From a reporting tool to an operational layer
GrönLoop tracked and reported waste for companies, and did it well. That was the problem: reporting is the layer general tools are absorbing. The work was moving the product into the process itself, where waste actually moves between companies, and building the two-track plan that makes that affordable.
Q4
Interface build starts here, not before
SITUATION
The MVP had done its job. Companies were willing to track waste when the workflow stayed simple, the time saved was measurable, and compliance reporting turned out to be a recurring pain rather than an occasional one.
But the product was sitting in the output layer: structuring data, producing reports, generating insight. That is a reasonable place to start and a difficult place to stay, because in 2026 an AI assistant can write the report without a dedicated product underneath it. The product was useful. It was not yet necessary.
If the value lives in the output, it can be replaced. If the value lives in the process, it becomes necessary.
THE SHIFT
The direction was to move the product closer to where the operational work happens, and out of the layer that only describes it.
WHAT I RECOMMENDED AGAINST
01
Do not build the interface yet. The strategic value was in the data flows and the compliance dependency, not in the experience. Structured forms and a light back office were enough to run the pilots. I named the conditions that would change that, and the sentence that settles it: we would use this, but our team cannot operate it.
02
Drop the AI-first framing. The product had been presented with AI in front. That framing pointed at exactly the layer that was losing its value. AI stays in the product where it removes work, in classification and in catching bad data early. It comes out of the narrative.
HOW IT WAS MADE EXECUTABLE
The original targets assumed one kind of customer. Work that crosses organisations involves several decision makers and moves slowly, so chasing it while trying to hit a near-term revenue number puts the two in conflict. Splitting them removed the conflict.
Track A, the revenue baseline. Single companies where the product delivers value on its own. These close faster and fund the work. Target: 15 to 20 companies.
Track B, the strategic pipeline. Three to five multi-party relationships where shared workflows can be piloted. Measured in validated workflows, not closed deals.
Revised for Q3: 15 to 20 paying companies, €6,000 to €10,000 MRR, and at least one cross-company workflow live with two or more organisations. Pricing follows usage and data volume rather than seat count.
Shared structure. Waste categories and reporting formats differ between companies, so data fragments at every border. A shared structure is a dependency rather than a feature.
Work that crosses companies. One company produces the waste, another moves it, a third processes it. Those handovers happen outside any system today.
From explaining to doing. The system described what happened. The next version guides what happens next.
A startup cannot announce an industry standard, so the only workable route is to line up with rules that already exist. The real question is which regulations and which industry contacts anchor the work.
This kind of change costs more. Making an existing workflow better is the cheap option. Changing how separate companies hand work to each other is the expensive one, and it is the only one that gives the product a position nobody can copy. Everything in the plan that looks slow, the two tracks, no interface yet, the manual back office, is there because the strategy is already the expensive choice.
Standards are about getting people to use them, not about building them. A shared way of classifying waste is easy to design and very hard to make anyone adopt. A startup cannot just announce a new industry standard. The workable route is to line up with rules that already exist, which means the real constraint is which regulations and which industry contacts anchor the work, not how much can be built.
Data that moves between companies costs more than data that stays put. Moving records between parties is the line most often left out of a cloud budget, and cross-company workflow is exactly that. So the price has to follow usage and data volume rather than number of users, or the product ends up paying more to run than it charges.