Choose an AI scenario. Before development.

For PMO, IT and engineering leaders whose reporting takes manual effort and who need to choose a useful scenario before choosing a tool.

RUB 120,000 3–5 business days
Discuss an AI discovery assessment
From a process sketch to an AI report, with sources, uncertainty and human review
Demonstration · conceptual interfaces, not client work

A de­ci­sion and a report. No implementation.

What we need

  • A current report and one project
  • Systems, roles and access restrictions
  • An interview with the process owner

What you receive

  • A process map and baseline measurements or clearly labelled estimates
  • 3–5 scenarios with requirements and limitations
  • A pilot recommendation or a reasoned case against using AI

Development, integrations and production operation are not included in discovery. A pilot and implementation are agreed separately.

Demonstration · fictional data

What a discovery report looks like.

A fictional team assembles a weekly status from tasks, a plan and meeting notes. This demonstrates the decision, not a client outcome.

Baseline: no measurement yet. The team’s estimate of preparation time must be checked with a time log, including review and corrections.

Weekly status

Choose for the pilot

Repeated task, reference reports, named reviewer. Compare full preparation and review time.

Document search

Consider later

A usable knowledge base and an access model are needed first.

Meeting actions

Test separately

Review transcription quality and consent before using recordings.

Automatic release decisions

Do not use AI here

A responsible person must resolve conflicting dates and authorise release.

Acceptance: compare the same reference reports, verify material facts, flag missing sources and record the reviewer’s decision. Agree numerical thresholds before the pilot. Implementation is not part of discovery.

Start with a task. Verify the result. Decide what comes next.

  1. 01

    Agree on boundaries

    One process, available materials, constraints and a responsible person on your team.

  2. 02

    Define acceptance

    What would be useful and which examples will demonstrate that it works.

  3. 03

    Deliver the chosen stage

    Discovery or a pilot with a clear scope. Scope changes are agreed separately.

  4. 04

    Hand over the result

    Materials, documentation and a decision: develop, change the scenario or stop.

Your team owns the result.

Before starting, we agree on hosting, access, source materials, documentation and handover. Support, response times and work with new sources are scoped separately.

Before we start

What if AI is not needed?

That is a valid result: we explain why a process change or conventional automation would be more appropriate.

Can we start without access to live systems?

We agree on sample materials and interviews for the initial assessment. We check that they are sufficient before starting.

Do we have to continue?

No. You receive the report and decide independently whether to commission another stage.

Choose your starting point.

Start with one useful result. The next stage is a separate decision.

Discovery and a pilot purchased separately start at RUB 370,000 in total. Discovery is not mandatory before a pilot; there is no automatic continuation or fee credit.

What needs to work better?

Leave a contact and a few lines about the task. We will discuss fit, materials and the next step during business hours.

info@craftvelox.com

An enquiry is not a purchase. We will respond about the service you selected.

Do not include passwords, clients’ personal data or confidential documents.

Email info@craftvelox.com