ARCHETYPE·INDUSTRY·ERA

The CEO Wanted AI; the Customers Didn't

Kardex

09/2023 – 02/2024

A CEO expectation to follow the competitor on advertised AI functionality — specifically predictive maintenance — without it having been tested whether the competitor had substance behind the website, or whether customers wanted the feature at all.

01 — TRIGGER

Kardex brought me in in September 2023 for two assignments at once. Coach for the software product-management team — four people, with the task of taking the function to a higher level. And product manager, hands-on, for two concrete products: the cloud-based customer support and IIoT platform with AI functionality, and a new intralogistics-workstation solution.

Six months. Two hats at once. Coaching a function while standing as a player on the same field is a constellation that sounds problematic in theory and in practice gives coaching what it needs: credibility. Whoever stands up and does it can say how it goes.

02 — THE REAL PROBLEM

The four-person PM team had a structural problem that had nothing to do with technical quality. It sat between a clearly more powerful business-development team and an established hardware-PM team, and its standing in both directions was lower than its contribution warranted. Software PM in a hardware-shaped intralogistics company has to re-earn part of its right to exist every quarter.

That became the actual question over the course of the engagement: not “how can we do PM better”, but “how can the team speak with a voice that carries its contribution”.

03 — A CHARACTERISTIC SCENE

The central moment of the engagement wasn't a coaching moment, but a substantive argument with the CEO.

The principal competitor advertised AI functionality on its website — specifically predictive maintenance — as a unique selling point. From the executive board came a clear expectation: Kardex must have it too. Not “someday”, but as part of the current IIoT product.

I asked for the chance to talk to customers before the roadmap decision. I got it. Through business development we could reach not only existing customers but also customers of the competitor. What the conversations brought back came down to two points.

First: the competitor had the terms on the website but no substance behind them. Predictive maintenance was advertised as a feature without the customers we spoke to having experienced it in daily operation. It was a marketing artefact, not a product.

Second — and this is the real point: the customers didn't want predictive maintenance at all. They wanted two other things: more efficient spare-parts mapping — i.e. faster knowledge of which part to get where — and preventive maintenance, i.e. planned replacement before failure.

The logic behind it, once heard, is hard to refute. When the machines stand still, they don't earn money. Predictive maintenance promises to forecast a looming failure before it happens. But that forecast must be followed by preparation and reaction time long enough to procure the right spare part and dispatch a service technician — otherwise the prediction lengthens the standstill rather than shortening it.

The recommendation to the CEO was accordingly: no AI feature into the product because the competitor advertises it. Instead — what the customers actually wanted: better spare-parts logic, cleanly implemented preventive maintenance.

The CEO went with that recommendation.

04 — WHAT WE BUILT

The other product — when sales doesn't come along. The second half of the hands-on work was the new intralogistics-workstation solution. To sales, the solution was a nice gimmick — pretty to look at, but not a product you take to market because day-to-day business pays for something else. To a sister company in the corporate group it was actually a differentiator; there it could have played a role.

Business development was simultaneously pursuing another solution — a mechanical solution to a different problem — and directed all available attention, energy and resources to bringing that mechanical solution into business planning. The workstation solution was as a result significantly curtailed. It wasn't shut down, but it didn't become what it could have been.

That isn't a defeat for the product. It is a lesson about the reality of sales organisations: a product the sales function doesn't want doesn't reach the market, no matter how good it is. Sales isn't the enemy — sales is the reality in which the product either lives or doesn't.

The coaching. In parallel ran the coaching with the PM team. From the team lead down to individual members. Partly also with the hardware-PM team and individual product owners, where the need was there.

What changed over the six months was less methodology and more self-positioning. The team had a clearer understanding of its own role at the end and an audibly improved standing in the much more powerful business development. Standing in such a constellation doesn't come from more power. It comes from a team knowing precisely what its contribution is, being able to formulate it clearly, and not giving it up under conflict.

05 — OUTCOME

Customer interviews showed that customers wanted preventive maintenance and a better spare-parts logic, not predictive; the roadmap was adjusted accordingly, the AI feature was not added to the product, and the four-person software-PM team was lifted over six months into a more sustainable position inside the hardware-shaped corporation and handed over to a successor.

06 — WHAT REMAINS

Saying no is a core PM discipline, not a stylistic choice. The hardest form of no isn't to a sales colleague who wants a custom feature for one customer. It's to a CEO under pressure from a competitor's website who takes the obvious as the right answer. Whoever rebuilds the competitor in that moment without testing the reality behind it is rebuilding a marketing artefact. Whoever asks customers in that moment — and goes back to the executive board with the answer — is doing the actual job of product management.
“All competitors have feature X” is the wrong question. The right ones are two. Do they really have it, or is it only on the website? And: do customers actually want it, or is the whole category a shared marketing reflex? Both questions can be cleared in two weeks of customer interviews. Those two weeks are the cheapest investment available in any roadmap dispute.
PM standing comes from understanding, not from power. In a hardware-shaped organisation the software-PM team is structurally the smaller partner. It doesn't get bigger through reorganisation, but by knowing its contribution more precisely and articulating it more clearly than the functions it speaks with. That is the form of capability building a coach can realistically reach in six months — and it is the form that still holds after the coach.

Working through a similar shift?

If something on this page resonated, the QuickCheck takes four minutes and gives you a structural read of where you actually are.

Take the QuickCheck Start with a Diagnostic →