Most leadership teams develop or refine their company strategy every year. The process takes significant effort across a couple of weeks, often months.
How well the process runs shapes the result. Decisions are clearer, ownership is set, and the strategy turns into work instead of staying on a slide.
AI now has a seat at the table. Used loosely, it acts like an intern: quick and willing, but generic. Instructed well, it adds real value. It drafts, sorts, keeps track and pushes back, while the leadership team keeps every decision.
Here is a seven-step process in three phases, and what AI can do in each step.
Phase 1: Frame
1. Setup
Decide who takes part. That can be a whole team, or specific people from across the company. Make their role explicit: they add ideas, weigh in on the key choices and approve the result. Also decide on visibility. Does the work stay private to the team, do you share only the results, or is the whole process open?
Done well: everyone knows their decision rights before the first idea is written down.
2. Context
Frame the work. What kind of strategy is this: revenue growth, product innovation, operational efficiency? Then gather the facts about your market, customers and performance, and about what teams see day to day. Set the ambition and the limits: the horizon, the maximum number of themes, headcount and the target.
The type of strategy frames the brief, the ideas you consider and the choices you will face. Pick up to three:
Revenue growth: Win new customers and grow existing accounts
Customer experience: Retention, satisfaction and ease of use
Operational efficiency: Do more with the same team and budget
Employee engagement: Culture, talent and leadership
Product innovation: New capabilities and bold product bets
Market expansion: New regions, segments or channels
Digital & AI transformation: Change how we work with data and AI
Sustainability & impact: Responsible growth and social impact
Done well: everyone starts from the same picture.
AI's role: the context is the instruction. AI is only as good as what you give it. Market notes, competitor scans and survey results turn a generic assistant into one that knows your situation. The whole leadership team and AI need to think from the same knowledge base.
3. Brief
Write a strategy brief that agrees the strategic question and the limits you work within. Four parts are enough: where we stand, where we are going, how we will win, and what I ask of you.
Done well: one page, written by the CEO, that names what you will say no to.
AI's role: AI drafts the brief from the context in minutes. Leadership edits it. Reacting to a draft is faster than starting from a blank page, and the editing is where the thinking happens. Be aware of framing bias, though. A knowledgeable AI can also be a good sparring partner, it’s always available, and dares to ask uncomfortable questions easily.
Phase 2: Explore and choose
4. Ideate
Go wide. Collect options and ideas from across the organization before you narrow down. One line per idea is enough, and rough is fine.
Done well: many ideas from many people, grouped so patterns show.
AI's role: AI seeds the board with its own suggestions, clearly marked as its own. The team keeps what fits and dismisses the rest. AI can also sort ungrouped ideas into existing groups. It adds breadth, and the people in the room decide what belongs.
5. Key choices
After going wide, you need to narrow it down. 'Doing everything' is in most situations not a strategy. Make the hard choices: what you will focus on, and what you will deliberately not do. Each choice rules something out, and that is the point. Take a 42-person SaaS company with a flat headcount. Does it deepen its home market or enter a new one? Does it sell direct or through partners? Does AI support the product, or become the product?
Done well: a handful of real either/or choices, each with a visible cost.
AI's role: AI plays two roles here, and neither one decides. As a neutral facilitator, it identifies which choices shape the strategy most, shows how many ideas depend on each one, and sets out the options side by side. As an objective supporter, it builds the fact base for each option: what market data, customer input and your constraints say for and against it. The team makes the call. Once a choice is made, AI tracks what that choice now rules out.
Phase 3: Focus and hand over
6. Synthesize
Bring your choices together into a few clear strategic themes.
Done well: themes you would still defend in three years, no more than the maximum you set, and sub-themes that are directions rather than projects.
AI's role: AI reads every idea and choice, flags overlap ("these two groups point the same way, merge them?") and asks its questions before it drafts. It also flags sub-themes that are really projects, because those belong under goals later.
7. Cascade
Hand each theme to the leaders and teams who will run it. Break it further down into team goals and KPIs. When we did this exercise with a 300-person tech scale-up, their annual OKRs turned into 100 to 150 tangible team goals for the next quarter.
Done well: every theme and sub-theme has an owner team and a lead, and approved items land as goals on each team's board.
AI's role: AI suggests an owner for each item, based on who already owns related goals. Leadership adjusts and approves. Nothing is published until every item has an owner.
Most common mistakes
Each step has its own way to go wrong.
| Step | Common mistake |
|---|---|
| 1. Setup |
|
| 2. Context |
|
| 3. Brief |
|
| 4. Ideate |
|
| 5. Key choices |
|
| 6. Synthesize |
|
| 7. Cascade |
|
Using AI well
The same pattern runs through all seven steps:
Give it context, not just a question. The brief, the facts and the constraints are the instructions.
Let it propose. Never let it decide. Mark what is AI's, and keep approval with people.
Use it to sharpen choices, not to make them. A clear either/or is worth more than a recommendation.
Let it do the bookkeeping. It can track which ideas depend on which choice, what has been ruled out and who owns what.
An intern becomes a valuable colleague through context and clear instructions. The same goes for AI.
A strategy process ends where the work begins: on each team's board.