Software and platformsthat go liveand stay live

Within 10 working days: two executive interviews, the analysis of one prioritised critical workflow, an opportunity and friction model, an economic hypothesis as a value-effort frame, a Pilot Decision Frame, and a decision document of 10–15 pages. Regulatory framework questions are identified and named, not legally assessed. Not included: revision loops, additional interviews, further workflows, architecture deep-dives and implementation work. All of these are available as change requests. Payment is 50 % on engagement and 50 % on delivery of the document.

The fixed price starts at 8,500 EUR net and depends on the scope of the selected workflow. A binding proposal follows a 20-minute introductory call. Payment is split evenly between engagement and delivery of the decision document — the price is fixed beforehand, not billed by effort.

The German KI-MIG has been in force since 29 July 2026 and the AI Regulation broadly applies since 2 August 2026; the Bundesnetzagentur becomes the central market surveillance and coordination body, while sectoral and state competences remain in place alongside it. For financial entities — in particular CRR institutions and Solvency II insurers — BaFin published explicitly non-binding guidance on 18 December 2025 covering ICT risks in AI deployment under DORA, across the full lifecycle from data acquisition to retirement. The assessment names these framework obligations as context for the leadership team; it does not assess them legally. That belongs with your legal department or a law firm. At Daimler Mobility I was Senior IT Manager for Enterprise Architecture & Information Security (11/2016–10/2020) in a BaFin-regulated financial environment; from 2019 as Lead Information Security Officer (ISB), accountable for the ISO team and for MaRisk, BAIT, ISMS and ISO 27001 governance, as part of the global cyber security team.

A fractional CTO carries ongoing operational leadership at several days per week, which is right when a role has to be permanently staffed. The entry point here is built differently: a sharply bounded diagnostic assessment with a fixed price, a fixed duration and a document that stays in the company. Pilot design, transformation mandates and — where the need is concrete — interim leadership responsibility can follow from it. The difference is not availability but sequence: a defensible basis for the decision first, commitment second.

The embedded experience starts at Tönnies Fleischwerk (2004–11/2016, first as a software engineer, from 2006 Head of Engineering & IT Development), a food group with seven production sites. There, over several years, I implemented machine control and warehouse-logistics processes myself – in C, part of the control on single-board computers, with JNI interfaces to the Java systems. The level of accountability followed at SMA Solar. Almost four years of interim responsibility (02/2022–12/2025) for product management, engineering organisation and the architecture of a business-critical embedded software platform for inverters and energy management, three products and around 30 features taken to market readiness across multiple interdisciplinary embedded product teams. Introduced and anchored SAFe, built microservices-oriented domain teams, and established the company-wide architecture platform ArCo with architecture principles, decision processes and governance. In between, Senior IT Manager for Enterprise Architecture & Information Security at Daimler Mobility (11/2016–10/2020).

Both are possible, and the SMA engagement was the operational case: not an advisory position but interim responsibility for product management, engineering organisation and system architecture across several teams, including roadmap, backlog, releases, dependencies and priorities. Which form fits — mandate, advisory or interim leadership — is settled in the 20-minute introductory call.

Structures that keep working without me. At SMA Solar those were the company-wide architecture platform ArCo with its principles, decision processes, documentation and governance; SAFe anchored organisationally rather than taught in a workshop; microservices-oriented domain teams; and end-to-end Jira and Confluence processes for portfolio management, architecture and software development. In the assessment it is the decision document, which stays in the company and can be argued with in my absence. An engagement that only works while the external person is present has missed its purpose.

Because the gap between AI strategy and usable AI implementation almost always sits in the same place: missing experience with the actual engineering operation. I combine three layers: built it myself (embedded platforms, system architecture, today agentic AI systems), then led it (engineering organisations and product accountability at interim leadership level), now orchestrate it (AI governance, transformation mandates). That turns an AI initiative from a tooling question into an architecture and organisation question. That is where it is decided whether it ever leaves pilot status.