How to choose an AI consultant in San Antonio: 10 questions to ask
Updated August 28, 2026 · by Edgar D. Reyna, Azuryc · Boerne, Texas
The AI consulting market is full of firms that produce beautiful strategy decks and no working software. Print this list, ask every firm you talk to the same ten questions, and compare answers side by side. The builders and the deck-sellers separate fast.
The 10 questions
1. "What are your prices?"
Good answer: numbers, on the spot or on their website.
Bad answer: "it depends, let's schedule a discovery workshop." Price opacity is a sales strategy, and you're the one it's used on.
2. "Show me a system you built that ran in production last week."
Good answer: a live screen, real interface, real workflow — even anonymized.
Bad answer: case-study PDFs and logos. Slides are not systems.
3. "Do I actually need AI for this?"
Good answer: "maybe not — a lot of problems are automation or process problems, and those cost less and break less."
Bad answer: AI is the answer before they've heard the problem.
4. "What happens when the AI is wrong?"
Good answer: human sign-off gates on consequential decisions, audit logs, and a defined escalation path. They can show you one.
Bad answer: accuracy percentages and a shrug.
5. "Who owns the system when you leave?"
Good answer: you do — code, credentials, documentation, and training for your team is part of the engagement.
Bad answer: a proprietary platform with a monthly fee forever and no exit path.
6. "What does Texas law require of me?"
Good answer: a plain-English explanation of TRAIGA and, if you're in healthcare, SB 1188 — proportionate to your size. (Here's ours.)
Bad answer: either a blank stare or a six-figure "compliance program." Both are telling.
7. "What's the smallest thing we could do first?"
Good answer: a scoped audit or a single pilot with a fixed price and a payback estimate.
Bad answer: the transformation roadmap, 18 months, seven figures.
8. "Which parts of my project are automation, and which are AI?"
Good answer: a clear split. Rules and integrations for the predictable work; AI only where judgment over messy inputs pays for itself.
Bad answer: everything is "AI-powered." That usually means they don't know where the line is.
9. "Who exactly will build this?"
Good answer: the person in the room, or named engineers you can meet.
Bad answer: the senior person sells, then the work ships offshore to people you never meet.
10. "What did a project like mine actually pay back?"
Good answer: honest mechanics — hours saved, errors avoided, cycle times cut — with the caveat that your numbers depend on your operation.
Bad answer: "300% ROI guaranteed." Nobody honest guarantees ROI on work they haven't scoped.
Local matters more than logos
For a San Antonio or Texas Hill Country business, a nearby consultant who answers the phone beats a national brand with a rotating bench. You want someone who can sit in your office, watch how the work actually flows, and stay close after launch — because the first month in production is where every system earns its keep.
Start with the diagnosis, not the pitch
Answer 5 questions and get your top 3 fixes — each tagged automation, AI, or process fix, with a payback line. Free, 2 minutes, no call required.
Get your Payback Map →