AI strategy consulting decides which AI work gets funded, in what order, and what gets cancelled. Hiring someone to build an automation is a different job: that work ships one specific system after the decision already exists. Most owners call me for the second thing when the first thing is what they are actually missing.
AI strategy consulting is the work of picking a small number of AI bets, naming a human owner for each, and putting a calendar date on the first one. If you already own AI tools and still cannot point to one revenue number or one hour that changed, your problem is not tooling. It is an unmade decision.
Key takeaways
- Strategy chooses and sequences the bets; automation builds the one that won.
- A usable AI strategy fits on a few pages and names what you will not build this year.
- The right recommendation changes with your employee count, not with industry excitement.
- Judge an engagement by its kill list and by whether the first funded bet has an owner and a date.
What is AI strategy consulting, and how is it different from hiring someone to build the automation?
The adoption numbers show the gap clearly. AI agent deployment sits in the single digits across nearly every business function even though 88 percent of surveyed organizations report using AI somewhere (Stanford HAI, 2026 AI Index). That distance between owning tools and changing operations is the entire market for strategy work.
Of the 30 pages competing for this term, 11 treat strategy as a subtopic and 5 treat services as one, with headings like "What is Marketing Strategy Consulting?" and "What are Digital Marketing Consulting Services?" That is why the phrase reads like a brochure category instead of a defined deliverable.
The strategy job: choosing, sequencing, and killing
Strategy work compares candidates against each other and refuses most of them. I have built 29 AI systems for my own consultancy, which is also how you learn which ones should never have been built in the first place.
The build job: shipping one working system
Build work is narrower and more honest about it: one workflow, one integration, one measurable outcome. It assumes somebody has already chosen the target.
Where the two get sold as the same thing
Trouble starts when a build quote arrives dressed as a plan. If the proposal names a tool before it names a decision, you are buying construction and calling it direction. My breakdown of an AI consultant versus an AI agency covers how those two offers get bundled.
What should you be holding in your hands when an AI strategy engagement ends?
A usable AI strategy fits on a few pages. Not a deck, not a maturity model, just the decisions:
- The two or three workflows that get funded this year
- The named human accountable for each one
- The systems of record each workflow touches
- The single measure that says it worked
- The list of things you have agreed not to build
A system of record is the software where a fact officially lives, such as your CRM, your scheduling tool, or your billing platform.
A ranked list of bets with a named owner on each
Ranked, not grouped. A tie means nobody chose, and a bet without a name attached belongs to no one by the second week.
A kill list, written down
A kill list is the set of AI ideas you are formally declining this year, with a one-line reason each. Porter made the same point about strategy generally: it requires choosing what not to do (Harvard Business Review). Almost nobody writes this part down, which is why the same abandoned ideas resurface every quarter.
A sequence tied to calendar dates, not phases
"Phase two" is not a date. For example, with a home services client the sequence starts where the owner's own week disappears, usually quoting and scheduling, not at whatever demo looked most impressive.
How do you decide which AI bets are worth funding this year?
A bet earns funding when it attaches to a measurable change in how your customers behave or how your costs run.
Pick against a market shift you can measure, not a capability demo
For instance, search behavior has moved. Pew Research Center found users who saw an AI summary clicked a result 8 percent of the time, compared with 15 percent for users who did not see one, and Ahrefs reports that the presence of an AI Overview correlates with a 58 percent lower average clickthrough rate for the top-ranking page. A service business can plan around a shift that size.
Compare that with spending driven by enthusiasm. Generative AI was already used in 11 percent of marketing activities while digital spending grew 11.1 percent in the year covered by the Fall 2024 CMO Survey (Duke Fuqua), which means plenty of money went into AI usage nobody had decided anything about.
Test each candidate against revenue, risk, and who owns the failure
Three questions, asked out loud: what revenue or what hour does this move, who owns it when it breaks, and what are we stopping to make room for it?
Why most AI pilots are strategy failures, not technology failures
The model usually worked. The pilot died because no one was accountable for the output and nothing was cancelled to fund it.
Does the right AI strategy change depending on how big your company is?
Yes, and the size bands are not subtle. The Census Bureau puts AI use at 19.8 percent of US businesses as of May 2026, with fewer than 20 percent of firms with four or fewer employees using it at all, while 32 percent of firms with 100 to 249 employees and 37 percent of firms with 250 or more do.
Under 20 employees: one workflow, one owner, no committee
Pick one workflow and skip the governance layer entirely. Because fewer than 20 percent of the smallest firms use AI, the first-mover position in a local market is still unclaimed.
This is the band where a second opinion pays for itself, which is most of what my guide to hiring an AI consultant is about.
250 and up: governance becomes the actual bottleneck
Approval paths and data access decide your timeline, not model choice. Across the industries I work in, government, hospitality, medical, franchising, and home services, the regulated ones invert the order: the kill list gets written first.
Why do so few AI strategy consulting pages publish a number you can budget against?
Across the 30 pages competing for this keyword, 12 publish no figure at all and 5 give ranges only, leaving 13 with at least one exact currency or percent figure. Their outbound links point mostly at YouTube, which 10 pages cite, and at LinkedIn, Twitter, and Facebook, each cited by 8. Zero of the 30 cite a .gov or .edu source.
What the pricing silence looks like across the field
Unsourced numbers and hidden pricing are a buying signal on their own. You are being asked to trust a category, not a quote.
Two numbers that do anchor the decision
The real comparison is an internal hire. The Bureau of Labor Statistics puts median pay for advertising, promotions, and marketing managers at $165,780 a year. Then there is run rate after a strategy becomes a system: Intercom prices its Fin agent at $0.99 per resolved outcome. Money is tight either way, since marketing spend fell to 7.7 percent of company revenue in the Fall 2024 CMO Survey, the lowest in more than three years (Duke Fuqua), so the budget has to come from something that stops.
What to ask for in writing before you sign
Require three things on paper: a fixed scope, a dated first deliverable, and an estimate of ongoing per-use cost. If you want more detail on how I work as an AI consultant in Las Vegas, you can read a little about me as well.
Short answers on AI strategy consulting
Can I skip strategy and just start building?
Yes, when exactly one workflow is obviously broken and you can live with the result. It stops being reasonable the moment two departments want different things from the same budget.
How long should an AI strategy engagement take?
Weeks, not quarters. If the proposal spans a quarter before anything ships, ask what ships in week three and who owns it.
Who from my team has to be in the room?
Whoever owns the revenue number and whoever will be blamed when the system is wrong. Everyone else can read the summary.
Is it too late to get an advantage from this?
Not at the small end. With fewer than 20 percent of firms with four or fewer employees using AI, most local markets still have no AI-fluent competitor.
How do I know the strategy worked?
The test is whether a task left someone's week, not whether a tool got installed. Ask the owner what they stopped doing.
A good AI strategy is judged by what it cancels
With 88 percent of organizations already using AI somewhere while deployment stays in the single digits (Stanford HAI), access is not the scarce thing. The scarce thing is a decision about which few uses matter, plus the discipline to stop the rest. So judge an AI strategy consulting engagement by the clarity of its kill list and by whether the first funded bet has a named owner and a date on the calendar. If you want to start there, book a 30-minute consultation and we will map where your hours are going before anyone proposes building anything.