According to a recent MIT NANDA report, 95% of organizations investing in generative AI are getting zero measurable return on it, despite $30–40 billion in enterprise spending. The reason isn’t the technology, the talent, or the regulation. According to the same study, it’s the approach.
That approach to the problem usually takes one of two forms: companies either pick the wrong process to pilot first or never get past picking at all. A strategic AI workshop exists to solve both by making the first decision well before any code is written or any budget is committed.
This article explains what a strategic AI workshop is, why leadership teams are running them in 2026, who should be in the room, and what you walk away with.
What Is an AI Strategy Workshop?
An AI strategy workshop is a working session where a leadership team looks at the company’s own processes and decides which ones are actually worth piloting with AI. Instead of teaching a tool or showing demos, it walks participants through a structured evaluation of their real work: what they do, how often, with what data, and at what cost.
The participants are usually the decision-makers and process owners. The session lasts anywhere between a few hours and a full day, depending on the format and how many processes are being evaluated. The output is typically some form of written record: a scored list, a ranked shortlist, or a recommendation per process. Different providers structure it differently, but the underlying idea is the same: turn a vague conversation about AI into a concrete first step.
It’s the step before implementation, where you decide what to build.
Strategic AI Workshop vs. AI Tool Training: What’s the Difference?
The term “AI workshop” gets used for two very different things, and companies often book one when they need the other. Most of the AI training on the market today is tool training. The agenda is something like:
- introduction to generative AI
- prompting basics
- hands-on with ChatGPT or Copilot
- a few use cases per department
- and a certificate at the end
The participants leave knowing how to write better prompts. That’s useful, but doesn’t tell anyone which process to automate first.
A strategic workshop is built for a different problem: not “how do we use AI” but “where should we use AI, and what would it take to start.” That difference runs across every dimension that matters:
| Tool Training | Strategic Workshop | |
| Who’s in the room | Wider team (marketing, HR, operations, support) | Executives, department heads, process owners |
| What participants leave with | A skill (better prompts, faster drafting) | A decision (which pilot to fund, or why not yet) |
| Format | Hands-on practice on sample tasks | Diagnostic, scoring, structured commitment |
| Success metric | Did the team start using the tools afterwards? | Did the leadership team commit to a named first step? |
The word “workshop” covers two completely different products. Most disappointment with AI training comes from picking the wrong one.
If your team has never worked with generative AI, they need tool training. But if your leadership is past the “how does ChatGPT work” question and starting to ask which AI project to fund (or whether to fund one at all), more prompt training won’t move them forward. They need a structured way to evaluate options and commit to one.
The most common mistake we see is companies booking tool training when the leadership team actually needed strategy. Both cost money, but the wrong choice costs twice. Once for the workshop. Once for the AI project that doesn’t happen, because no decision was ever made.
4 Reasons to Run an AI Strategy Workshop in 2026
Knowing what an AI strategy workshop is and isn’t doesn’t yet tell you whether you need one. The case for running one comes down to what it gives you that other formats don’t. Below are the four reasons that come up most often.
1. You pick the right first project, not the loudest one
The 95% failure figure from the MIT NANDA report we opened with has a specific shape. MIT found that companies tend to fund the most visible AI use cases (around half of all GenAI budgets go to sales and marketing) while higher-ROI opportunities in back-office functions stay underfunded. In practice, this means pilots get chosen for visibility, not suitability. A strategic workshop is built to flip that order. It scores candidate processes against the same criteria, so the comparison is structured rather than based on whoever made the loudest case in the last meeting.
2. You walk out with a written commitment, not another discussion
Most AI initiatives die because nothing was actually decided, just discussed. A workshop done well produces a written record: a verdict per process, an owner, a measurable goal. (What “done well” means in practice is something we’ve broken down in a separate piece.) That artefact survives the next quarter’s reshuffling of priorities, where a vague “let’s explore AI” almost never does.
3. You catch readiness problems before they sink the project
Most AI pilots fail because the company wasn’t ready to use the technology, not because the technology didn’t work. The data wasn’t where it needed to be, the process owner wasn’t named, and the success metric was never defined by anyone who could actually be held to it. A well-run workshop tests readiness across data, ownership, processes, and decision-making before any contract is signed. That way, the gaps surface during a one-day session, not three months into a paid project when someone has to explain them to the board. If you want to test one process against this framework yourself, our AI Opportunity Canvas walks through the same readiness questions on a single page.
