For anyone unsure where to start
AI only pays off once you know how the work really flows
Discovery is six weeks of looking over the shoulder of the people doing the work: interviews, shadow days and a workflow analysis. It produces a validated route to build, with the places where time leaks away set down in black and white.
Book a conversationWhy look first
AI doesn't work in a process nobody knows
The tools are rarely the problem. What it runs aground on is that nobody knows exactly how the work flows: which steps are actually taken, where things sit waiting, and which detours people have built for themselves over the years. Automate a process you only know from paper and you automate the paper version. So we look first, at the people doing the work and not only at those who describe it.
What we find
Four things that only surface when you look
We meet these four in almost every department, regardless of sector. None of them appear in any process document, and they are exactly where the time goes.
The documented process isn't the process
Nearly every organisation has a description of how the work is supposed to go. That description is usually a few years old and no longer accurate in several places. Build on it and you are building for a situation that does not exist.
The workarounds are the real process
Where the system does not cooperate, people invent something. Their own spreadsheet, a message to a colleague, a folder on a local drive, a step they skip because nobody checks it anyway. Those detours are not sloppiness, they are information: they point at where the design does not fit the work.
Time leaks between the steps, not inside them
Ask people where their time goes and they name the tasks. Watch them work and you see the waiting: for an approval, for a file, for someone to call back. That waiting appears in no time sheet and is often the largest part of the lead time.
Nobody sees the whole chain
Everyone knows their own part well and the parts before and after it only halfway. It is exactly at those handovers that information gets lost and work gets retyped. It is also where AI pays off most, and where nobody points on their own, because it belongs to nobody.
The cycle
The discovery cycle, week by week.
Interviews
We talk to the people doing the work and to those describing it.
Shadow days
We walk along on the work floor and see where the work really gets stuck.
Workflow journey
Every step, every handover, every workaround drawn out, so the friction becomes visible.
Impact / effort
Which use cases really earn themselves back, and which are just a good story.
Pilot plan
One validated route, ready to build or hand over to whoever builds.
Practical
Duration, investment, result.
6 weeks per department
From signal to validated pilot plan.
From €35,000
Per department, excluding VAT.
Discovery report
An impact/effort matrix and one validated route to build.
When discovery isn't the right instrument
- Do you already know exactly which application you want to build, and why? Then discovery is a detour. Skip it and start building.
- Is there no mandate to actually redesign the work? Then discovery produces a report nobody acts on. Sort out the mandate first.
- If it turns out along the way that a process problem is just a process problem, we say so. Not every question is an AI question, and that is a usable outcome too.
Our expectations were high, and they were exceeded. A fun, accessible and highly effective way to get to know the possibilities of AI.
At Veris, redesign after discovery returned up to 20% of the work week.
Frequently asked questions
A few things upfront.
01 What does a discovery cycle cost? +
From €35,000 per department, excluding VAT. Every proposal splits the investment per building block.
02 And if no AI opportunity comes out? +
Then we say so. Discovery regularly finds that a process problem is just a process problem; that too is a valuable outcome.
03 What happens after the pilot plan? +
Building: ourselves as an agent or PoC, or via a clean handover to a development partner, with knowledge and quality secured.
How much of our people's time does this take? +
About an hour per interview, and on a shadow day we walk along without anyone having to do anything other than their job. That last part is deliberate: the moment people start explaining their work instead of doing it, the very detours we are looking for disappear.
Why not just a workshop with the management team? +
Because the management team knows the process as it was intended. That is valuable information, but it is not where time leaks away. A workshop produces a list of ideas; watching produces a list of places where the work gets stuck. That second list is shorter and more usable.
Which department should you start with? +
Usually the department where the most handovers come together, because that is where the friction concentrates. We look at that together in the introductory call; it is often clear within twenty minutes.
Where does your time leak away?
Book a conversation; we look together at which department lends itself first to discovery.