feliXart

Service

AI Readiness Assessment

Setting out where AI belongs in your company — and where it does not.

Most AI projects stall in the pilot because they started on the wrong process. The purpose of this work is not to build a model but to measure which process is genuinely suitable, whether the data you hold is sufficient, and what it will cost per month. The report states what you should not attempt as clearly as what you should, and in most companies the second part is the more valuable one. Scope and duration follow the size of the job. You are left with a document that holds value whether or not you continue with us.

What it covers06

Process selection

Candidate processes scored on frequency, cost of error and measurability, with the chosen one justified.

Data sufficiency

Where the data sits, whether it is reachable, whether it is clean, whether it is enough to ground on — and what it would take to make it enough.

Cost and architecture estimate

Provider or self-hosted; latency, unit cost and a monthly usage estimate.

Risk and compliance

Where data travels, KVKK and GDPR scope, policy-layer requirements and the audit record you will need.

Measurement plan

How success will be measured: the evaluation set, the acceptance threshold and the condition on which a pilot goes to production.

Phased roadmap

Pilot, production and rollout phases, with dependencies, effort estimates and a fixed-price implementation proposal.

Process04
01

Groundwork

Kick-off, candidate processes listed, existing systems and data sources reviewed.

02

Field

Interviews with the team doing the work by hand today, collecting real examples and edge cases.

03

Trial

A small, measurable trial on the chosen process, so quality and cost are discussed from measurement rather than opinion.

04

Synthesis

Findings presented to management with a measurement plan and a fixed-price implementation proposal.

FAQ05
Why is it paid?
Because it is engineering work, not a demo. Field interviews, data review and a measurable trial take real effort. What you pay for is the report you keep, not a quote.
Can the answer be “do not do this”?
It can, and sometimes it is. On a process that does not repeat, carries a low cost of error or has no usable data, AI produces expense. We say so in the report, with reasons.
Do you need to see our data?
A sample is enough for the assessment; full access is not required. A confidentiality agreement is signed beforehand, and which data is examined for what purpose is put in writing.
Do you decide which model we use?
We decide together, balancing quality, latency, cost and data residency. We are not tied to a single provider, and what we build is designed so the model can be changed.
Is the fee credited against the project?
Where the work proceeds to implementation we credit all or part of the fee against the project price, with the proportion and the window written into the contract.

Looking for where AI belongs in your business?

Tell us about the process — and if it is not a fit, we will say so plainly.