Strategy & Use Cases · Insight
The Use Cases HandsOn Turns Down
A strategy sets the Guardrails — it selects what not to do before it selects what to do. We believe: The refusal set is the first movement of any serious AI strategy.
5 min read
May 12, 2026
HandsOn Insights
“We have an AI strategy portfolio now.” We often hear this statement from executives—and it’s tinged with a certain pride and the conviction that they’re “keeping up with the times.” We then quickly become the party poopers, because: An AI project portfolio and an AI strategy are not the same thing. A strategy differs in that it selects what will not be done based on defined parameters before selecting what will be done. An AI portfolio is nothing more than an initial shortlist of potential use cases—but by no means a strategy.
To make the issue more tangible, consider the example of a fictional client: an industrial supplier with €1.2 billion in revenue, three plants, and a division that has been losing margin for the past two years. On the management team’s desk lies a list of 34 AI use cases. Each case has a sponsor, an estimated business case, and a traffic light indicator. Thirty-one of them are green. The head of strategy calls the document the “AI portfolio.” The HandsOn response to such a list is not to assist with prioritization. We start by eliminating use cases: In this particular engagement, after three hours of methodical work on the refusal set, 7 cases remain, and 27 are dropped. This usually leads directly to discussions with the head of strategy, who painstakingly compiled the list in the first place. He is taken aback—because he has defended the document for nine months.
What the market does with use-case strategy
The standard exercise is the 2×2 matrix with circles on it. X-axis business impact, Y-axis feasibility, circle size the estimated investment. Top right is the quick-win zone, and that is where the cases the sponsor can present well end up. The hard cases — usually the ones with the highest structural impact — drift into the bottom right and get “pushed to next quarter”.
That’s not a strategy. A document is produced that makes the executive board feel good, while at the same time a pilot portfolio is created that ultimately fails to scale. The 2024 BCG/MIT survey estimates that the proportion of mid-sized AI initiatives that move from pilot status to full-scale operation is in the range of 25 to 30 percent. The rest are maintained in subsequent quarters, occasionally renamed, and quietly pushed into the line at the end of the year – failing to deliver measurable impact.
The reflex we’ve observed—and one we must overcome—is dangerous: the argument that “we’ll do everything because everything is important.” The portfolio remains at 34 cases, the roadmap stretches out to 18 months, and the scarcity inherent in any strategy—scarcity of attention, leadership bandwidth, data architecture, and employee adoption capacity—is glossed over.
No one says no. So no one has decided.
How HandsOn does use-case selection
The HandsOn method is refusal-first. We build the portfolio in two movements, and the first movement is deletion. Selection comes second.
Movement one: the refusal set. Before a single case lands on a matrix, we write down which classes of applications the company will not pursue this fiscal year. In this exercise, we use three filters:
What survives these three filters is the real portfolio. Applied to the fictional industrial supplier: 7 of 34 cases. Four of them carry the strategic core, three are learning cases, chosen to build internal capability rather than for ROI. The other twenty-seven are not “deferred”. They are deleted. That is the difference.
Movement two: selection within the remainder. Only here does matrix logic come into play, and it is tighter than usual. We score the remaining cases against the HandsOn AI Operating Model — six domains from Strategy to Operations — and check, for each case, in which domain the bottleneck sits. When the bottleneck is not in Strategy or Use Case but in Governance or Literacy, the strategy-engagement answer is: not yet. Address the bottlenecked domain first, then start the case.
What good looks like
A use-case portfolio after this cut is recognisable by three things.
Anti-patterns
The most common anti-pattern is the matrix with circles, described above. The second, more dangerous one, is the sequencing lie — all 34 cases stay in, they are just distributed across 18 months. No one has decided, no one has lost anything, no one has stuck their neck out. The portfolio looks like a strategy and is a distribution. In Q3 the board realises that the bandwidth is enough for 4 cases in parallel, not 12. The other 8 fall quiet. Whoever had decided that in Q1 would have saved nine months of attention.
The third anti-pattern is the quick-win reflex. An AI chatbot for HR self-service produces a nice press release and zero cents of margin impact. A use-case portfolio that consists of three quick wins and a roadmap has dodged the hardest part — the strategic choice.
When a Strategy & Use Cases engagement is the right format
Not every company that needs AI advisory needs use-case strategy first. Five signals make the difference — three or more hits, and use-case selection is the right door.
When the bottleneck sits elsewhere — no inventory of running applications, no clarity on high-risk classification, a capability gap that hollows out any strategy — the strategy engagement is not the first movement. Then the diagnostic leads.
Close
Three weeks after the cuts—again, in this hypothetical example—the head of strategy presents the rejection list to the board. There are 27 cuts listed on a single slide. The discussion during the meeting is the longest of the year. Most board members find it difficult to prioritize. Activity and a long list of cases to be implemented provide a sense of security. At least until the end of the year, when the results are tallied and it becomes clear that cases were not implemented—or, much worse, were implemented without measurable ROI or adoption by the organization. Then you have to face the much more uncomfortable question: What did implementing those 34 cases actually accomplish? That’s why at HandsOn, we always start by trimming the use case list, even if that means things get uncomfortable right away.
Strategy & Use Cases · Engagement
Which cases would you delete?
The HandsOn AI Strategy Engagement helps you identify the right use cases and then move on to focused implementation.