Senior Program Leader · AI / M&A / IT
Impact emerges when concept and execution fit together. I plan cleanly, steer agile to the right degree, and decide quickly and clearly which method the situation requires. I use technology deliberately, as a means to an end.
People are at the centre: whoever deliberately leads and develops employees wins satisfied customers and results that last. I have shown this in crisis management during COVID, which I led for an operations centre of more than 750 people, and likewise in more than 16 M&A IT integrations and in AI process automation and program recovery projects.
The situation
“We bought the licences and gave the team access. What we see: rising costs. What we don’t see is cost reduction elsewhere.”
“The acquisition completed six weeks ago, but the integration plan is not in place yet.”
“It has been reported green for three cycles. Everyone in the room knows it isn’t.”
What I do
Your AI spend is not producing a measurable return. There are licences, pilots and a steering committee — but no process that runs differently from two years ago. Process and technology are roughly 40% of the effort. The other 60% is the people who run the process every day: their objections, their expertise, their fear of automating themselves away. Skip that and you get a technically successful pilot nobody uses. Take them along and the same technology reaches production.
IT is delaying your synergies and extending the period in which you keep paying the seller for its IT (Transition Services Agreement, TSA). IT is the critical path of every integration and the last function to get a plan. Applications, identity, service desk, infrastructure and data all have to land on one date, while both organisations keep serving customers. Cross-border, the failure point is rarely technology. It is language, time zone and working culture.
Your program is red and nobody is saying so. Programs rarely fail loudly. Scope drifts, dates slip a fortnight at a time, and the status stays green because nobody is rewarded for saying otherwise. By the time it turns red, the options are expensive and the sponsor has lost the room.
Business case
A representative core process, costed the way I put it in front of a steering committee — before a single line of scope is agreed.
How to read thisThirty-five per cent sits below what I have delivered — one core delivery process went from 20 steps to 12, a reduction of 40%. The figures above are a worked illustration, not a forecast, an offer or a commitment. They exclude licences, infrastructure and change effort. And the caveat that decides it: released capacity only becomes value once you can name the revenue or cost line it lands on. Your numbers are established in the use-case portfolio, against your own volumes.
How I work
If a program is amber, it is reported amber — in week two, not week twenty. Sponsors can handle bad news. They cannot handle finding out late.
Volume, cost, error rate, cycle time — and the line the value lands on. If the number does not hold, I say so before you spend, not in the closing report.
The playbook, the standards and the capability stay in your organisation. A mandate that cannot end is a mandate that failed.
No engagement pyramid, no junior team learning on your budget. You get the person you met, in the room, holding the date.
How we contract
NDA signed before the first substantive conversation. In live transactions, deal-sensitive material stays on your systems — I work inside your environment, not alongside it.
Swiss law, place of jurisdiction Zurich. Your own framework agreement is fine; I have signed on client paper before and will not hold up a start over boilerplate.
Single point of accountability is the offer — and the risk. It is managed by documenting decisions as we go and keeping your own people close to the work. For engineering capacity beyond one person I work with NudgeIT.
Every mandate has a defined end. Playbook, standards and reporting stay with your team. A mandate that cannot end was scoped wrong.
Programs succeed or fail on what happens between people.
Keep learning. Ask what’s next.
Reimagine the work. Build what’s new.
Lead with empathy. Put people first.
Boldly. Learn as you go.
Better together. Solve it as one.
Who I am
Over twenty years in the volunteer fire department, several of them in a command role. That is where I learned to decide with incomplete information and a running clock — and to name the problem while it can still be fixed.
Volunteer Fire Brigade Kilchberg-Rüschlikon · Officer rank · since 2005Most recently I sat in the global core management team (N-2) of SoftwareONE, a software and cloud provider with 9,000 employees across more than 60 countries — leading the delivery model for a CHF 70 million business line. Before that I built the M&A IT Integration function and the PMO practice, and led IT infrastructure program management. That I now lead AI programs into production is not an accident — it is the result of a standing habit: asking what comes next before the market demands it, and then going to learn it.
What runs through all of it: technology was never the bottleneck. The question was always whether the people came along. You do not roll out a global delivery model to 1,600 people by memo — you build it with them, listen before you redesign, and create the conditions in which others get better. That is why I trained for it — Certified NLP Master, Coaching for Transformation. Not as a shortcut, but because I don’t trust instinct alone.
The part that is hard to replicate: I integrate across language borders and name what others prefer to leave unsaid. A 1,400-user integration in Colombia. Business continuity across India, Germany and Mexico. Steering call in Spanish, board report in German. And if a program has been reported green for three cycles when it is clearly red — that goes on page one of my report.
I was an operator, not an advisor. The integrations were mine, with my budget and my name on the steering report. Today I ask the same question I always did: which process costs the most, which people run it every day — and what does it take for both to look different in ninety days? That is the craft I bring.
Straight answers
Because the model runs in weeks and the organisation takes months. A process only changes once the people who run it understand the new flow, trust it, and apply it without supervision. Skip that and you get a technically successful pilot with zero adoption — while the cost runs anyway. In my experience process and technology account for roughly 40% of the effort; the other 60% is involvement, training, governance, and an honest answer to what happens to the time that is released.
That belongs on the table in the first workshop, not after go-live. There are three honest answers: the capacity lands on a named revenue or service line, it is absorbed through natural attrition, or there are job losses. Which one applies is a decision for the executive team — not for me and not for the tool. My job is to make sure the decision is taken before the start and communicated openly. Unspoken intentions are the most reliable way to lose an AI program.
The integration readiness assessment is a fixed fee from CHF 28,000 and takes three weeks. Running the integration program through to cut-over is charged at CHF 2,000 per day. Total cost depends on the number of users, applications and legal entities involved. The assessment produces a cost envelope you can defend to the board before you commit to delivery.
A Transition Services Agreement (TSA) is the contract under which the seller keeps providing IT and other services to the buyer for a defined period after closing. If the integration runs late, the Transition Services Agreement may have to be extended — depending on the deal at a cost of CHF 50,000 to CHF 150,000 per month, on top of delayed synergies. The exit date from the Transition Services Agreement is therefore one of the most expensive variables in a post-merger IT plan.
Yes. Mandates run internationally and are delivered in English by default — the working language of most global programs. German, Spanish or Portuguese are available in addition where a local team works faster in its own language. Delivered examples include a 1,400-user integration in Colombia and business continuity leadership across India, Germany and Mexico.
The work is done personally. No engagement pyramid, no junior team learning on your budget. Where a mandate needs engineering capacity beyond one person, I work with my delivery partner NudgeIT. Otherwise additional capacity comes from your own organisation — which is also how the capability stays in-house after the mandate ends.
It means deciding, case by case, where the model acts on its own and where a person reviews or signs off before anything is committed. The rule of thumb is the cost of being wrong: low-value, high-volume steps run automatically, while exceptions, financial commitments and anything a regulator or auditor could ask about keep a named human approver. Every checkpoint is documented, so the process is defensible after the fact — not just fast.
One mandate is taken at a time; next availability on request. A first conversation of 30 minutes is free and results in a straight answer on scope, timing and price — including whether the engagement is a fit at all. A written proposal follows within five working days.
Next step
Bring the situation, the deadline and the constraint. You will leave the first conversation with a straight answer on scope, timing and price — including if that answer is that you do not need me.
Based in Rüschlikon near Zurich, working on site across Switzerland, DACH, Iberia and Latin America — and remotely where that serves the program better.