Friedrich Jakob Söhn

The first question
is never about
the technology.

Most decisions get made before anyone has looked at the full picture. The technology, the organisation, the commercial reality, the second-order effects — stripped away until the decision feels manageable. That's usually where things go wrong.

Aviation, finance, manufacturing, legal, chemical, technology — different industries, different problems. Same questions underneath. The approach travels because the questions don't change — only the people and the context around them do.

Current role
DPM Lead Global Functions & GPO IT
Previously
Head of Application Development & Automation · COO · Consulting
Languages
German · English
Based
Battenberg, Germany
Background

Both sides
of the table.

Currently DPM Lead for Global Functions — IT portfolio ownership across HR, Finance, Legal, and Procurement, working directly with the executive leadership of each. Before that, Head of Application Development and Automation, and before that Head of Workplace Development and Automation: in both roles, building teams and capabilities from scratch, not inheriting them.

What made that possible: starting in consulting, then becoming COO of an IT and AI company. Writing proposals, winning clients, understanding what customers actually need underneath what they say they want — and then being the person who had to make it all real on the other side. That gap between what gets sold and what has to work doesn't go away once you've seen it from both directions.

The commercial instinct, the ability to read a customer, the organisational judgment, the habit of not cutting the picture down before it needs to be — all of it comes from having stood in both places. And the habit of working with new technology directly — not waiting until it's mainstream — comes from the same place: you don't understand what you're deciding about until you've done it.

How I think

Culture first.
People second.
Technology last.

01
Build the picture. Keep building it.
Most decisions get made on a simplified version of reality — because holding full complexity is uncomfortable and a cleaner picture is always available.

The alternative isn't waiting for certainty. It's absorbing as much as possible, deciding at 80%, and continuing to learn after the decision is made. Not every decision will be perfect. But the picture keeps growing — and that's what makes 100% delivery possible even when no single decision was complete.
02
Lead people. Don't manage their tasks.
Servant leadership is a posture, not an abdication. Listening, supporting, removing obstacles — that's where trust starts. But a leader is accountable for strategy, for translating that strategy into what it means for each role, and for understanding where each person stands and where they need to go.

Empowerment without direction is just comfortable ambiguity with a better name.
03
Technology is the answer. Never the question.
"We need a new system" is a conclusion, not a problem statement. Licence models, vendor lock-in, maintenance costs, indirect usage implications — these aren't finance department problems added after the architecture decision. They are inputs to it.

Bringing commercial reality in early changes what gets built. Leaving it out changes what it costs.
In practice

Not case studies.
Outcomes.

Ownership, approach, and the discipline of seeing it through.

Mindset
Holding complexity long enough to act on it well
Culture shapes how people work together. People determine what an organisation can actually do. Technology enables — but only what the culture and the people are ready for. In that order. Always.

That sequence changes how you approach almost everything. Understanding where someone actually stands, where the organisation needs them to go, and what it will take to get them there — that comes before any conversation about tools or systems. A team that doesn't have direction and real ownership won't be fixed by a better platform.

And when the technology question does arrive, it arrives with full context — the commercial reality, the organisational dependencies, the second-order effects that don't appear in the business case. The picture keeps building, the decisions keep correcting, and the goal stays the same throughout:

Enough clarity to act well — and enough honesty to keep learning after the action is taken.
Architecture & commercial reality
When nobody owns the full picture
On a large transformation programme, the architecture team was doing its job — evaluating options, aligning on standards, making defensible calls. The commercial team was doing its job — managing budgets, tracking spend, reviewing contracts.

The problem: the consequences of a technical decision can land entirely in the commercial domain. Neither team will catch it unless someone is across both at the same time.

That's the gap. Not incompetence on either side — a structural blind spot that only closes when someone is genuinely working across both. Running the architecture team while keeping the commercial picture in the room isn't a process. It's how you think.
Programme leadership
Turning an IT project into a business project
The typical application replacement programme is driven by IT and received by the business. IT defines the solution, manages the delivery, and hands over something the business didn't fully shape and doesn't completely own. The result is predictable: 80% delivered, systems that never quite get switched off, workarounds that become permanent.

