AI Transformation

    AI that makes it from pilot into everyday work.

    Leadership sets the direction, the teams bring their knowledge of daily work, and a first use case arrives where it helps. I set an ambitious direction and start with a small, low-risk step.

    88%

    of organisations already use AI in at least one function

    McKinsey, The State of AI, Nov 2025
    ≈3×

    as often, AI high performers redesign their workflows

    McKinsey, Nov 2025
    57%

    of companies in Germany use AI, the majority for the first time

    Bitkom, Sept 2026

    Leadership sets the direction. Teams find the lever.

    The value comes when both meet: a clear vision from the top and the knowledge of the people who do the work. Then the first use case makes it into everyday work, and the next ones follow faster. I also help teams think a little bigger than the budget request they would have dared to write.

    An example: AI-generated quotes, in production since 2018
    01Top-down & bottom-up

    The two forces of an AI strategy

    Executives & top management

    Executive sponsorship and a clear vision give AI initiatives their direction. Leadership makes sure they pay into business goals, get the resources they need and have backing across the organisation, while giving teams the freedom to experiment.

    Mid-level managers & ICs

    The people close to daily work and customers know where AI helps most. They find effective, feasible solutions for real problems in their own workflow, carry adoption and make sure, through feedback, that tools fit into existing processes.

    02Multi-directional strategy

    Six dimensions of every initiative

    For every AI initiative, whether it starts at the top or in a team, I work through six dimensions and clarify the role of leadership and of the teams in each.

    DimensionLeadershipTeams
    Strategic focusSet the vision, align use cases with business goals, define guardrailsPrioritize the most pressing pain points from daily work and customers; start human-centered and low-risk
    ExplorationCreate space, budget and permission to experimentTest tools in practice, share findings, surface successful experiments upward
    Responsible AISet binding AI standards, align with values, anchor governanceAdhere to standards, anticipate risks early, test for bias, safety and ethics
    ResourcingProvide platform, data access, skills and budgetLeverage existing data/tools; make the ROI case when more is needed
    ImpactDefine KPIs, connect impact to business goalsQuantify effects (efficiency, satisfaction, cost), track KPIs, make value visible
    Continuous improvementInstitutionalize learning loops, enable scalingGive feedback, iterate solutions, spread what works
    03Prioritization

    Finding the right use cases

    Good use cases are found systematically. I connect AI capabilities with your business priorities and start where the benefit is high and the risk is low.

    1. 1

      See the real process

      With Event Storming I map the real sequence of work together with the people who run it, including handovers and waiting times. That shows where AI can help most.

    2. 2

      Collect ideas with the Creative Matrix

      A grid maps business goals such as customer experience, efficiency, innovation and productivity against AI capabilities. Each cell produces concrete ideas.

    3. 3

      Prioritise by impact and feasibility

      The first use case is the one with high value and low risk. Every idea is validated before it is built.

    04People at the centre

    AI makes people stronger

    AI works best when it extends human thinking and takes over the repetitive. People gain time for judgement, creativity and relationships.

    Augment: people get stronger

    • Critical thinking & problem solving – AI provides data, humans interpret and decide
    • Creativity & innovation – AI generates options, humans push the boundaries
    • Relationships & collaboration – AI shares information, humans build trust
    • Strategy & vision – AI forecasts, humans set the long-term direction

    Automate: people get time back

    • Repetitive, rule-based tasks – data entry, retrieval, formatting, basic code
    • Time- and resource-intensive tasks – research, analysis, summaries, first drafts

    A person stays in the loop at every step: Data selection & preparation · Prompt design & refinement · Output review & correction · Continuous monitoring & feedback

    05Data Readiness

    Good data makes AI reliable

    I clarify early which data you already have and how it can be used. You can often start with what is there and improve the rest along the way.

    Quality
    Accuracy · Completeness · Representativeness · Consistency · Relevance
    Accessibility
    Availability · Cost · Format
    Structured and unstructured
    Tables, images and free text each need their own approach. That is the first thing I clarify.
    06AI Adoption Framework

    Your next step, clearly defined

    I locate where you stand today and define the next realistic step, along the themes Learn · Lead · Access · Secure · Scale.

    01

    Tactical

    First experiments, individual tools, learning effects

    02

    Strategic

    Aligned use cases, governance, first scaling

    03

    Transformational

    AI as part of the operating model, organization-wide

    My approach in five steps

    1. 1Assessment & Event StormingMaturity level, data situation, real process bottlenecks
    2. 2PrioritizationCreative Matrix → impact × feasibility
    3. 3PilotLow-risk use case, human-in-the-loop, clear KPIs
    4. 4Responsible-AI checkBias, safety, compliance, values
    5. 5Scale & learning loopRoll out, measure, iterate
    07EU AI Act Readiness

    Ready for the AI Act, together with your legal counsel

    The EU AI Act introduces its obligations step by step until 2028: AI literacy since 02.02.2025, transparency from 02.08.2026, high-risk Annex III from 02.12.2027 and Annex I from 02.08.2028. I build the technical requirements into the system design from the start, so compliance comes with it.

    01

    Risk classification workshop

    Map your AI use cases against the Act's risk tiers (minimal, limited, high, prohibited), before you build.

    02

    Technical documentation & logging

    The artifacts Article 12 expects (architecture specs, eval records, audit trails) are produced as part of how I build.

    03

    Human-oversight design

    Review gates and override points, built directly into the system architecture.

    04

    Vendor & model due diligence

    Assessing the third-party AI providers you integrate against the obligations you carry as a deployer.

    My contribution is technical and does not replace legal advice. Final compliance sign-off stays with your legal counsel.

    08FAQ

    Frequently asked questions

    Top-down or bottom-up – which first?
    Ideally both in parallel. The vision sets the direction, and practice makes sure something arrives in everyday work and grows.
    We don't have a clean data foundation yet. Too early for AI?
    No. I start with low-risk use cases and build data readiness in parallel.
    What if the AI makes mistakes?
    That is why human-in-the-loop is always part of my approach: in data selection, prompting, output review and monitoring.
    Will AI replace our people?
    The focus is augmentation: people are freed up for strategy, judgment and creativity, while the repetitive work is automated.
    Do we have to commit to a specific AI platform?
    No. My approach is tool-agnostic. The methodology (Creative Matrix, maturity levels, Responsible AI) works independently of platform and product.
    Do you provide legal sign-off on EU AI Act compliance?
    No, final compliance sign-off stays with your lawyers. I turn the Act's technical requirements (risk classification, documentation, logging, human oversight) into system design, working closely with your legal counsel.

    What would you build if nobody said no?

    I start with an honest look at where you stand and a use case that shows its value quickly. The direction can happily be bigger than the first step.

    Methodological framework based on the Google Cloud AI Adoption Framework and the Responsible AI Principles. Figures: McKinsey, The State of AI (November 2025); Bitkom press release (September 2026). Figures as of the respective publication.