Marcel Springer · Senior Advisor · Munich
AI strategy and enablement. Transformation and performance programs for boards, executive teams and investors: with tools that stay in-house.
Start the conversationHow it runs inside your organization
AI services
AI Transformation & Enablement
A simple start to working together: four ways in.
Talks
Briefings and orientation for boards, executive teams and committees.
Positioning review
Positioning per function along the AI positioning framework.
Enablement
Training, skill libraries and working standards for day-to-day use.
Build-out
Core workflows rebuilt with GenAI and agentic-AI levers: tools that stay in-house.
Rolled out in mid-sized companies and corporate groups, with a baseline taken before the first implementation. How the positioning is derived and how the work runs: Positioning · Approach.
Enterprise AI platforms, including Anthropic Claude (Enterprise). Advice is vendor-agnostic.
AI positioning framework · Functional level
The vertical axis describes where building capability sits, the horizontal one how deeply the work is integrated. The two are independent and follow the same logic: integration effort and contribution to differentiation rise together. The question is therefore not how far an organization gets overall, but for which functions the higher integration effort creates a competitive advantage.
Integration depth A
Off-the-shelf tools
Integration depth B
Process and system integration
Integration depth C
Your own data
Stage 3
Distributed building
Distributed, on standard tools
Right where task variety is high and no single application carries differentiation.
Integration effort
Distributed, process-integrated
Right when core workflows have many variants and nobody central knows them all.
Integration effort
Distributed, on your own data
Right for core workflows that carry differentiation. Highest effort, lowest substitutability.
Integration effort
Stage 2
Central building
Centrally provided standards
Right when consistency outweighs speed, for instance in external communication and reporting.
Integration effort
Central process integration
Right for regulated or auditable processes where control matters more than speed.
Integration effort
Central data products
Right when a few highly aggregated questions matter on sensitive or tightly regulated data.
Integration effort
Stage 1
Assisted work
Pure use, no coupling
Right at low volume and high variance. The cheapest state that still makes sense.
Integration effort
Integration without building
Right when a standard system already brings the coupling and nobody needs more.
Integration effort
Structurally unstable
Deep coupling without building capability in the function leaves assets nobody uses or maintains.
Integration effort
Positioning is not decided company-wide but per function, say product management, marketing or procurement. Within the function, the design centers on core workflows, which bundle several processes. The sound outcome is a deliberate spread across the matrix: most functions may well stay at low integration depth in plain use, a few with differentiation-critical core workflows earn more. What is expensive is not the wrong position but a missing positioning strategy. Shared components, guardrails and cost rules make these decisions defensible and reversible.
Which functions justify which integration depth?Axis 1 · Distribution of building capability
Three stages, distinguished by where building capability sits. Each stage delivers real value.
Stage 1 · Assisted work
Where capability sits
Nobody
Contribution
Genuine individual leverage in drafting, research, analysis and preparation.
Structural limit
The value stays tied to one person and one session. Nothing accumulates.
Stage 2 · Central building
Where capability sits
One central specialist team
Contribution
Governed, production-grade automation and the first assets worth reusing.
Structural limit
The constraint is throughput, not capability. Small requests never clear the backlog.
Stage 3 · Distributed building
Where capability sits
A minority in every function
Contribution
Process knowledge and building capability in the same person. The operating model changes.
Structural limit
Duplication, sprawl and unowned artifacts. Without maintenance the asset base decays.
Constant across all three stages
The central unit does not disappear. Its remit shifts.
Its task in stage 1
Access, licensing, policy and the boundary of the data rooms
Its task in stage 2
The applications themselves, and the entire budget
Its task in stage 3
Shared components, guardrails and evaluation. Cost attributed per application, spend thresholds in place of case-by-case approvals, a named owner per artifact.
The stages are not mutually exclusive. As a rule all three run in parallel. What matters analytically is which stage carries the majority of the work.
Axis 2 · Integration depth
The second axis measures not building capability but coupling to processes, systems and data. As integration depth rises, substitutability falls and integration effort rises accordingly.
Integration depth A · Off-the-shelf tools
What it is
Licensed, generic assistants with no coupling to your own systems.
Contribution
Short time to introduce, individual relief that is felt immediately.
Structural limit
Fully substitutable. Competitors buy the same capability on the same terms.
Integration depth B · Process integration
What it is
Coupling to planning, quality assurance, billing and other core applications, including the rules held there.
Contribution
Output follows your process logic. Not substitutable, because your process logic is not.
Structural limit
Lead time of several months. Requires domain knowledge and deterministic guardrails around a probabilistic method.
Integration depth C · Your own data
What it is
Analysis, scoring and prediction on data that only your organization holds.
Contribution
The application becomes an asset rather than a running cost item.
Structural limit
The constraints are data quality, legal basis and maintenance, not the model.
A fourth level, developing your own models, is regularly named as the target state. For most organizations it is currently the wrong ambition. Differentiation today comes from exclusive data and from entanglement with processes, not from proprietary algorithms.