The shift that changes the outcome: rebuilding the project logic around business processes rather than technical systems. Not as a governance change on paper — by actually mapping each process end to end, understanding it from backend logic to front-end experience, and transferring real ownership to the people who run those processes every day.

When the business owns the outcome — not the ticket, the outcome — they make different decisions. They flag problems at the boundaries. They push for the last 20%. They don't accept workarounds because they understand what the workaround actually costs them.

Strategic ownership stays at the centre — watching the connections, holding the priorities, making the calls that nobody else can make because nobody else sees all of it. That's what gets a programme to 100%.
Curiosity
Getting there before it has a name
Machine learning in production environments in 2018 and 2019 — TensorFlow, Python, R, Hadoop, image analysis, process monitoring. Not because it was on a roadmap. Because it was clearly going somewhere important and the only way to understand what that meant was to work with it directly.

The same instinct runs through everything: build the MVP, not the concept. Show something working before asking anyone to trust it. Use the tool to design the process that will use the tool — because that's where you find out what the theory missed.

Last night: a Linux partition, the Claude API, a Telegram bot that pushes articles directly to the website. Not because it was necessary. Because understanding what you're deciding about requires doing it — not reading about it.
Industries

Same approach,
different context.

No industry-specific playbook. The questions that matter — about technology, organisation, and what decisions actually cost — don't change with the sector.

Aviation
>35,000 – 130,000 employees
Financial Services
>200,000 employees
Manufacturing & Heating
>10,000 employees
Chemical Industry
>500 employees
Legal & Compliance
>1,100 employees
Networking / Technology
>75,000 employees
Cooling & HVAC
>3,000 employees
IT & AI Consulting
Built & led from ground up
Expertise

The scope
in practice.

Two decades across every seat in the room — consultant, COO, architecture lead, programme director, portfolio owner. Organisations from 500 to 200,000 people, industries with nothing in common except the questions that keep coming up regardless.

Currently working at Director level. Open to the right permanent role where the mandate actually matches the scope — and available for consulting engagements where the situation calls for someone who thinks across the whole picture.

People & leadership development
Accountable for where each person on the team stands and where they need to go — not just as a resource, but as someone with a trajectory. That means understanding the gap between current capability and what the organisation needs, developing a target state together, and having the honest conversations that make development real rather than performative. Servant leadership is a posture, not an excuse to step back.
IT portfolio & governance
Structuring portfolios so leadership can make real decisions — not manage the appearance of making them. Budget ownership, prioritisation, target architecture, and the governance layer that ties business demand to what actually gets built.
Programme & transformation leadership
Full ownership of complex IT programmes — accountable for outcomes, not status updates. The first weeks go into understanding what's actually happening, not pitching what comes next. That's the difference between getting a stalled programme moving and adding another layer to the problem.
Enterprise architecture & platform decisions
Technology decisions made with full awareness of their commercial and organisational weight. Led architecture teams. Made platform calls that held up — because they factored in licensing, vendor risk, what the organisation could actually absorb, and what the team could realistically ship.
Commercial & licensing intelligence
Vendor contracts, licence models, indirect usage implications — input to decisions, not cleanup after them. The cost of missing this doesn't show up immediately. It builds quietly until it does.
AI & automation strategy
Working with machine learning in production since 2019 — before most organisations had an AI position. Still working with it directly today: building, testing, integrating. Helping organisations get a clear picture of what AI actually means for their structure, their roles, and how they deliver. Process by process, not as a headline programme.
Interim & fractional leadership
Step in where leadership capacity is missing — programme owner, IT lead, COO ad interim. Ready from day one. Every role here has been done, not just advised on.
Selected projects

Work that
required something.

Structured by what each situation actually demanded — not by size, date, or title.

