Disclaimer: Independent product concept by Kaushal Khodifad. Not a live commercial product.return to portfolio
Disclaimer: Independent product concept by Kaushal Khodifad.

FactoryLens · Natural-language intelligence over factory video

Ask your factory a question, get a grounded, timestamped, video-cited answer.

Safety unlocks the budget. Operations justifies the renewal. Natural language is the wedge.

Working live demowarehouse-sample.mp4
Is anyone missing a hard-hat?
Grounded answer

8 people, all in hi-vis vests, 0 wearing hard-hats - headwear is caps/beanies only.

0:03 0:08 0:12
Try it - ask a video what happened

The thesis

Every factory runs cameras. Almost none of that footage is queryable.

A plant manager who wants to know how many times someone entered the press cell without a lockout last week has two options today: scrub hours of footage by hand, or buy a rigid detector that answers only the one question it was pre-configured for.

FactoryLens turns the camera estate into a question-answering surface: ask a question in plain language, get a grounded answer with clip citations. No detector to configure.

Why now

Three tailwinds converging in 2026

01

The market is large and compounding

Computer vision in manufacturing is ~$7.9B in 2026 at 12.7% CAGR to ~$16.2B by 2032. Edge deployment is already ~47% of CV spend and the fastest-growing slice.

02

The category leader vacated the field

Drishti, the reference brand in manufacturing video analytics, was acquired by Apple in September 2023 and is no longer sellable. Demand is stranded.

03

The interface paradigm shifted

Video foundation models (Twelve Labs Marengo 3.0 + Pegasus, on Amazon Bedrock as of Dec 2025) made natural-language video retrieval production-viable. Factory-safety incumbents still ship dashboards and alerts.

Market opportunity

Sizing the bet

Computer vision in manufacturing~$7.9B (2026)12.7%~$16.2B by 2032Core TAM
AI video analytics (cross-industry)~$5B (2025)22.7%~$17B by 2031Fastest-growing adjacent
Workplace safety (total)~$26.2B (2026)17.2%~$93.3B by 2034Budget pool we tap
AI workplace-safety software~$122.6M (2026)7.4%~$189M by 2032Direct competitor slice
EHS software~$2.5-3.0B (2026)10.7-14.6%~$3.9-10.2BIntegration + co-sell

Manufacturing is the largest CV vertical (~28% of CV spend) and edge deployment leads at ~47% share and the highest growth. That validates an edge-first architecture, not cloud-streaming.

The AI workplace-safety niche is small (~$122M) but sits inside a $26B+ safety budget growing 17%. The play is not to fight for the niche; it is to expand what video can do and pull from the larger safety + operations budgets.

Provenance legend- how to read every number on this page
  • Third-party estimate

    Market sizes: directional 2026 figures (Research and Markets, Mordor Intelligence, Genetec/Intellisee) - for sizing, not precise claims.

  • Publicly reported

    Drishti → Apple (Sept 2023) and Twelve Labs Marengo 3.0 + Pegasus on Amazon Bedrock (Dec 2025); competitor funding from public trackers.

  • Incumbent-reported, not our data

    The ~77% injury-reduction / ~$1.1M-per-site ROI anchors are Voxel/Intenseye-class incumbent results - FactoryLens has no pilot data.

  • Measured live in this demo

    Answer faithfulness is no longer just a target: the live demo ships a working evals rubric that measures it per answer.

The structural whitespace

One empty quadrant. That's the whole thesis.

OPEN WHITESPACEInteraction model → natural-language queryDomain depth → manufacturing-nativealerts / dashboardsVoxelAlerts + dashboards · safety-onlyIntenseyeConfigured detectors · safety-onlyProtex AIDetection + reports · safety-onlyTwelve LabsTrue NL query · horizontal (our supplier)DrishtiAcquired by Apple 2023 · demand strandedFactoryLensNL query · manufacturing-native

Everyone makes you configure alerts on a safety-only tool, or offers open query with no factory depth. FactoryLens is alone in the manufacturing-native, ask-anything corner - the quadrant Drishti vacated.

Competitive landscape

Everyone makes you configure. No one lets you ask.

Voxel

Safety vision
~$61M raised ($44M Series B, Jun 2025); 5 continents
Interaction: Real-time alerts + dashboards
Gap vs. us: No open NL query; safety-only; ops ROI secondary

Intenseye

Safety vision
~$64M; ~$300M val
Interaction: Hazard detection + analytics
Gap vs. us: Configured detectors, not free-form Q&A; enterprise-only

Protex AI

Safety vision
~$54M; Salesforce Ventures
Interaction: Detection + gen-AI reporting
Gap vs. us: Report generation, not a query interface; safety-only

CompScience

Safety + insurance
$27.6M Series B
Interaction: Monitoring + workers' comp
Gap vs. us: Insurance-led, narrow; not a query product

Twelve Labs

Video foundation model
Marengo 3.0 + Pegasus on Bedrock
Interaction: True NL video search/Q&A
Gap vs. us: Horizontal; no factory ontology, integrations, or compliance. Our likely supplier, not competitor.

Invisible AI / Retrocausal

Mfg ops analytics
Active
Interaction: Cycle-time / quality dashboards
Gap vs. us: Ops-only dashboards; no safety hook; no open query

Rockwell VisionAI (Elementary)

Industrial automation
Incumbent
Interaction: Quality inspection, PLC closed-loop
Gap vs. us: Inspection-specific; not a cross-cutting query layer

Drishti

Mfg video analytics
Acquired by Apple, Sep 2023
Interaction: (Was cycle-time + traceability)
Gap vs. us: Gone as a vendor. Demand is stranded.

Our defensible wedge

Four moats, compounding

Interaction moat

Natural-language query becomes the daily habit; dashboards become a feature, not the product.

Domain moat

A manufacturing event ontology (PPE, LOTO, zone breach, forklift proximity, spill, cycle complete, micro-stoppage, idle, ergonomic strain) horizontal VLMs do not have.

Trust moat

Edge-first + privacy-by-design + an evals layer that makes answers auditable. This clears a works council and an EU AI Act review.

Wedge economics

Safety opens the door on existing budget; operations proves hard-dollar ROI that drives the renewal and the expansion.

↑ Independent product concept by Kaushal Khodifad. FactoryLens is an independent product concept by Kaushal Khodifad; it is not a real company or a commercial product. It explores the natural-language intelligence over factory video space. Not a live commercial product. Data is illustrative.

Return to portfolioOpen the live demo