Data and AI readiness assessment

An AI roadmap that can say "not yet"

In two to three weeks, the engineers who would build it tell you which of your AI ideas your data and your AWS estate can carry today, what has to be fixed first, what it costs to run, and in what order.

Built for teams already running on AWS

The assessment is for a CTO, VP Engineering or founder who owns the decision and can act on a roadmap in the next 6 to 12 months. Three situations we see most often:

The pilot works in the sandbox

Your Bedrock or SageMaker pilot works on the demo data. The board is asking when it ships. Nobody has written down what stands between the two.

The data lives in six places

The use cases are clear. The data behind them sits in separate systems with no catalog and no owner, and every AI plan stalls on that layer.

Governance is blocking scale

PHI, GDPR data or customer contracts mean you cannot hand an AI system data you cannot govern.

Not the right fit

If your production workloads are not on AWS yet, this is a different conversation, and we will say so on the call.

If nobody on your side can act on the roadmap in the next 6 to 12 months, it is too early.

What you get

Five deliverables, written for the people who will decide, then presented live to your sponsor.

  1. Pattern discovery

    Which of your AI ideas map to which data and AI patterns, and where you stand on each: current state against target state, usually for two to four patterns.

  2. Data and AI readiness scorecard

    Data availability, quality, governance, architecture, skills and organisation, each scored out of 4, with one overall verdict: proceed, proceed with conditions, or not yet.

  3. AWS architecture recommendation

    A service-level design per pattern, with the controls your data needs (PrivateLink, KMS, IAM boundaries) designed in rather than added later. Validated with your engineers before it goes in the report.

  4. 12-month roadmap

    Quick wins in months 0 to 3, foundation in months 3 to 6, advanced work in months 6 to 12, with a go/no-go gate between each phase and the dependencies made explicit.

  5. Business case

    Cost of the status quo, AWS run cost for the recommended architecture, payback period and a multi-year view. Numbers your finance team can check.

READINESS SCORECARDKnowledge assistant over internal documents1234Data availability3/4Data quality3/4Data currency2/4AI-specific preparation1/4Architecture readiness3/4Governance and security2/4Skills and operating model2/4Organisational readiness4/4Overall2.5 / 4Proceed with conditions
Illustrative scorecard for one use case. AI-specific preparation at 1 of 4 (no curated knowledge base, no evaluation set) is what turns a yes into a proceed with conditions.

Three weeks, about five sessions, one readout

Read-only access to your estate. No production credentials, no agents installed, no change to anything.

Week 1

Discovery

Interviews with your sponsor, platform lead and data owner. A read-only look at the AWS estate and the data landscape. We write down what the status quo costs you in your own numbers.

Output

Prioritised list of use cases or infrastructure priorities, with the patterns they map to

Week 2

Patterns and architecture

Gap analysis per pattern, target architecture, sequencing. One working session with your engineers to check the design against what they know about the systems.

Output

Scorecard and architecture recommendation, validated by your team

Week 3

Business case and readout

Run cost, payback, the roadmap with its gates. Then a live readout to your sponsor, where the report is presented and questioned, not emailed.

Output

The full report and a roadmap you can start on the following week

What we need from you

  • About five sessions of 60 to 120 minutes over the three weeks
  • Read-only access to the AWS estate and architecture documentation
  • The executive sponsor present at the readout

What happens after

The roadmap is yours. If you build the first quick win with us, that is a separate piece of work, priced from the roadmap, not bundled into it. If your team builds it, the architecture is specific enough to hand over.

Managed operations can appear as an option in the third phase of the roadmap when a governed AWS foundation is what the plan needs. It is an option on a page, never the ask on the call.

AWS Partner AI Services Competency badge

Who does the work

The same engineers who write the assessment build the systems. AWS reviewed the AI systems we delivered for the AI Services Competency, with the case studies public and the customers on the record.

Frequently asked questions

What people ask before booking the first call

It is a fixed fee for the fixed scope above: two to three weeks of senior engineering time, five deliverables, one readout. We confirm the fee in writing on the first call, before any work starts. Anything outside the scope is a separate, separately priced piece of work, and we say so up front.

Want a straight answer on your AI plans?

30 minutes with Cosmin to see whether the assessment fits your situation. If it does not, you will hear that too.