The problem is real, and everyone feels it
Talk to anyone running a car dealership and you'll hear the same thing: their teams are drowning. Not from lack of effort — from constant interruptions. Email notifications. Phone calls. Customers walking in. DMS updates. Service requests. The day fragments into a hundred tiny context switches, and by closing time, everyone's exhausted but nothing important got done.
Vienna AI Company built LISA to address exactly this. Their AI assistant handles the routine stuff — triaging emails, pulling documents, scheduling appointments, answering common questions at 3am. One pilot customer said it best: training a new salesperson to do what LISA does takes six months. LISA does it in five minutes.
The institutional knowledge piece is particularly clever. LISA Scribe records meetings, transcribes them, and makes everything searchable. Someone asks "What did the customer say about their trade-in preference last month?" and you just ask LISA instead of scrolling through email or hoping someone took good notes. This isn't flashy, but it's the kind of thing that prevents the small failures that make dealerships look disorganized.
The skeptic problem (and the solution hiding in plain sight)
Here's where it gets interesting. LISA is expanding at a controlled pace — two new dealerships per month — and hitting the same pattern repeatedly. Office staff love it immediately. Sales and service teams? Skeptical. They see it as software for desk jobs, not relevant to actual car work.
Then someone convinces them to try it once. Upload a technical PDF, ask a question, get a professional answer. Instant conversion. The aha-moment is real and repeatable, but it requires someone manually walking the skeptic through it.
This suggests the product works beautifully, but discovery is broken. At two dealerships per month, you can afford to personally onboard every user. At twenty per month, you can't. The opportunity here is building a self-service flow that surfaces that aha-moment without human handholding. A five-minute guided tour with pre-loaded sample documents — service manuals, parts catalogs — that walks users through uploading, asking, and seeing results. Make the quick win unavoidable, and adoption stops being a people problem.
The integration gap nobody notices until it matters
LISA captures incredibly valuable context — customer preferences, conversation history, the kind of details that turn a transaction into a relationship. But right now, that data lives in LISA while customer records live in the dealership management system. When a salesperson opens a customer file, they don't see what LISA knows unless they remember to check separately.
This creates exactly the kind of fragmentation LISA is supposed to eliminate. The logical next step is bidirectional sync with the DMS. When LISA learns something about a customer, it should write back to their record automatically. When someone opens a DMS file, they should see LISA-captured context right there. Start with one DMS (CROSS 3 appears in several sources), add conflict detection so nothing gets overwritten accidentally, and suddenly LISA becomes the system of record for customer intelligence instead of just another tool.
The company has strong early traction — 25+ customers, major Austrian automotive retailers, pilots booked through Q3 2026. They're clearly solving a real problem, and their controlled expansion model keeps quality high. The opportunities here aren't fixes — they're about removing friction from what's already working. Make the aha-moment self-service, eliminate the DMS gap, and you accelerate adoption without changing what makes the product valuable.
If you're curious about the full breakdown — theme patterns, evidence clusters, the whole analytical view — we built it with Mimir and it's worth exploring how the pieces connect.