The AI implementation pilot
Two September places · Selected for fit
One problem. A working AI system.
We build and test it with your team, then hand it over.
- Implementation fee
- EUR 0
- Delivery
- Typically five weeks. Scope and timing agreed before kickoff.
- Ownership
- You own the bespoke work and IP
You cover approved external costs. We ask for a case study you approve before publication.
An illustrative pilot · Manufacturing
Can we still ship Friday?
A supplier is two days late. See how a different production sequence can protect the delivery date.
Explore by industry
Materials, production schedules, and delivery commitments.
Interactive example · Fictional data
Friday’s dispatch
Two deliveries slip into next week.
Cutoff · Fri 11 Sep, 15:00
Production line · 8 hours / day
Mounting brackets
1,200 pieces · 12 hSupport plates
800 pieces · 16 hPanel covers
600 pieces · 8 hA’s material arrives Wednesday, 08:00
B and C already have their own material reserved.Keeping A → B → C leaves the line waiting for A’s material. B and C miss Friday.
One line. Three orders. A plan the production team can review.
How we would measure it
Measure how early delivery risks are found, time spent replanning, and whether the approved schedule holds. A first pilot uses an order book, material availability, a production plan, and supplier updates for one line. The numbers here describe a fictional scenario, not customer results.
Build it. Test it together. Make it yours.
Before kickoff
Agree the scope
Define the result using real samples. Agree field meanings, scope, timing, access, costs, and measures.
Typically 3 weeks
Build, test, and train
Build the workflow, test ordinary and exception cases with your users, and train the internal owner.
Then around 2 weeks
Stabilise and hand over
Review live use and fix weak spots. Hand over the tested workflow, instructions, known limits, and a manual fallback.
We build. You bring the context.
Bring one recurring problem, representative examples, and people ready to test. We agree the scope and timing together.
Apply for a pilotWhat does our team need to contribute?
- Decision-maker
- 30 min/week
- Approves the project direction and clears internal blockers.
- Problem owner
- About 2 hrs/week
- Shares real cases and reviews progress.
- Users
- 30–60 min/session
- Test the system and take part in training.
- Systems contact
- 1–2 hrs at setup, then as needed
- Arranges approved access and resolves connection or permission questions.
Estimated involvement. Time varies by role and phase; we agree what’s needed before kickoff. Roles can overlap.
What do we keep?
The bespoke implementation and IP, tested rules, a manual fallback, training, and instructions for your internal owner. We document known limitations. Your team runs the system after handover; further support is optional and agreed separately.
How do costs and the case study work?
You approve external subscriptions, usage, hosting, or other services before we use them, and cover the agreed costs. We agree case-study terms before kickoff. It can name your company or remain anonymous; you approve the evidence and final text before publication.
Is this a fit for our company?
We welcome focused operational problems in manufacturing, finance and accounting, healthcare, staffing, and retail, including a first AI implementation. Company-wide rollouts, core-system replacement, major data cleanup, and autonomous high-risk decisions need a different scope.
Tell us what gets in the way.
Describe one problem and a useful result. We’ll work through the technical details together.
We reply within 24 hours with a discovery invitation or a clear decline.