Not a deck about trends. A document you can make a decision with.
Process recording
We look at how a process actually runs, not how the manual describes it. That means sitting with the people who do it.
Data review
Which systems hold the information, how clean is it, and can it be reached technically. This decides success more often than the choice of model.
Benefit calculation
For each candidate: time saved, errors avoided, effort to build and to run. With the assumptions stated openly.
Prioritisation
An order with reasons. What first, what later, what not at all. The first step is deliberately kept small.
Regulatory classification
Which risk class under the EU AI Act, which obligations follow, and what that means for documentation.
Fixed price
For the recommended first step we name a price before you decide. No renegotiation halfway through.
Five stops in two weeks
The assessment is deliberately short. Anyone who needs four weeks for an opinion is selling occupancy.
All results collect in your client portal, including appointments and documents.
First call
Thirty minutes. What should improve, what is in the way.
Process recording
Half a day at your premises, with the people in the process.
Data review
We check systems, interfaces and quality.
Assessment
Benefit, effort and risk per candidate, in writing.
Handback
A two hour presentation, then the document is yours.
Consulting, agency or both
The terms get used loosely in this market. Here is how we mean them.
| Criterion | Classic consultancy | Pure delivery | Kuroko Labs |
|---|---|---|---|
| Result | Recommendation and slides | A system built to spec | Concept plus a running system |
| Who carries the risk | The client | Shared, depending on the spec | We do, through the fixed price |
| Billing | Day rate | Time and materials or fixed | Fixed price based on the concept |
| After go live | A new engagement | Maintenance optional | Support including portal |
| An honest no | Rarely in their interest | Not foreseen | Part of the assessment |
Scroll sideways for all columns →
What consulting costs
How to spot a usable assessment
Four points worth checking with any provider. Including us.
- Does it contain a no? An assessment without discarded ideas is a proposal wearing the costume of an assessment.
- Are the assumptions visible? Every benefit calculation rests on estimates. They belong in the document so you can check them.
- Was anything measured? Without a baseline there is no way to prove success later. Handling time and error rate belong before the start.
- Who does the building? Ask to meet the person who will work on the project.
AI consulting Munich: common questions
Related topics
Tell us about one of your processes.
Thirty minutes, free of charge. Afterwards you will know whether an assessment is worth it, and we say no when it is not.