Enterprise portal replacement — 230 applications, 100% delivered
Took over a stalled programme to replace a company-wide application and microservice landscape. Rebuilt the project logic around business processes rather than systems, transferred real ownership to the business, and coordinated architecture, delivery, and commercial decisions in parallel. Most programmes stop at 80%. This one didn't — including a mid-programme decision to replace a purchased solution with a custom build after the global rollout cost made the vendor's model unworkable. That went to the board. It was approved.
Manufacturing
>10,000 employees
4 years
Corporate integration — application strategy across a 50,000-person acquisition
When a much larger organisation acquires a smaller one, the assumption is usually that the larger one's technology wins. That's not always the right answer — and it's rarely as straightforward as the integration roadmap suggests. Defined the application strategy for the acquired entity, challenged integration assumptions that looked solid on paper but had no realistic delivery path, and negotiated the actual terms of what would move, what would stay, and what would be retired. The TSA exit and contract negotiations that followed were built from scratch — the original agreements weren't fit for purpose.
Manufacturing
12,000 + 50,000 employees
Platform exit under live operations — full Google workspace decommission
Full responsibility for decommissioning a Google-based application and platform landscape across a subsidiary group, while keeping the business running throughout. The definition of "done" kept shifting as the target architecture evolved. Delivered anyway.
Manufacturing
Group-wide
Organisational restructuring into self-organisation — built from scratch
Restructured a consulting organisation of ~50 people into a holacratic model — self-organising teams, distributed authority, no traditional management hierarchy. Developed the team concept, appointed technical leads, reorganised cross-divisional collaboration, and adapted the tooling and process structures to match. First time building something like this from the ground up. It worked.
IT & AI consulting
~50 people
ERP separation — the work behind the work
The official programme had a big-name consultancy attached. The actual clarification of what needed to happen — the architecture decisions, the process logic, the open questions that the official documentation hadn't resolved — happened elsewhere. Sometimes that's how it goes.
Manufacturing
Enterprise-scale
RPA & automation — building a permanent capability, not a pilot
Established RPA technologies and built the organisational and technical foundations for sustainable process automation. Strategic direction, use case selection, and integration into existing landscapes. The goal was a capability that would outlast the programme — not a proof of concept that would need rebuilding later.
Manufacturing
>10,000 employees
AI & data science — manufacturing, aviation, legal, financial services
Machine learning and data science projects across four industries, starting in 2018 — TensorFlow, Python, R, Hadoop, data pipelines, image analysis in production, process monitoring in regulated environments. MVPs instead of POCs: something working that people could challenge and decide to trust, not a slide deck about what might be possible. AI also used as a door-opener for data conversations that wouldn't have started otherwise.
Multiple sectors
2018 – 2020
Horizon 2020 — AI in critical care ventilation
Authored the IT and AI components of a successful EU Horizon 2020 funding proposal for applying AI to the ventilation of critical care patients. Several million euros secured. The project was implemented — by others, after moving on. A pocket knife arrived as a thank you.
Healthcare / EU
Research funding
SAP implementation — architecture designed around the licence model
Full SAP rollout at a new production facility, including enterprise architecture, international coordination, change management, and operational handover. The architecture was designed with the indirect access licensing implications in scope from day one — not flagged by an audit afterwards.
Cooling industry
>3,000 employees
12 months
Fuel Data Management & Business Intelligence — aviation
End-to-end BI platform for a major aviation customer — data architecture through to reporting layer. Product management, international coordination, customer presentations at industry events.
Aviation
>130,000 employees
2 years
Process Monitoring (classified)
Technical design, process definition, and full project management for a sensitive monitoring system. Regulated environment, stakeholder complexity, long delivery timeline. Everything defined from scratch.
Legal sector
>1,100 employees
20 months
Azure IAM & Identity Platform — global rollout including China
Central identity and access management on Azure, coordinated across multiple countries. Application testing coordination in China. Internal, nearshore, and external teams across time zones.
Manufacturing
Global
4 months
Contact

If this reads like
the right person
for what you have —

Whether it's a programme that needs real ownership, an organisation navigating a transformation that isn't going as planned, or a Director-level role where the mandate matches the scope — direct contact works best.

jakob@proccit.com
Battenberg, Germany

Currently operating at Director level and open to the right permanent role — where the mandate matches the scope and decisions are made on substance. Also available for consulting engagements where independent, full-picture thinking is what the situation calls for.