Most descriptions of a market entry pilot stop at “Discovery, Pilot, Scale.” That schema tells a prospective client nothing about what actually happens, and it tells an internal team even less about what to deliver by when. Here is what the first four weeks look like in practice, with the deliverables and decision gates attached to each one.
Week 1: Buying Committee Mapping, Not Sales Calls
The first week is not spent selling. It is spent building an accurate map of who actually decides. In DACH enterprise procurement, the visible contact rarely holds unilateral authority, and the real gatekeeper is often in Compliance, Datenschutz, or Einkauf rather than the department that initiated contact.
Deliverable: a documented buying committee map with roles, veto power, and budget authority identified by name. Decision gate: if the champion cannot identify who signs, the pilot does not proceed to Week 2. This single gate eliminates a large share of pilots that would otherwise stall silently in month three.
Week 2: Compliance Gate Identification
Every sector has its own procurement gate, and it is rarely stated upfront. For SaaS and AI vendors, this typically means EU AI Act risk classification, DSGVO data residency questions, or sector-specific certifications (BaFin for financial services, BFSG/EAA for consumer-facing digital products). Surfacing this in Week 2, rather than discovering it during contract review in Week 10, is what separates a structured pilot from an improvised one.
Deliverable: a compliance gate checklist specific to the prospect’s sector and system classification. Decision gate: any gate requiring more than 30 days to satisfy is flagged to the client before deployment begins, not after.
Week 3: Success Metrics, Defined Before Deployment
By Week 3, deployment scope and the specific metrics that will determine pilot success are fixed in writing and signed off by both sides. This is the single most commonly skipped step in improvised pilots, and it is also the reason so many pilots drift into an undefined “let’s see how it goes” state that neither side can exit cleanly.
Deliverable: a one-page success criteria document with quantified thresholds. Decision gate: no deployment begins without signed-off criteria. This protects both the client, who gets a clear evaluation basis, and the vendor, who gets a defined off-ramp rather than an open-ended obligation.
Week 4: Controlled Deployment Begins
Only now does technical or operational deployment start, against the scope and metrics fixed in Week 3. Because the first three weeks did the structural work, deployment itself is narrower and faster than in pilots that begin with technology and work backward toward alignment.
Deliverable: deployment against fixed scope, with a documented weekly check-in cadence for the remaining pilot period. Decision gate: the check-in structure feeds directly into the Day 60 controlled deployment review and the Day 90 reference decision (see our earlier piece on the 90-Day Pilot Model).
Why the Sequence Matters More Than the Steps
None of these four weeks is complicated in isolation. What makes the difference is the sequence, and the discipline of not skipping ahead. Enterprise buyers in DACH are not skeptical of pilots. They are skeptical of ambiguity. A pilot that arrives with a buying committee map, a compliance gate checklist, and signed success criteria reads as institutionally serious before a single feature has been demonstrated. A pilot that arrives with a demo and a hope reads as exactly that.
This is also where most international SaaS and AI vendors lose time they don’t realize they’re losing. The instinct is to move fast into deployment, because that is what the US or UK playbook rewards. In DACH, the four weeks spent on structure before deployment are not overhead. They are the mechanism that gets the pilot to a genuine reference decision instead of a quiet, undocumented lapse.
For the deliverables and metrics we anchor to structure the 90-day evaluation, and how the reference decision itself gets made, see our earlier piece on the full model.
