Who is behind Diaglo
I'm Ionuț, and I've worked in automotive engineering for 15 years.
I started building Diaglo after watching the AI field advance week after week. I used my knowledge of vehicles to validate how the system guides the model through a constrained, multi-step analytical process that prioritizes causal reasoning, mechanical relationships, and failure probability rather than surface-level pattern matching or keyword search like an ordinary chatbot.
Diaglo was built and refined by testing it against real-world diagnostic cases and iterating until the conclusions matched confirmed repairs and known failure mechanisms.
What makes it different
Instead of handing you a long, generic list of "could be", Diaglo runs every case through a strict engineering workflow:
- 01
Starts from your exact car: make, engine, year, mileage — which we identify in the database so we can extract all the technical information needed.
- 02
Looks for the mechanism: which physical process can produce the exact symptom described. If the AI proposes a hypothesis, it cannot move forward until the physical processes and mechanical logic are verified.
- 03
Compares the causes: by probability and rules out the ones that do not fit, with clear reasoning.
- 04
Checks real public data: part numbers from the parts catalogue, official recall campaigns, and official technical service bulletins.
- 05
Tells you what to test: so you can confirm the diagnosis before replacing any part.
Includes the mechanic negotiation script: Alongside the technical analysis, the report gives you a practical shop conversation and negotiation script: the exact questions to ask your mechanic, which checks to request before approving expensive part replacements, and what parts and labour budget to expect.
After every report, I follow up to see whether the issue was resolved at the shop and which part turned out to be at fault, continually improving the system.
The role of AI and the human in the loop
I truly believe in the future of AI — not as a replacement for people, but by integrating them as the human in the loop. The goal was never to automate repairs, but to reproduce structured human reasoning and narrow down the most likely root causes so you and your mechanic have a clear starting point.
What Diaglo is NOT
- Not an OBD scanner. You do not need any hardware to use it; if you already have fault codes from a scan, you can include them.
- Not a replacement for a physical shop inspection. Every technical conclusion must be physically confirmed on the vehicle by a mechanic before parts are replaced.
- Never a reason to drive an unsafe car. If the issue involves brakes, steering, fuel leaks, a burning smell, or smoke, Diaglo uses a safety score and tells you clearly when to stop driving the car.
I'd love to hear your feedback. Write to me anytime at contact@diaglo.ai.
— Ionuț Prunes, founder of Diaglo