You don’t know what you don’t know. I do.
Part car guy, part software developer. I give you clarity — what to look for, what to ask, and where it breaks. Here’s the kind of question I bring to the table that a vendor hopes you never ask.
The vendor gut-check
A demo always works. I ask the question that shows whether the integration holds up at 2 a.m. when your DMS is slow — because I’ve shipped the code that has to survive it.
The hallucination guardrails
These models are known to hallucinate exactly this kind of context. There are three guardrails that keep one honest. I know in one question whether a vendor has them — or whether their “AI” will confidently make things up to your customers.
The architecture underneath
The slick interface isn’t the product. The architecture under it decides whether this works on your floor or falls over at scale. That’s the layer I surface — the part you can’t see in a sales deck.
I understand what you’re trying to add, build, or implement. Let me be the segue between the idea and the execution.
The vendor treadmill
- Ten vendors, ten dashboards, zero strategy tying them together
- Salespeople who’ve never desked a deal telling you what your store needs
- Long contracts for tools that look great in a demo and stall on the floor
- “AI” that’s really just a chatbot — and your team quietly stops using it
- No one accountable for the outcome, only the invoice
One person on your side
- A single advisor who’s actually run dealerships — front and back
- Vendor-neutral counsel — I tell you what to skip, not just what to buy
- A real dev team that can build what no vendor sells you
- A blueprint mapped to your store, not a generic playbook
- Accountable for results — measured the way dealers measure: PVR, gross, closing ratio
You know your store. I know the stack.
A salesperson can read you a spec sheet. They can’t tell you which model breaks on dealer data, which “AI” is a glorified chatbot, or what a real integration with your CRM actually costs to maintain. I can — because I’ve built it.
The AI stack
- Which models fit which jobs — and which are overkill you'll overpay for
- Agents vs. chatbots vs. automations: what each actually does on a floor
- Where your dealer data lives, and how to use it without leaking it
- Build vs. buy vs. fine-tune — the real cost and lock-in of each path
Frameworks & integration
- CRM, DMS, and tool integrations that hold up — not demo-ware
- How an AI workflow plugs into DealerSocket, eLeads, R&R, Reynolds, ADP
- RouteOne / DealerTrack realities for finance and compliance flows
- What "no-code" really means — and where it quietly falls apart
Vendor reality
- The questions that expose whether a vendor's "AI" is real
- Contract terms, data ownership, and exit clauses that protect you
- What's worth paying for vs. what you already own and aren't using
- When the honest answer is "don't buy this" — and I'll say it
The gotchas
- Adoption failure — the tool works, your people don't use it
- Hidden recurring costs, per-seat creep, and usage-based surprises
- Hallucination & guardrails: where AI confidently gets it wrong
- Compliance landmines in finance, advertising, and customer data
Enterprise scale
- What works at a single point vs. across a 20-store group
- Standardizing process so AI is consistent store-to-store
- Roll-out sequencing so you're not betting the group on one pilot
- Reporting that rolls up the way ownership actually wants to see it
The numbers that matter
- Tying AI spend to PVR, gross, closing ratio — not vanity metrics
- Lead-to-show, show-to-close, and where AI actually moves them
- Fixed ops, BDC, and finance impact — not just the showroom
- ROI you can defend to a dealer principal, in dealer language
I know the gotchas — and where the code fails.
I can spot the failures in the code and understand the architecture underneath it. I know what will actually hold up and what won’t — before you sign, not after.