Compliance Reviewer
LOC proves the work happened. Compliance Reviewer proves it was done right.
Trained AI QA agents read the photo evidence your crews submit and check it against your acceptance checklist, field by field: right equipment, right label, right reinstatement. Every install gets reviewed, not a sample, and verdicts land in seconds with the reasoning in writing.
The problem
Thousands of install photos. Who is actually checking them?
Every operator runs its own QA standard, and every build partner submits evidence against it: thousands of photos a week across aerial, buried, no-access and frozen-ground installs. Manual review cannot keep pace with that queue, so most programs check a sample.
That means a low percentage of the work ever gets reviewed. The rest ships on trust, and the problems the sample missed come back later as rework, disputes and fault tickets, when the crew is long gone and the trench is closed.
Compliance Reviewer reviews every submission. QA agents trained to your standards do the first pass on every photo, your reviewers keep control of the verdicts, and vendors get a clear, disputable reason instead of a black-box rejection.
How it works
From the crew's phone to the vendor scorecard.
-
You set the standard
Configure the photos every install must include and the pass criteria for each field. Your checklist, not a template.
-
Crew submits the evidence
Crews capture the required photo set in the field; incomplete submissions do not slip through.
-
Agents check every field
QA agents trained to your standards review every photo. Screenshots, duplicates and re-used photos get caught and reported to you.
-
Verdicts with reasons
Pass, fail or flag in seconds, rationale in writing. Your reviewers override, vendors can dispute.
-
Patterns become action
Scorecards, coaching notices and trends by vendor, state and field.
Trained to your standards, not ours
- You define the required photo set and the acceptance criteria for every install type. Checklists are configuration, not development.
- The agents are trained against your standards and recalibrated with your reviewers' feedback: agree, disagree, re-run the review.
- Four installation types out of the box: aerial, buried, no-access and frozen ground. Pass rules configurable per vendor.
What you get out
Operational output, not just verdicts.
- Vendor performance scorecards, generated automatically from real submission data: who fails what, where, and which way it is trending.
- Coaching notices that put what the crew submitted next to what the standard requires, with practical notes they can act on.
- Integrity reports: screenshots, duplicates and re-used photos, flagged and reported the moment they appear.
- A reviewable record for every install, with verdict, rationale and evidence kept together.
Where it fits
- A standalone quality-control platform for FTTH field installations, powered by trained AI QA agents. It runs on its own today.
- Integrating with LOC as a module, so evidence captured at the asset will flow straight into review.
- Built for the people who own acceptance: quality teams, vendor managers and build supervisors.
The accountability layer
Sampling was a compromise with manual review. When every install can be checked, the sample stops being the standard.
See Compliance Reviewer run against your own acceptance checklist.
Book a demo