→ AI readiness assessment

AI readiness assessment for companies deciding whether to build at all

Two weeks, a fixed fee, and a written answer to the only question that matters first: is there anything here worth building, and what would it actually cost you.

Request an assessment→All consulting options

Overview

How we approach this work.

An AI readiness assessment tells you whether your data, your processes and your systems can carry the AI project somebody has proposed. We look at what you have rather than at what a vendor deck assumes you have, and we write down what we find.

Most AI projects that fail did not fail during engineering. They failed because nobody checked whether the data existed in usable form, whether anyone could define a correct answer, or whether the process being automated was stable enough to automate. Those are two week questions, not six month questions, and they are much cheaper to ask first.

This is the assessment engagement. If you already know what you want built, the service pages cover that work directly and you do not need this.

What you get

Concrete deliverables.

Data inventory, against reality

What data you hold, where it lives, how clean it is, and which of it is actually usable for the use cases on the table.

Use case shortlist with expected return

Every candidate scored on value, feasibility and effort. Usually two survive, and one of those is not the one people expected.

Build or buy, per use case

Where an existing product covers it, we say so. Building something a vendor already sells for a few hundred a month is a poor use of your budget.

GDPR and EU AI Act position

Where your data would be processed, under what agreement, and which obligations attach to the use case rather than to the hosting choice.

Architecture recommendation

Provider API, EU region endpoint, private VPC, or open weights on your own hardware. With the cost difference attached to each.

A costed first build

One proposal, scoped and priced, for the use case that survived. Including the option of not doing it.

How we do it

The process, in four phases.

Week one, what you have

Sessions with the people who run the processes, a look at the real data rather than its documentation, and a review of the systems anything would have to integrate with.

Week one, what is possible

We score the candidate use cases against feasibility and value, and discard the ones that do not survive contact with your data.

Week two, the written answer

The assessment document: findings, shortlist, compliance position, architecture recommendation and a costed first build.

Presentation and decision

Ninety minutes with whoever is deciding. You keep the document regardless of what you decide next.

Scope

What we look at.

What we assess against

Anthropic Claude, OpenAI, Google Gemini, open weights on vLLM, AWS Bedrock, Azure OpenAI, Vertex AI, and the EU region options for each.

Data and retrieval

Your existing databases, document stores, CRM and ERP records, pgvector, Qdrant, Azure AI Search, and whether hybrid retrieval would actually help you.

Compliance framing

GDPR processing basis, data residency, EU AI Act risk classification for the proposed use case, and what documentation each obligation implies.

Outcomes

What you walk away with.

A written assessment that a board can read and act on, not a slide deck of possibilities.
A shortlist of use cases with the reasons the others were dropped, which is often the more useful half.
A costed first build, so the next decision has a number attached to it rather than a range.

References with names available after a first call.

In depth

What it costs, and how the fee comes back

€9,500 to €14,000 for the two week assessment, fixed. The number does not move because the scope does not move.

The same commercial mechanic applies here as on the consulting side of Resser Solutions.

The full fee comes off your first invoice if we go on to build what the assessment recommends.
Valid for six months from the day you receive the document.
Applies to projects above €50,000, one time, not a running discount.
If you take the assessment to another supplier or build it in house, it is yours and you owe us nothing. That happens, and it is a fair outcome.

Why an assessment is cheaper than a pilot

The common alternative is to fund a small pilot and see what happens. It sounds prudent and it usually costs more.

A pilot answers whether a model can produce plausible output on a sample, which is almost always yes and almost never the question. It does not tell you whether the data holds up at volume, whether the process is stable enough to automate, whether anyone can define correct, or what the production version would cost to run. Those are the things that kill projects at month five, and an assessment surfaces them in week one.

If the honest answer is that you should run a pilot, the assessment will say so and scope it. We are not against pilots. We are against pilots funded before anyone checked whether the data exists.

When you do not need this

Three cases where we will say so on the first call, and save you the fee.

You already know the use case, the data is clean, and someone internally can define a correct output. Go straight to a build.
The task is deterministic. A rule or a query already produces the right answer, and a model would add cost, latency and doubt.
The budget for the whole initiative is under €25,000. Spending half of it on assessing the other half is not a sensible split.

Related services

Other places this work shows up.

AI inside your CRM and ERP Private and on-premise AI infrastructure All AI consulting options

FAQ

Frequently asked.

How much does an AI readiness assessment cost?

€9,500 to €14,000 for two weeks, fixed. The fee comes off your first invoice if we go on to build what it recommends, on projects above €50,000, valid for six months.

What do we get at the end?

A written document: data inventory, shortlisted use cases with the reasoning for each, the ones we dropped and why, your GDPR and EU AI Act position, an architecture recommendation with costs, and one costed first build. Plus ninety minutes presenting it to whoever decides.

Can you just tell us if AI is worth it for us?

That is the question this answers, and sometimes the answer is no. We have written that report. A supplier who has never delivered a negative assessment is not assessing anything.

Do we need clean data before we start?

No. Working out how clean the data actually is forms part of the assessment, and it is frequently the most valuable finding in the document.

How is this different from a proof of concept?

A proof of concept shows a model producing output on a sample. An assessment tells you whether the thing survives production: data at volume, integration, cost, compliance and whether anyone can define a correct answer. A PoC is often the right second step, and the assessment scopes it when it is.

Do you sign an NDA?

Yes, as standard, before anything is shared. We work under EU law and your data stays in the EU throughout.

Not sure it is worth building? Ask us first.

Tell us what you are considering. We reply within one business day with whether an assessment makes sense for you, which one, and what it would cost.

Request an assessment→