Three months ago, a dental group owner in Scottsdale told me: "I searched 'AI business consultant' because I know I need one. But I have no idea what I'm buying. Is it a guy who installs ChatGPT on my laptop? Is it a McKinsey engagement that costs $200K? I genuinely don't know."
Fair question. The term "AI consultant" covers everything from freelancers charging $50/hour to build Zapier flows to Big Four firms selling year-long transformation programs with 14-person teams. The market is confusing on purpose, because confusion keeps prices high.
Here's what the engagement actually looks like when you hire someone who builds things instead of someone who writes about building things.
Week 1: The diagnosis
A real AI consultant doesn't start by talking about AI. They start by shadowing your operation. Either physically walking your office or spending 3-4 hours on calls with the people who actually do the work — the office manager, the sales lead, the ops director. Not the owner. The owner's view of the operation is optimistic by default.
The goal is mapping every process that touches manual effort: how leads come in, how quotes go out, how jobs get scheduled, how invoices get sent, how follow-ups happen (or don't). The consultant is building a heat map of where time, money, and revenue are leaking.
I spent a full day at a property management firm in Jacksonville last November. Their owner swore their maintenance request system was tight. Two hours in, I watched a tenant call the front desk, leave a voicemail, wait 4 hours for a callback, realize nobody called, and call back again angry. That one loop was happening 30+ times a week. The owner didn't know because the front desk never told him.
By the end of Week 1, you should have a prioritized list of 5-8 automation candidates, each scored by estimated time savings and revenue impact. If your consultant delivers a 90-page PowerPoint instead of a actionable priority list, that's your first red flag.
Weeks 2-4: The build
This is where the real consultants separate from the advisors. An advisor hands you the list and says "good luck." A builder picks the top 1-2 items and starts engineering the solution.
Operational blind spots, AI-driven ROI breakdowns, and industry deep dives — delivered to operators who build.
The first build should be fast — days, not weeks. It should target the single most painful bottleneck identified in Week 1. For that property management firm, it was an automated maintenance request pipeline: tenant texts a dedicated number, describes the issue, the system categorizes it by urgency, auto-assigns the right vendor, and sends the tenant a confirmation with an ETA. The front desk didn't touch it unless the system flagged something unusual.
Build time: 6 days. Result: 14 hours per week freed for the property managers, and tenant satisfaction scores jumped from 3.2 to 4.1 on their quarterly survey.
The build phase also includes training your team. I won't sugarcoat this — staff resistance is real, and it shows up differently in every company. At a minimum, you need 30 minutes with each person whose workflow changes. Not a group meeting. One-on-one. Show them exactly what changes, what doesn't, and where their new time goes.
Weeks 4-8: The stack
Once the first automation is running and your team trusts it, you move to the next item on the priority list. Then the next. Each build gets faster because the infrastructure from the first one (API connections, data pipelines, notification systems) carries forward.
A typical 8-week engagement delivers 3-4 fully operational automations. By the end, the client usually has a clear picture of what their operation looks like with AI baked in, and they can decide whether to keep building or maintain what they have.
What it actually costs
I'm going to be more transparent about pricing than most consultants are comfortable with, because I think the ambiguity in this market is actively harmful.
For a small business ($1M-$5M revenue), a focused 4-8 week engagement with actual builds typically runs $3,000-$8,000 per month. That's for diagnosis, custom automation builds, staff training, and 30-day support after launch.
For mid-market ($5M-$15M), you're looking at $8,000-$15,000 per month because the systems are more complex, there are more integrations, and more stakeholders need to be managed.
If someone quotes you $50,000+ for an "AI readiness assessment" that produces a deck and no working software, you're paying for a very expensive opinion. Deloitte's 2024 report on mid-market AI adoption found that 62% of companies that hired strategy-only AI consultants never implemented the recommendations. Strategy without execution is shelf-ware.
I'll be upfront about one thing I'm still figuring out: ongoing maintenance pricing. Some automations run for months without needing a tweak. Others need constant adjustment as the client's processes evolve. We're still calibrating the right model for long-term support, and I'd rather admit that than pretend we have it perfectly figured out.
The red flags
You're hiring the wrong person if they:
- Talk about "digital transformation" for more than 30 seconds without naming a specific process they'd change
- Won't give you a ballpark cost within the first conversation
- Have never built a working automation — only advised others to build one
- Propose a 6-month "discovery phase" before any software gets deployed
- Can't show you a before-and-after metric from a previous client
The right consultant talks in hours saved, dollars recovered, and errors eliminated. Not in "strategic vision" or "organizational readiness maturity models."
Do you actually need one?
Honestly? Maybe not. If your business is under $500K in revenue with a team of 3 people, you probably don't need a consultant. You need a good Zapier tutorial and a weekend. The math doesn't justify the spend until you're dealing with repeatable, high-frequency workflows that multiple people touch.
But if you're north of $1M, you have 6+ employees, and your team spends more than 10 hours a week on tasks that are purely mechanical — data entry, follow-ups, report building, scheduling — a consultant pays for themselves in the first month. Not the second. The first.
Curious what the priority list looks like for your operation? Our free assessment gives you the exact breakdown — what to automate, in what order, and the projected savings. Three minutes. If we're not the right fit, we'll tell you that straight.