AI strategy

Sequence first, technology second.

We look at your processes, assess use cases for value and feasibility and determine where you start. With architecture, model choice and the rules of the EU AI Act.

Discuss a strategy sprint
Small, gnarled olive tree on a stepped travertine block; a cable bundle runs down the trunk and the cables plug into sockets in the stone, with moss on the lower step and a silicon leaf in the crown

Use cases by impact

ValueQuotinghigh
EffortQuotingmedium
RiskQuotinglow

Sample view

Outcome

Which use case comes first determines everything else.

Cross-section of olive wood, rough-sawn on the left, sanded on the right with visible grain; a cable bundle held by three champagne clamps runs across the sanded half

Use cases, prioritized

QuotingHigh value
Service knowledgeMedium value

Sample view

01

A roadmap with clear priorities

Use cases ranked by value, feasibility and risk, with a clear first step.

Precision balance in champagne metal on white Calacatta: weights of travertine and Verde marble on the left, a cushion of moss with lichen, an olive branch, a silicon leaf and two green olives on the right; a braided cable plugs into the marble base

Business case

Effort
Value
Assumptions

Sample view

02

A business case that holds up

Effort, value and assumptions are out in the open, so management and IT see the same numbers.

Hexagonal stepped pyramid of travertine, sandstone and Calacatta with a tip of dark Verde marble and champagne edges; moss, gray-green lichen, olive shoots and two green olives grow at the bottom while the tip stays bare; a braided cable plugs into the base, in front of it a sheet with a blind-embossed hexagonal seal

EU AI Act classification

QuotingObligations checked
Service knowledgeObligations checked

Sample view

03

Clarity on the EU AI Act

You know which application triggers which obligations before anything is built.

Building blocks

What a strategy sprint covers.

A fixed scope, three weeks, a roadmap at the end.

  • Assessment

    of processes and systems

  • Evaluation

    of use cases

  • Architecture and model choice

  • Data and interfaces

  • EU AI Act and data protection

  • Roadmap and business case

Five mistakes we watch for in the sprint.

  1. 1

    Pilot in isolation

    runs in one department but isn't connected to any system.

  2. 2

    The data question comes too late

    only after the build does it turn out the data is missing.

  3. 3

    Tool before use case

    the license is bought before anyone knows what it's for.

  4. 4

    Nobody is responsible

    operations have no owner.

  5. 5

    Rules as an afterthought

    EU AI Act and data protection are only checked after the build.

Travertine block, left rough with moss at the front and milled into steps behind; a bundle of cables runs over the edge and plugs into two sockets in the step

Roadmap · Q1–Q4

Q1Orders from the inbox as draft sales orders
Q2Answers from manuals for the service team
Q3Draft quote from the inquiry
Q4Dealer materials from product data

Sample view

A strategy that can actually be built.

We don't write strategies for someone else to implement. What we recommend, we can build ourselves, and we test that during the sprint.

Explore integrations

Start with a strategy sprint.

Three weeks, a clearly defined scope, and a prioritized roadmap with a business case at the end.

Discuss a strategy sprint
More on AI consulting for mid-sized businessesWhat does AI consulting cost?

What's included

  • Assessment of your processes
  • Evaluation of up to 20 use cases for value and feasibility
  • Architecture and model recommendation
  • Roadmap and the top five use cases with a business case