Want To, Not Have To

AAACHT AG was a German early-stage joint-stock company with the ambition to sell AI consulting to mid-sized German companies. Six people on the team. Three worlds in their CVs: operational IT (system management), software engineering, sales.

What was missing on the team: someone who knew from experience how to turn an idea into a product, how to structure a consulting portfolio that could actually be sold, how to build a function that would carry that work durably.

Closing that gap was my brief as Chief Product Officer on a fixed term. Three months, June through August 2024.

The brief broke down into two threads.

First: a consulting portfolio that doesn't begin and end with “we do AI”, but that has a staircase a mid-sized firm can walk up.

The lowest step: a standard offering via an online tool, with a chatbot delivering an instant teaser result. Low-friction. A managing director in a mid-sized firm clicks through it in the evening, gets a first read ten minutes later, knows whether the topic is relevant for their business.

Above that, in stages: workshops for needs assessment. Coaching for the leadership team. Trainings for the workforce. And at the top: bespoke AI development for clients whose use cases left the standard solution behind.

Second: the product-management function as a lasting structure — method, roles, handover points between sales, engineering and customer, reporting.

Both threads were set up at the end of the three months and handed over to the management board.

Three things did not run, in the months after my handover, the way they would have needed to run for what we built to reach the market.

The AG incorporation was delayed. Constituting a joint-stock company in Germany is a regulatory and procedural journey. The formal hurdles took longer than the business plan had foreseen.

Equity-based financing did not come through at the required level. The business model assumed that, after incorporation, tranches would be financed through share issuance. That did not carry at the volume needed.

The market for AI consulting in the German Mittelstand turned out to be tougher than the business model had assumed. Long buying cycles, high evidentiary thresholds, scepticism about external AI promises. That is the 2024 reality in that segment, and a business plan that doesn't price that reality into its tempo and cash burn enters stress.

The sum of those three points: the company was subsequently dissolved.

1. A product function can be built in three months, if you build the portfolio first and the methodology after. The reflex in young teams is to start with methodology — frameworks, processes, tools. That delivers nothing as long as the portfolio isn't concrete. The order is: first the staircase the customer walks onto, then the maintenance of that staircase as an internal function. At AAACHT this order was right, and it was achievable in three months.
2. Selling AI consulting to the German Mittelstand is change selling, not efficiency selling. The standard sales story for AI consulting — “we lower your costs, we accelerate your processes” — is an efficiency story. It misses the actual lever. The lever is: AI, like any new technology, is a change topic. The efficiency gains AI promises can only be realised if employees want to adopt the technology, not have to adopt it. Business models that don't put that at the centre sell a promise that doesn't materialise on use — and the Mittelstand feels that gap faster than other markets, because every managing director is on the shop floor on Wednesday and can see whether the technology is being used.
3. A small early-stage AG is more a business-model construction than a product construction. The order of risks is not “can the product work” → “does it find a market” → “can it be financed”. In an AG early phase the order is reversed: the financing and incorporation risk decides first, the market risk second, the product risk third. A function that works only on the third risk cannot save the second or the first. Whoever enters this setup as a product owner should know that — and should make it clear to the founders on day one, not on day ninety.