Signs Your Company Is Ready for an AI Roadmap (and Signs It Isn't)

Not every company that wants an AI plan is ready to act on one. Half the operations we see would benefit more from three months of workflow cleanup than from a shiny roadmap. This post is a short honest checklist for figuring out which side of that line you're on.

Five signs you're ready for a real AI roadmap

1. You have an AI mandate from the top, and a deadline attached

Not a general enthusiasm. A specific expectation from the board or CEO that a plan exists by a specific date. If someone owns the outcome and someone will be held to it, the audit turns into action. If neither exists, the audit turns into a document.

2. Your senior team spends more than two hours a week debating AI without deciding anything

This is the single most reliable indicator we see. When calendar time on AI conversations exceeds calendar time on AI execution by 10 to 1, the operation has run out its ability to solve this internally. That's exactly when an outside diagnostic pays for itself.

3. You can name three workflows that "obviously" need AI

Support triage. Proposal drafting. Invoice reconciliation. Sales research. Meeting notes. If you can name three specific processes without thinking, you have a portfolio to prioritise — which is what a roadmap does. If you can't name any, the readiness gap is deeper and the first exercise is workflow discovery, not AI strategy.

4. Your team has tried at least one AI tool and it didn't stick

Counterintuitive, but this is a good sign. It means the organisation is willing to experiment. The reason a tool didn't stick usually isn't the tool — it was applied to the wrong workflow, or dropped into a team that didn't have adoption support. That's exactly the problem a good roadmap solves.

5. You have budget authority and a rough number in mind

A rough figure — $50K, $250K, $1M — that leadership would sign off on if the plan justified it. Without a ballpark, the audit produces a plan that stalls at approval. With a ballpark, phase-one funding is a small conversation instead of a large one.

Four signs you're not ready yet

1. Your operational processes change every quarter

Automating a moving target burns money. If your top workflows have been restructured in the last six months and are up for another rewrite, the readiness score is low across the board. Stabilise first, automate second.

The good news: stabilisation isn't a year-long project. It's usually one or two SOPs written well, a couple of decisions ratified by the owner, and a short window of "no more changes." Then the roadmap works.

2. You have fewer than 20 employees

Not a hard rule, but the math is honest. AI ROI at that size mostly shows up as time saved for the founder or a small team. That's real, but it's usually addressable with off-the-shelf tools ($200/month of software) instead of a custom roadmap. Save the audit for the moment payroll makes automation matter.

The exception: a small AI-native company building AI as its product. That's a different conversation entirely — it's product engineering, not internal transformation.

3. Your data lives in email threads, screenshots, and someone's head

AI is downstream of data. If the answers to "how many tickets did we handle last quarter" or "what's our average deal size by segment" require Slack archaeology, phase one of any roadmap will be data plumbing, not AI. That's still worth doing — but call it what it is and don't confuse it with an AI strategy.

4. You want AI to do something no team currently does

The single most common failed AI project we see: the CEO wants AI to handle a job the company doesn't yet do at all. AI is good at automating existing patterns, not inventing new ones. If nobody on the team could describe the job step by step, no model can either.

Solve the "who does this today, how do they do it" question first. Then the automation question is real.

The in-between case

Most operations aren't a clean yes or no. They're in the middle — some workflows ready, some not; some data clean, some messy; some leadership aligned, some not. The role of a diagnostic is to sort those things without a six-month strategy retainer.

A useful heuristic: if you can honestly answer yes to at least three of the "ready" signs and no to at least three of the "not ready" signs, the audit will produce a plan. If not, the audit will produce a much more useful outcome — an honest picture of what needs to change before an AI roadmap is worth commissioning.

Where teams get this assessment wrong

Confusing enthusiasm with readiness

Enthusiasm makes you want a roadmap. Readiness makes the roadmap succeed. They're not the same. Enthusiastic companies with unstable processes end up with expensive shelved plans.

Waiting for perfect data

The mirror mistake. Data is never clean enough. If your data is 70% usable and one specific workflow is stable, you can start there. Perfectionism disguised as prudence is why some operations wait years to move.

Assuming size is the main variable

It isn't. A 40-person operation with a strong ops lead and one big obvious workflow is more ready than a 400-person org with warring departments and no data plumbing.

Reading "not ready" as a permanent verdict

Most "not ready" companies become ready in one to three months of specific, unglamorous work — usually a workflow cleanup or a data consolidation. It's not a rejection, it's a sequence.

A short self-assessment

Give each item a score from 0 (not true) to 2 (definitely true):

1. The board or CEO expects an AI plan by a specific date
2. Senior team spends significant time on AI discussions without deciding
3. You can name three workflows that obviously need AI
4. Your team has tried at least one AI tool
5. You have a rough budget number in mind
6. Your top workflows have been stable for at least 6 months
7. Your operational data lives in queryable systems, not email
8. You have 4–5 people who can commit 10–15 hours to a diagnostic

Roughly: 12 or higher and you're ready for a roadmap. 8 to 11 and you'd benefit from a diagnostic that includes a "get ready" phase. Below 8 and the honest answer is that AI isn't the first problem to solve.

What "getting ready" actually means

When the score comes back low, the interventions are usually specific and short:

• Document the top 3 workflows step by step, in one sitting
• Consolidate operational data into one system, even imperfectly
• Assign a single owner for the AI decision (not a committee)
• Set a real deadline for a plan (not "sometime this year")

Each of these is a two-to-four-week project. Done together, they take a "not yet" and turn it into a "ready" inside a quarter.

Frequently asked questions

Can a small company benefit from an AI roadmap?
Below 20 people, usually no. The right move is buying two or three good AI tools and configuring them well. Above 20 people, the math starts to work — especially if there's one clear operational workflow that costs meaningful hours.

Does our data need to be clean before we start?
Not entirely. Data needs to be usable for at least one target workflow. If none of your data is usable, phase one is data plumbing. That's a valid first phase, and it's often faster than expected.

What if leadership isn't aligned on what "AI" means?
That's a common starting point. A diagnostic is often the alignment exercise — the process of scoring opportunities together tends to end the terminology debate faster than a strategy offsite ever will.

How long does "getting ready" take?
Usually one quarter of specific, unglamorous work: workflow documentation, data consolidation, and one owner assigned to the AI decision. It's not a year.

Where to go next

If the self-assessment came back high, an AI Audit is the fastest way to translate readiness into a plan. If it came back mid, the audit still works, but expect one "get ready" phase in the roadmap. If it came back low, our post on what an AI audit actually looks like covers what to prepare before commissioning one.