An AI Roadmap Template: The First 90 Days and the Next 12 Months

An AI roadmap has two halves. The first 90 days are planned in detail: one or two workflows, each with an owner, a baseline and a decision date. The next 12 months are planned loosely, by quarter, from a ranked shortlist. Each part carries owners, measures and the rules everything runs under.

What is an AI roadmap?

An AI roadmap is the timeline that turns an AI strategy into work. It says which opportunities run in which order, who owns each one, what they will be measured on and when each will be judged. It sits after the strategy, which decides where AI pays and what it may touch.

People often confuse the two. A strategy is a set of decisions. The roadmap puts those decisions on a calendar, so the team knows what happens on Monday. The guide to building an AI strategy covers the decisions themselves.

Plenty of businesses have neither. According to Microsoft and LinkedIn's 2024 Work Trend Index, 60% of leaders say their company lacks a vision and plan for AI. Meanwhile staff use AI anyway, on their own terms. A roadmap is how a leadership team takes that energy and points it somewhere.

What goes into an AI roadmap template?

A complete template has six parts: the ranked shortlist, the rules, the owners, the 90-day plan, the 12-month view and the review rhythm. Each fits on a page or less. Copy the headings below into a shared document, fill them in with your leadership team, and keep it where everyone can see it.

Most templates online skip the rules and the review rhythm. However, those two parts decide whether the roadmap survives contact with real work. Without rules, every workflow argues the same questions again. Without a rhythm, the plan goes stale by month three.

How do you fill in the opportunity shortlist?

Write one short card per opportunity, then rank the cards. Each card names the job, who does it now, how often, the current measure, the knowledge it depends on, the risk if AI gets it wrong and the owner. Rank by return, effort and risk, and put the most frequent, checkable job first.

Halcyra is a fictional company, used here to show the format. Its report scores well because it is weekly, defined and easy to check. The guide to the weekly workflow test explains the scoring.

Problem definition matters most at this stage. According to RAND's 2024 study, the most common root cause of AI project failure is that the problem is misunderstood or miscommunicated. A card that cannot name its current measure is not ready to rank.

What does the first 90 days look like?

The first 90 days run the top one or two workflows properly. In weeks one and two, set rules and baselines. Weeks three to six load the knowledge each workflow needs and build it. Then run it on real work in weeks seven to ten, training the people using it. Measure and decide in weeks eleven to thirteen.

Training belongs inside the plan, not after it. According to BCG's 2025 AI at Work survey, only one-third of employees say they have been properly trained. It also found regular use is far higher among people with at least five hours of training.

The decision at week thirteen is the point of the whole exercise. A workflow that never reaches a decision drifts into pilot purgatory: still running, still costing, never judged.

How do you lay out the next 12 months?

Lay out the next 12 months by quarter. Quarter one is the 90-day plan. Quarters two to four take the rest of the shortlist in rank order, one or two workflows per quarter, each starting only once the last is running. Keep later quarters loose: name the workflow and owner, not the weekly tasks.

The condition column does the real work. It stops the roadmap promising dates the business cannot keep. Each new workflow also reuses what the earlier ones built: the rules, the shared knowledge and the habit of approval.

Be wary of judging too early. MIT NANDA's 2025 study counted an AI project as a success only when it showed measurable results six months after the pilot. So a quarterly roadmap should look back two quarters, not one.

How do you keep the roadmap current?

Check it monthly and rewrite it quarterly. The monthly check asks whether each running workflow is on track against its baseline. The quarterly rewrite uses the latest results to re-rank the shortlist. Only the AI owner changes the roadmap, and every change is written down with its reason.

Three habits keep it honest:

1. Results move the ranking. A workflow that beat its baseline may earn a second phase; one that missed drops down.
2. New ideas join the shortlist. They are scored on the same card before they compete for a slot.
3. Stopped work is recorded. What was stopped, and why, saves the next person from trying it again.

Skein's strategy diagnostic runs listen, map, rank and plan, and ends in a ranked shortlist of three to five opportunities, a 90-day plan and a 12-month roadmap; the AI strategy page shows how it runs. That suits a growing B2B business with more AI ideas than time to judge them. A team that already knows its first weekly workflow can fill in this template itself.

Frequently asked questions

What should an AI roadmap include?

It should include a ranked shortlist of three to five opportunities, the rules for data and approval, and named owners. Add a detailed plan for the first 90 days, a quarterly view of the following 12 months and a review rhythm. Each opportunity needs a baseline measure and a decision date.

How do you create an AI roadmap?

Start with your AI strategy's ranked opportunities. Put the top one or two into a 90-day plan with owners, baselines and a decision at the end. Place the rest by quarter, each starting once the previous workflow is running, then review monthly and rewrite quarterly.

What is the difference between an AI strategy and an AI roadmap?

The strategy decides where AI will pay, what it may touch, who owns it and which tools to use. The roadmap turns those decisions into a timeline, with workflows placed in order, owners named and decision dates set. Strategy comes first; the roadmap schedules it.

How long should an AI roadmap cover?

Most growing businesses plan the first 90 days in detail and the following 12 months by quarter. Planning further ahead rarely helps, because AI tools and the business both change quickly. Each quarter's results should reshape the next, so the roadmap is rewritten rather than followed rigidly.

Who should own an AI roadmap?

One senior person should own it, usually a founder or operations lead with authority to change priorities. Each workflow on it also needs its own owner. Australia's National AI Centre recommends a senior leader as overall AI governance owner, which fits the same role naturally.

Related reading