Company profile · 4 min read
Bernard: AI front and back office for appliance repair
Bernard connects appliance-repair calls, scheduling, dispatch rules, diagnosis and parts preparation in one workflow.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
Bernard is building an AI front and back office for appliance repair companies. The useful buyer fit is a servicer whose phones, warranty rules, schedules and parts knowledge live in separate systems—and whose first truck roll too often becomes a diagnostic visit instead of a completed repair.
What it does
The current Bernard homepage describes one workflow for calls, scheduling and parts. It says the system answers business-hours, overflow, after-hours and voicemail calls; identifies COD, warranty, redo, status and escalation requests; and applies the servicer’s own pricing, coverage, fee, service-area and escalation rules. It can also read live availability and write creates, reschedules and cancellations back to the field-service system.
The product’s distinctive layer is Pre-ID. Bernard says it combines the exact appliance and symptoms with repair history, similar completed visits, manufacturer guidance, parts data and the servicer’s operating knowledge. The resulting diagnosis, ranked parts and follow-up questions are meant to reach the technician before the visit. The YC launch description says the company is already live with some large authorized service companies; that is a company-reported status, not an independently verified customer result.
Why I’d look closer
The advantage is operational continuity. A call-center answer is only useful if it produces a bookable appointment, a plausible technician match and the parts context needed to finish the job. Bernard’s YC profile identifies Dan Katzman’s service-heavy strategy and operations background at Nava Benefits and David Goodfellow’s engineering leadership in fraud and geo-compliance at Radar Labs. Those backgrounds fit a rules-and-workflow product better than a generic chatbot pitch.
The tradeoff is that diagnosis and scheduling are consequential. Wrong parts, wrong warranty treatment or a booking that the team cannot fulfill create cost and customer frustration. The company’s public pages do not publish pricing, independent first-visit completion data or a detailed error policy.
What I’d ask
Which field-service systems are supported? Can the dispatcher inspect the evidence behind a diagnosis and override ranked parts? How are manufacturer updates, warranty exceptions, technician skills and customer data kept current? What happens when the agent cannot distinguish a repair, redo or escalation?
My editorial take
Shortlist Bernard if appliance-service volume is high enough that missed calls and repeat visits are the real margin problem. Start with one service line and explicit dispatch rules, then compare booked-work quality and parts readiness against the current process. The product is promising because it connects the phone to the truck; that same connection is why every automated change needs a visible audit trail.
Quick facts
| Field | Sourced detail |
|---|---|
| Product | AI call handling, scheduling, dispatch context and repair preparation |
| Buyer | Appliance servicers, manufacturers, retailers, warranty companies and service networks |
| Core workflow | Calls → job classification → scheduling → diagnosis/parts → technician handoff |
| Pricing | Not published in the checked pages |
| Main question | Can Pre-ID improve technician readiness without hiding uncertainty? |
Sources checked
| Source | Checked |
|---|---|
| YC company profile | 2026-09-19 |
| Bernard homepage | 2026-09-19 |
| Bernard YC launch | 2026-09-19 |
Cohort context
Bernard is listed in Summer 2026. In our 2026-09-18 directory snapshot, 119 of 232 listed companies in that cohort have YC’s primary industry label B2B (51.3%). This is a current-directory comparison, not an original intake count or a performance ranking. Nine-cohort dataset.
Public website snapshot
Observed 2026-09-19T16:18:37.795Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.
| Signal | Homepage observation |
|---|---|
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Observed |
| Typed structured data | Observed |
| Docs/developer link | Not observed in this response |
| Pricing link | Not observed in this response |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |
Public observations · Collection method. Missing links here do not establish that a capability or file is absent elsewhere.
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.
First1000 ↗ · X ↗