The AI questions you're asking are the right ones. Here's how to press further.
See how Revizto fits your workflow
Key takeaway: If you've stopped asking "what does your AI do" and started asking "what does it actually change," you're already ahead of where most buying conversations were two years ago — and that instinct is worth pushing further before you commit to anything.
A year or two ago, most conversations I had with customers about AI were broad and theoretical. People were curious, a little skeptical, and mostly asking what it meant for them without a specific problem attached. That's changed. Now the questions are practical: will this reduce admin, will it support a real decision, will it help my team manage project information better. If that's the question you're already asking, you're not being difficult. You're asking exactly what you should be.
"[My customers] are less impressed by the word AI now, and more interested in whether it solves something real."
Gustavo de Bardeci
Three questions worth asking about AI inside your project
Based on what separates a real shift from a talking point, here's where I'd push if I were sitting on your side of the table:
- Does this change a decision, or just describe one? A lot of AI in construction software still summarizes information you already had. That's useful, but it's not the same as changing what you'd actually do differently on a project.
- Would this survive contact with your busiest week? AI that only works when someone has time to set it up carefully doesn't hold up on a live project. Ask what it looks like when your team is already stretched.
- Is the salesperson talking about your workflow, or their roadmap? If every answer routes back to what's coming next quarter rather than what works today, that's worth noticing.
Where AI genuinely helps — and where your tech vendor should be specific
AI is legitimately reshaping research, account planning, summarizing information, and drafting — inside vendor teams and, increasingly, inside AEC teams too. That's real, and it's useful. What it hasn't replaced is the harder part: someone who understands your project context well enough to know what you actually need versus what you're asking for on the surface.
That matters for how you evaluate a vendor's team, not just their product. Once everyone has access to similar AI tooling, the differentiator isn't who has it first — it's who can apply it to your specific situation without you having to translate for them. If a vendor's people can't do that translation themselves, the AI features underneath won't make up the difference.
What to expect from a vendor relationship that's actually working
A healthy relationship with a technology vendor should include some friction — not constant agreement. If every question you ask gets a yes, that's usually a sign you're not getting a real answer about fit. The vendors worth working with will tell you when something isn't the right match for how your team actually operates, even if that costs them the deal in the short term.
That's also where feedback should be structured, not just collected. If you're raising the same issue repeatedly and it disappears into a support queue, that's worth asking about directly — the best vendor teams route recurring patterns back to product decisions, not just individual tickets.
If you've already started asking that question of every AI claim that lands on your desk, trust that instinct. It's the right one.
If you want to put that same question to us directly, talk to our team about how AI works inside Revizto — we'd rather you ask it than assume the answer.
