Table of Contents
  1. The one question that decides it
  2. What each type of engagement actually delivers
  3. The two most common mistakes
  4. How long should strategy take?
  5. Which one is right for your situation

The one question that decides it

Before evaluating any AI consultant, answer this:

Can you describe the specific workflow you want to automate in one sentence?
If yes
"We want to automate our lead follow-up sequence."
You need automation consulting — someone to build it. You already have the strategy. Don't pay for someone to tell you what you already know.
If no
"We know AI could help us but we're not sure where to start."
You need strategy first — someone to identify the highest-ROI opportunity and sequence the work. But keep it short. Two to four weeks, not twelve.

That's a useful starting framework. The rest of this post fills in the details — what each type of engagement actually delivers, the mistakes businesses make when they pick wrong, and how to evaluate whether the proposal you've received matches what you actually need.

In practice

Many small businesses that think they need strategy actually need automation. They know the workflow — "we spend 3 hours every week building this report manually" — they just haven't named it as an automation candidate yet. In that case, you may be able to move directly into implementation without a formal strategy phase.

What each type of engagement actually delivers

AI Strategy Consulting
Answers: Where does AI fit in my business?
Maps your workflows against automation potential. Identifies the highest-ROI opportunities. Sequences projects by impact and implementation risk. Defines governance — who decides what gets automated, who reviews outputs. Produces a prioritized roadmap you can execute against — in-house, with a freelancer, or with an implementation partner.
AI Automation Consulting
Answers: How do we build this specific thing?
Takes a specific, defined workflow and builds it into a working automated system. Maps the process, selects the tools, builds the integrations, tests the output, and hands it off running. Includes documentation so someone on your team can maintain it. The deliverable is a working automation, not a plan.

The clearest way to think about it: strategy is for when you have multiple possible AI projects and need to pick the right first move. Automation is for when you know the workflow and need someone to build it.

In practice, the two often happen sequentially. You do a short strategy sprint to identify and prioritize. Then you move into implementation on the first project. The mistake is treating them as separate, long engagements when they don't need to be — especially for a small business where the right answer is usually obvious once someone looks at the workflows.

The two most common mistakes

→
Hiring a strategy consultant when you need a builder
You know you want to automate your lead follow-up. A strategy consultant runs a 12-week discovery process, produces a detailed report recommending you automate your lead follow-up, and invoices accordingly. In some cases, a focused implementation could have delivered a working automation faster and at lower cost. This can reflect a mismatch between advisory and implementation capabilities. Ask any consultant you're evaluating: "Can you build what you recommend, or do I need to hire someone else for that?"
→
Hiring a builder before the process is defined
The opposite mistake. You hire someone to automate your client onboarding process before that process is actually consistent — it varies depending on who handles it, what the client needs, and which day of the week it is. The builder builds something, it breaks on every edge case, and nobody maintains it. The fix is spending time documenting exactly how the process should work before any automation is written. When automations fail, it's often due to inconsistent inputs rather than limitations of the tools themselves.

How long should strategy take?

For a small business: often two to four weeks, depending on scope.

A proper AI strategy engagement for a small business involves reviewing your workflows, assessing your technology stack, identifying automation candidates, estimating ROI for each, and producing a prioritized roadmap. For a business with 5–30 employees, that work typically takes a few weeks rather than months.

If someone is proposing a 12-week strategy engagement before any building starts, it's worth understanding why — possible explanations include broader scope than you anticipated, a different methodology, or uncertainty around what implementation will require. None of those are automatically disqualifying, but they're worth discussing before you commit.

The right strategy engagement for a small business looks like the AI Readiness Assessment or the AI Workflow Audit — fixed scope, fixed price, specific deliverable, two weeks maximum. After that, you should be in implementation, not still in discovery.

A useful benchmark

A two-week strategy sprint can typically identify the best workflow to automate first, estimate ROI, and define risk controls. After that, you should have a clear direction on what to build. If you're still in discovery after several weeks, it's worth checking whether the scope has grown beyond what your situation requires.

Which one is right for your situation

Run through these scenarios — the first one that matches your situation tells you what to do:

"We have no idea where to start with AI."

Strategy first — but keep it short. A two-week readiness assessment will identify your highest-ROI opportunity and tell you exactly what to build first. Don't let strategy drag into months of discovery. By week three you should be building something.

"We know we want to automate X, but we don't know how."

You may be able to move directly to automation consulting. You've already answered the strategic question — you know what to build. Find someone who can build it. That said, a brief scoping call to confirm the workflow is well-defined before starting can save rework later.

"We tried to build an automation and it didn't work."

This one usually means the process wasn't defined before the build started. Before hiring anyone new, spend a week documenting the process exactly — every step, every exception, every edge case. Then bring in a builder. The automation failed because the inputs were inconsistent, not because AI doesn't work.

"We have five different AI ideas and don't know which to pursue."

This is the genuine strategy case. You need someone to evaluate all five against your workflows, team capacity, and existing tools — and tell you which one to build first based on ROI and implementation risk. That's a two-to-four week engagement, not twelve weeks.

"We want ongoing AI automation across multiple workflows over 6–12 months."

This is the Fractional AI Consulting case — strategy and implementation under one engagement, built around execution rather than advisory alone. It's the right fit when you have enough work to justify ongoing engagement rather than a series of one-off projects.

The bottom line: strategy and automation aren't opposites — they're often sequential. Keeping the strategy phase focused and moving into building as soon as you have a clear direction tends to produce better outcomes than extended discovery. The businesses that get the least from AI consulting often spend too long in planning and not enough time building and iterating.