Success factors · Independent of position
The position sets the effort. Five success factors decide whether it pays off.
Proprietary data
Identify, structure and legally secure proprietary data: it is what separates your answers from everyone else’s in your category. The AI applications must be able to access it and understand its context and data logic.
Encode knowledge
Embed domain expertise, judgment and problem-solving patterns into skills and plugins, with named ownership and incentives. Encoded knowledge lifts performance: gains arrive immediately, even for brand-new users (e.g. early-career professionals).
Verification in the workflow
System- and tool-based verification as a fixed step in every workflow; additional human sign-off where errors are irreversible, regulated or external-facing.
Recut roles
Actively recut roles, responsibilities and task profiles for AI-changed workflows; freed capacity is named, then redeployed or harvested. Active change management counters uncertainty and raises motivation and engagement.
Baseline measurement
Before the first implementation comes the baseline; after it, impact measurement per application. Without a baseline, every claim of impact stays unproven.
Approach
A track record shows what is possible. This section shows how it runs inside your organization: who builds, what gets measured, and what remains on the final day.
The handover package as it would sit with you
Handover list · excerpt
What stays in the house on the final day
01
The deliverable is not a recommendation report but a clear positioning strategy and working tools: skills, calculators, trackers, decision-ready reporting. Built on your real cases and in use from the first week.
02
We build together. Your experts define what good looks like and hold the pen on at least one tool themselves. Whoever has built one method can build the next alone.
03
Before the first implementation comes the baseline: what the AI produces on its own, and what your team produces today. Only then can the difference be evidenced.
04
The goal is not an extension. What remains is a working library, a named owner for every tool, and a team that can evaluate and extend it.
05
No success rates without a source. Figures come with source, date and limitation, or they do not come at all. That applies to market data and to the effect of my own work.
06
Tools nobody approves have no effect. Cost attribution per application, thresholds instead of case-by-case approvals, clear ownership, so that IT and finance can go along.
One function and one task type are enough to start: measurable results within the first six to eight weeks.
Let’s talk about your program.Further consulting fields
Programs that get delivered
Transformation, turnaround and separation programs: from the target picture to Day-1 readiness.
Reference project: the operational and financial split of ADAC into the three-pillar model (e. V. · SE · foundation).
From the deal into the line
Negotiation leadership, due diligence and post-merger integration in one pair of hands.
Share transfer of 51 % of ADAC Autoversicherung AG (> €100M) including the first integration phase; +100 % sales growth in switching business in year one.
Steering that holds
Organizational design, steering models and KPI systems matched to strategy.
Supervisory board of Arisa Assurances SA & Arisa Ré, Luxembourg (2017 to 2019), chairing the audit committee; further governance mandates since 2024, including DAX-listed groups.
From idea to legal entity
Group and business-unit strategies, and building new businesses through to a standalone entity.
CO₂-certificate trading for EV owners: 170,000 customers and €100M revenue in twelve months, then spun off.
Track record
Project lead for the operational and financial separation into e. V., SE and foundation (Oliver Wyman).
Set up and led the Europe-wide turnaround program across eight country units: program governance, steering cadence with the managing directors, board reporting (2026).
Business development and project delivery under a framework agreement.
AI transformation programs in a corporate environment: process redesign, interactive decision-maker reporting, reusable method libraries.
Hands-on training programs, skill libraries and working standards; designed and rolled out within the Allianz group.
Share transfer from Zurich to Allianz, including leadership of the first post-merger integration phase; +100 % sales growth in switching business in the first year.
Consolidation of the roadside-assistance and legal-expenses insurance entities, with full commercial responsibility.
Mandates with responsibility for outsourcing and chair of the audit committee.
For a leading European assistance and travel insurer: account segmentation, cross-functional team structures, steering model; advisory followed by line responsibility.
Corporate strategy, controlling, accounting, shareholdings management and strategic projects with signing authority; outsourcing contract owner for HR, procurement, finance and tax.
Developed the insurance-group Strategy 2022; developed and rolled out the group-wide ADAC Strategy 2030, including brand extension into new ecosystems.
Built the business from scratch: 170,000 customers and €100M revenue in the first twelve months, then spun off into a standalone legal entity (ADAC).
Wallet-based digital membership card, connected-car pilot with 8,000 test users, e-charging partnerships with EnBW and Aral pulse.
Profile
Independent Senior Advisor in Munich, own advisory practice since 2024, interrupted by a line role leading the Europe-wide turnaround program at Allianz Partners (2026). Before that, 20+ years in consulting and line roles: eleven years at Oliver Wyman, eight years at ADAC in director-level and full commercial responsibility.
Mandates in insurance/assistance, energy, automotive/mobility and HR transformation: from board-level projects at DAX-listed groups to negotiation and workshop mandates. Today’s focus is transformation and performance programs with AI-led execution: decision-ready reporting, high execution speed, lean program structures.
Contact