Vendor evaluation · billed hourly
You don't know what you don't know. Especially across the table from a software vendor.
Vendors know their product better than you ever will, and the demo is built to show its best side. Before you sign, renew or expand, I go through the proposal and the demo with you, find the questions nobody has asked yet, and put them in plain language. Then I check that you're solving your actual problem the simplest way available. Sometimes that's the vendor. Sometimes it's a setting in a tool you already pay for.
Hourly, with the hours estimated on the call before you commit
What I look at
Four questions behind every good software decision.
Features are the easy part to compare. These are the parts that decide whether you're happy a year later.
I map the problem first, then the product. Most mismatches are visible before the demo ends, if you know where to look.
A file format shifts, a rule changes, a person leaves. Does the product stop and tell you, or quietly produce a wrong answer? Who fixes it, and how fast?
Time savings and ROI claims, checked against what the product actually touches. If the baseline counts work the tool doesn't do, the savings shrink.
Who owns your data and the logic built on it, how you get both out, what leaving costs, and whose AI model sits inside the product.
A real example
The questions an AI vendor hadn't been asked yet.
A private equity firm was deciding whether to expand an AI vendor's month-end reconciliation tool across dozens of entities. As part of a case study built on that live decision, I reviewed the vendor's proposal, its product recordings and the operating procedures. The product was well built, and I said so. But nothing in the material showed how it would hold up under change. I wrote down the questions I'd want answered before recommending the expansion. The firm took them straight back to the vendor.
- Testing "Can you replay six months of past closes through the current version and show us what changed?"
- Failure "When a source file's format changes, does the run stop, or does it produce a quietly wrong result?"
- Audit trail "If an auditor asks which version of the rules produced March's numbers, what's the answer?"
- The math "Does your time-savings baseline include work your product doesn't touch?"
- Design "This system has a free API. Why log in through the screen instead?"
- AI "Which AI model is inside, from which provider, and does it touch the arithmetic?"
Not gotchas. Most of these have good answers. If the vendor gives them, your confidence should go up. If they can't, you've learned that before signing instead of after.
How it works
A free call, a review, then questions you can use.
Free 30-minute call
What you're evaluating, what's at stake and when you need to decide. I'll tell you whether I can help and estimate the hours before you commit to anything.
I review
The proposal, the demo or recordings, the contract terms and how you work today. If it helps, I'll sit in on a vendor call and listen.
You get the questions and my read
Written questions ranked by importance, what a good and a bad answer sound like, and my recommendation: buy, negotiate, wait, or solve it another way.
Fair questions
What people ask before booking
What kinds of software?
Anything that runs part of your operations: CRMs, scheduling and booking systems, automation and AI tools, finance and back-office platforms, industry-specific software. The questions change by category. The habit of asking them doesn't.
Are you going to try to sell me a build instead?
No. If the vendor is the right answer, I'll say so. If building something yourself would be cheaper, I'll show you the math, and you're free to hire anyone to do it, or nobody.
We already signed. Is it too late?
Not at all. Renewals and expansions are often the better moment to ask, because you have real usage to point to and the vendor wants to keep you.
Evaluating a vendor right now?
Book a free 30-minute call. No pitch.
Tell me what you're looking at and when you need to decide. I'll tell you whether I can help and roughly how many hours it would take.