Robot failure analysis
Deepen AI's trained reviewers go through your robot's failures, find where and why each one happened, and back it with evidence from the logs and video.
Rate in your plan before you pay A plan within 2 working days No sales call
What a failure analysis looks like
Each failure comes back with its cause, the evidence and a tag, so you can see which causes are most common.
- Failure
- Pick failed in cell 4 at 14:02
- Error log said
Grasp timeout.- Cause found
- Perception: wrong targetThe robot reached for the box's shadow, not the box.
- Reason
- Front camera frames 41 to 58 show the grasp point on the shadow. Tagged “perception, shadows”.
- Failure
- Mobile robot stopped in aisle 7
- Error log said
Path blocked.- Cause found
- Map: out of dateThe aisle is clear on camera. The map still shows a pallet that was moved on Monday.
- Reason
- Map version and camera frame attached. Tagged “stale map”.
- Failure
- Collision warning in cell 2
- Error log said
- Collision.
- Cause found
- Confirmed: light contactThe arm touched the bin edge at 4 cm/s, inside your safe-contact limit.
- Reason
- Telemetry and the side camera agree. Tagged “planning, tight clearance”.
- Failure
- Item dropped during place, cell 6
- Error log said
- No error.
- Cause found
- Sent to youThe drop lines up with the conveyor starting early, which your logs don't record. Can you share the conveyor timing?
- Reason
- The case stays open until the cause can be confirmed with evidence.
How we start on your work
No sales call. A plan in your inbox, then a short paid pilot on your real failures.
What we need from you
- Your failure cases: logs, video and telemetry
- Your failure categories, or the questions you need answered
- Access to the viewers and dashboards your engineers use
- 5 minutesTell us the task
What your robots do, how many failures a week, and what you need to know.
- Within 2 working daysGet a plan and a price
Reviewers, hours, the rate, start date and the total price.
- 2 weeksRun the pilot
We train reviewers on your system and categories, analyse real failures and send two weekly reports. Then you decide.
Who analyses failures, and how causes are confirmed
Trained reviewers, working in English. Each one learns your system, your logs and your categories before live work, and every cause is backed by evidence you can check.
Every reviewer's accuracy is measured
Items you've already answered are hidden in the real work, so we know how accurate each person is.
Nobody guesses on unclear cases
The team lead decides, or asks you. Your answer joins the guidelines.
You see the results every week
A weekly report shows work done, QA results and what's improving.
Standard analysis covers manipulation robots, mobile robots and vehicles, from logs, video and telemetry. Analysis that needs engineering judgement on your code is quoted in your plan.
How much does robot failure analysis cost?
Failure analysis is quoted in your plan, because the rate depends on your system and the depth of analysis. You pay for the hours reviewers work, with a team lead, QA checks and a weekly report included.
Set before you pay. Your plan shows the rate, team size, hours and total price.
Starts with a two-week pilot on your real failures, paid upfront, with two weekly reports.
Your data stays protected. We work under NDA with the access you set. Deepen AI holds SOC 2 Type II, ISO 27001, ISO 9001 and TISAX.
Questions
What do we get for each failure?
The cause, its category, the evidence with timestamps and frame numbers, and a tag. The weekly report counts failures by cause, so you see which ones to fix first.
How does this help our engineers?
It gives them a head start. Reviewers do the first pass through logs and video, so engineers start from a confirmed cause and the evidence, not a raw error log.
What if the cause isn't clear?
The case goes to your contact with what's known and what's missing, and stays open until the cause can be confirmed.
Can failures feed our data work?
Yes. A recurring cause can become a collection spec, a curation tag or a benchmark scenario.
How fast are failures analysed?
You set the turnaround you need when we plan the work. We staff the team to meet it, and it's written into your plan with the price.
How is our data protected?
We work under NDA, with the access you set, and you can remove it at any time. Deepen AI holds SOC 2 Type II, ISO 27001, ISO 9001 and TISAX.