Five design invariants — pre-meeting check (v0.4 §1.3)
Each invariant must hold before the M0 conversation. Mark [x] only when verifiable, [!] if there's a known gap to surface in the meeting.
○Real authority, fractional presence — Mandate scope (Q1) drafted; authority asks (Q2) listed; pending charter signature at M0.
- Why it matters: Without explicit board mandate + finance access + portfolio decision rights, Partner becomes another expensive advisor. Verify before committing.
○System-setter, not daily decider — Partner role = Michal as Interim CAIO + VV delivery. Standing recusal on commercial framing.
- Why it matters: Partner designs forums + mechanisms others operate. If Partner is making daily calls, the engagement won't exit cleanly — successor will inherit Partner's decision queue.
✓Successor-building from Day 1 — Internal AI Lead (Kateřina) named, internal, named+active. Authority Erosion × Elevation counter-mechanisms required throughout (visibility-elevation rituals; attribution discipline; recusal on Lead's domain decisions).
- Why it matters: Internal AI Lead (Kateřina) is the engagement's exit vehicle. Every week of weak mentorship buildup pushes exit readiness back. R4 — engagement is for the successor, not the Partner.
○Exit criteria agreed upfront — 8 hard + 5 soft criteria drafted (Q6); pending sign-off at M0.
- Why it matters: Q6 hard + soft criteria signed at M0, non-negotiable thereafter. Without upfront agreement, "done" becomes a renegotiation at M5 — Partner loses leverage.
○Clean commercial interface — 3-layer fee structure drafted (Q7); recusal SLA 5 BD; CFO Bornemann named as commercial counterpart when Partner recused. Pending charter signature.
- Why it matters: Recusal protocol (L3-321) + 3-tier fee structure (T1/T2/T3) keep Partner out of VV scope decisions. Without it, every VV scope expansion looks self-dealing.
1. Engagement context
Archetype: Elevation — Amplify an existing internal AI lead — capability exists but lacks authority, scope, or executional discipline. Partner extends, never replaces.
If unresolved, currently-active obstacles cumulatively introduce 34–84 weeks of delay. Illustrative € exposure: €1.5M–€4.8M across 6 measured obstacles.
5 obstacles carry CRITICAL or LARGE engagement-value-at-risk.
Note: 7 of 8 active obstacles have impact data; 1 are not yet quantified.
2. People — named, inactive, unknown
Already engaged (2):
Miguel Martins da Silva (CSCO) — Already AI-active (VusionGroup, Geek+, Rinkai, dark warehouse).
Kateřina Lesch (HeadOfAI) — Newly appointed Head of AI at Dr. Max (internal). Was the engagement introducer in her former Deloitte / EmbedIT role; now holds the seat we needed her to influence. Reports to CIO (likely; verify with Kateřina). Methodology implication: O-031 External Champion Conflict no longer applies.
Named but not yet engaged (6):
Penta Operating Partner (PEOperatingPartner) — UNKNOWN — Penta operating partner not identified; open question for validation.
Leonardo Ferrandino (CEO) — CEO since 2016, target sponsor.
Thomas Bornemann (CFO) — Capex sign-off, Heavy Gate 2 counterparty.
Luca Demarchi (CIO) — Per 2024 sustainability report; tenure UNCERTAIN — possible move to Bata Group.
Why: When 5+ HIGH-severity obstacles fire concurrently, the engagement is structurally fragile. Consider whether this is the right week to commit to M0, or whether to invest 1–2 weeks in pre-mitigation.
Evidence: 7 obstacles with severity=high are currently active
2. Pre-stage 30-min CEO briefing before M0
When: Before M0
Who: AR
Why: Sponsor is not actively driving (Board state = Aware). M0 will not get the air cover it needs unless the CEO arrives pre-aligned on what is at stake.
