The classic stage-gate with five gates was basically a savings program: the business case, the market estimate and the strategic fit took a lot of work, so they were done piece by piece and prepared again at every gate. With AI, that reason goes away. So you cut the reporting overhead first by automating the bureaucracy, and then by getting by with one or two gates.
From the idea to the tracked impact of a project, there are often five tools involved, and all of them say they are the single source of truth.
Where does the reporting overhead come from?
The most common work I see in innovation management is copy-paste: from one Excel file into another, from Excel into a web interface and from there onto the slides for the next review, so between tools that don't really talk to each other. Done once, it doesn't seem like much. But you have to do it ten times in a row, because something small keeps changing at the front and every change then has to go through the whole chain. That takes time and is extremely error-prone, because people naturally try to be quick about it.
If changing one parameter in the business case means three hours of work for someone, they may simply not change it and think: good enough. I have seen a number of decisions where people put as much time into the basis for the decision as they could or wanted to, and in the end the decision still had to be made on weak numbers. If you have worked on something for a long time, you tend to think it must be good, but that isn't automatically the case.
Isn't one platform with an innovation portfolio dashboard enough?
The usual answer to why anything still gets copied is: one central platform, a dashboard on top, and the effort goes down by itself. That is what most tool vendors say, and the common AI answers too. That's true as far as it goes, but very few people really have just one platform. Especially in innovation management, from the idea to the tracked impact of a project there are often five tools involved, and all of them say they are the single source of truth, and the innovation manager still carries the data from one to the next to do the whole job.
But even with connected tools, the work repeats at every gate. At the start there is a low-threshold form so that more ideas come in, then after the first gate a first business case has to be calculated, which finance wants in one format and the decision makers in another. A lot of this work is basically communication, getting the information into a format the stakeholders can consume, and that happens again for every gate. So the number of gates is itself a lever on the reporting overhead.
The steering committee doesn't want to see a dashboard either, it wants a PowerPoint or a memo. You shouldn't try to bend the organization too quickly, though, most people resist that very strongly. If the board wants slides, give it slides. But build a process that generates those slides with one click instead of three hours of work, AI makes that possible now.
How many gates does an innovation process still need with AI?
If every gate is another round of the same work, it's worth asking why there are so many. The original stage-gate process with five gates was there to save work. Calculating and reviewing a good business case took a lot of effort, and so did a market estimate and looking at the strategic fit (and whatever else was a gate criterion). That's why only part of the work was ever done, followed by a check whether the rest was worth it at all.
You don't have to do that anymore. AI makes the slides cheap, and building the business case, the market estimate and the fit assessment gets much cheaper as well. So you can build a business case right from the start, decide whether it is worth validating, validate if needed and then decide straight away. The fewer gates, the faster the process, and ideally you bring all decision makers into one or two gates so the momentum doesn't fizzle out.
Automate first, then restructure
Because the organization resists being bent quickly, I wouldn't reverse the order: first automate all the bureaucracy that can be automated, meaning the chain from the idea form to the slide, and only then adapt the process to the new AI reality and talk about the number of gates. That first step is what Bayescase, innovation management software, is built for: the idea, the business case and the tracking of the impact sit in one place, so a changed assumption no longer has to be carried by hand through three files.
As an innovation manager, it's worth asking yourself more often what the job actually is: finding, evaluating and supporting innovations, or building slide decks and digging through Excel files. If the time you spend doesn't match your job title, you are probably doing something wrong.
Frequently asked questions
- How can I cut the reporting overhead in innovation management?
- First automate the bureaucracy: the chain from the idea form through the business case to the review slide, so that a changed assumption doesn't have to be carried by hand through several Excel files and tools. Then adapt the process: with AI you can often get by with one or two gates, and every gate less is one round less in which the same information gets prepared again.
- Is one central platform enough to simplify innovation portfolio reporting?
- In theory yes, but in practice one platform rarely covers everything. From the idea to the tracked impact of a project there are often five tools involved that all want to be the single source of truth, and the data still gets carried over by hand. That is why Bayescase, innovation management software, covers the whole way from the idea to the measured value.
- Why does the steering committee want slides instead of a dashboard?
- Because that is the form it is used to receiving information in, and an organization's habits change slowly. Instead of fighting that, you can generate the slides automatically from the same data, with one click instead of three hours of work.
- Does a stage-gate process still need five gates today?
- The five gates were there to save work, because the business case, market estimate and strategic fit took a lot of effort and were therefore done piece by piece. With AI, that reason largely goes away. One or two gates with all decision makers at the table make the process faster and keep the momentum.