4. You get a vendor-neutral recommendation, not a sales pitch
Most AI training is run by companies that also sell the tools used in the training. The recommendation at the end is rarely surprising. MIT’s research shows that the top thing enterprise buyers look for in AI vendors isn’t features or price; it’s trust, and trust is hard to build when the vendor’s recommendation and their revenue happen to align every time. A strategic workshop, delivered by a partner not tied to any specific platform, can recommend an off-the-shelf tool, a custom build, or no AI at all, depending on what your processes actually need. That neutrality matters most when the obvious answer turns out to be the wrong one, which happens more often than most vendors admit.
Who Should Attend an AI Strategy Workshop?
A strategic AI workshop only works if the right people are around the table. Two groups need to be there, and each brings something the other doesn’t:
- Decision-makers (executives, department heads). They bring the authority to fund a pilot and the perspective on strategy and budget.
- Process owners (managers who run the workflow day-to-day). They bring the knowledge of what would actually work and what won’t survive contact with real operations.
When either group is missing, the workshop tends to produce a decision that the other side quietly ignores.
Group size matters too. A well-run workshop typically fits around fifteen to twenty participants, spread across functions where AI might actually change how work gets done.
- Too few people, and you miss the range of perspectives needed to surface real bottlenecks.
- Too many, and the conversation stops being a decision-making session and starts being a broadcast.
Just as important as the total number is the balance between the two groups. Fifteen executives and three process owners produces a decision the operators will ignore. The reverse produces a decision without a budget.
One pattern worth noting from MIT’s research: the strongest AI deployments tend to come from bottom-up sourcing, where line managers and process owners surface the use cases, not central labs or executive teams working alone. A workshop is one of the few settings where both directions meet in the same room on the same day.
4 Phases of a Well-Run AI Strategy Workshop
Formats vary between providers, but most strategic workshops share a similar shape. They don’t happen in one day. They happen across several weeks, with the actual working session being the middle part of a longer process.
With the right people in place, the phases usually look something like this:
1. Calibration (before the workshop)
A short prep call where the provider learns what your team already knows about AI, which processes are on the table, and what the leadership wants out of the day. This is when the agenda gets tailored to your business, not the other way around.
2. The workshop itself
A single working session with the leadership team in one room. Candidate processes get scored against a set of readiness criteria, discussed, and ranked. By the end, the group should have a clear shortlist and rough verdicts. Length varies from half a day to a full day, depending on how many processes are being evaluated.
3. Written readout
A structured document capturing the scoring, the reasoning, and the recommendations for each process. This is what gets shared with people who weren’t in the room, and it’s the thing that survives when priorities shift next quarter.
4. Results presentation
A follow-up session where the provider walks the leadership team through the readout and answers questions. The decision is more likely to actually get acted on when it’s discussed, not just delivered on paper.
In our own AI strategy workshops we run this as a five-hour session, about two weeks after a one-hour calibration call, with a two-hour results presentation afterwards. The whole process typically spans a few weeks. Long enough for careful thinking, short enough that momentum doesn’t die between meetings.
5 Deliverables You Should Get From an AI Strategy Workshop
The point of a strategic AI workshop is what you take with you when it ends. Different providers package the deliverables differently, but a well-run workshop should leave you with a few concrete artefacts that survive the day:
- A decision on every process discussed. For each candidate, a written verdict: Go (worth piloting now), Not yet (fix something first), or Not here (AI isn’t the right tool for this problem). The value is in the “not yet” and “not here” answers as much as the greenlights, because they save you the cost of a project that was never going to work.
- A scored, ranked shortlist. Candidate processes ordered by business value, feasibility, and how quickly a result would show up. This is what turns “we have twenty ideas” into “here are the two we’re starting with.”
- A rough ROI estimate for each pilot. Not a promise, but a realistic range of what the business impact could look like if the pilot works. Enough for a board conversation, honest enough to hold up under questioning.
- A 6–12 month roadmap. What to pilot first, what comes next once the first results are in, and what to set aside for now. A plan that leadership can act on without going back for another workshop.
- A written report. A document capturing the scoring, the reasoning, and the recommendations, ready to share with anyone who wasn’t in the room. This is what the workshop leaves behind when memory fades.
The common thread across all of these is that they’re written, specific, and actionable. That’s what turns a workshop into a starting point for real work, rather than another meeting to remember fondly.
Ready to Turn “We Should Do Something With AI” Into a Real Project?
If a strategic workshop sounds like the right next step, check out our AI strategy workshop designed around your processes, your data, and the decisions your leadership team actually needs to make. We’ll help you score your actual processes, produce a written shortlist, and leave with one to three pilots ready to go in the next few weeks.