Evidence: L0.board_state = Aware; O-013 (Sponsor Escalation Stalled) is active
Reduces / resolves: O-013
3. Stand up L3-329 (AI conformity assessment process) within 30 days
When: Next 30 days
Who: CIO + GC
Why: EU AI Act pressure is Months_3_to_6 and the conformity assessment process is not yet stood up. The legal exposure window is closing; this is a hard deadline, not a soft target.
Evidence: L0.ai_act_deadline_pressure = Months_3_to_6; L3-329 state = not_stood_up; O-022 (AI Act Compliance Stalled) is active
Reduces / resolves: O-022
4. Negotiate 3 authorities during M0
When: Before M0
Who: AR + Sponsor
Why: 3 L1 authorities are currently Out and are gating active obstacles. These are the actual horse-trade items for the M0 conversation.
Evidence: Authorities to negotiate: L1-108, L1-129, L1-113
Why: Multi-market engagement with no cross-market sync mechanism. Markets will diverge (different vendors, different decision cadences) within the first quarter unless the sync is stood up early.
Evidence: L0.multi_market = true; L3-326 state = not_stood_up; O-014 (Multi-Market Parity Drift) is active
Reduces / resolves: O-014
6. Confirm Luca Demarchi still in seat (or identify successor)
When: This week
Who: AR
Why: CIO is the structural reporting line for the Internal AI Lead and the convening counterparty for Track 3. If the seat changes during the engagement, the AI Council charter has to re-anchor.
Evidence: people.CIO notes flag tenure uncertainty: "Per 2024 sustainability report; tenure UNCERTAIN — possible move to Bata Group.…"
4. Authority asks — ordered for the M0 conversation
Methodology says 5 authorities should be Must for an Elevation engagement. Currently all are Out. Order suggested by category for the conversation:
Information (1)
L1-113 — Right to interview anyone (incl. cross-market)
- What: Authority to interview anyone in the organization (cross-market, cross-function) without permission gating.
- Why ask: Engagement diagnosis requires direct access to people — frontline, Initiative Leads, sponsors, peers. Permission-gated access slows everything to a crawl.
Hiring (2)
L1-116 — Co-design Internal AI Lead role
- What: Authority to co-design the Internal AI Lead role — JD, scope, reporting line, target compensation.
- Why ask: Successor-building from Day 1 (design invariant #3) requires the Lead's role be designed properly. Co-design ensures it isn't just bolted onto an existing role.
L1-117 — Veto on succession candidate
- What: Authority to veto succession candidates for the Internal AI Lead seat.
- Why ask: Wrong-fit Lead is a Realignment situation. Partner needs the ability to block a bad hire before it happens.
Communication (2)
L1-129 — Direct board access on AI
- What: Direct authority to communicate with the board on AI matters (no IT/CIO gating).
- Why ask: Board literacy + sponsor coaching require direct access. Mediated communication dilutes the message and creates gatekeeper dynamics.
L1-130 — Speak to regulators on AI
- What: Authority to speak to regulators on AI matters (or accompany Legal/CRO).
- Why ask: Regulator engagement is now AI-specific (EU AI Act, sectoral). Partner often has the deepest AI context; regulator wants to talk to who actually knows.
5. Active obstacles to surface in M0
O-001 — Pipeline Stall at Gate 2 [severity: high]
Initiatives accumulate at Pilot-Validated; Scale decisions don't happen
Triggered conditions:
L1-108 is absent_or_denied
peer.CFO is missing_or_inactive
L3-303 is not_stood_up
peer.CIO is missing_or_inactive
Recovery actions:
Negotiate L1-108 mid-engagement
Escalate to L3-318 (Exec sponsor)
Document scoping limitation if irrecoverable
O-003 — Risk Function Disengaged [severity: high]
Regulated profile but CRO/DPO not integrated; risk classification skipped
Triggered conditions:
peer.CRO is missing_or_inactive
peer.QMHead is missing_or_inactive
peer.DPO is missing_or_inactive
Recovery actions:
Force trigger via L1-124 (veto) on next initiative requiring classification