Visual Intelligence

Turn your blind spots into insights

Video footage becomes production data, so you can see which processes cost you the most, and why.

Camera view of a bottling line where Visual Intelligence detects the filler, the shrink wrapper, line speed and the people at each station

Why Visual Intelligence

Find the minutes your machine data can’t explain

Visual Intelligence uses cameras and AI to capture what happens around your machines. AI processes the footage right at the line and converts it into numbers that show up in Factbird next to your downtime and performance data.

Measure what your sensors miss

Manual cycle times and starved or blocked stations become numbers that show which steps cost you output.

Get to the root cause faster

Replay the thirty seconds around every detected event, so nobody has to stand at the station taking notes.

Know which loss to tackle first

See whether to automate a manual step, change the material supply, or rebalance the line, before spending budget on the wrong fix.

How it works

Better understand the reasons behind downtime, speed loss, and scrap

Visual Intelligence adds camera data to Factbird, so you can see how long manual cycles take, when a station runs out of material, and where product backs up on the line.

Use cases

What Visual Intelligence captures

Factbird Live view of a manual station with the uptime chart, the camera frame and the Station Utilization and Cycle Time cardsFactbird Live view of an infeed conveyor with stops on the uptime chart and the Number of Stops, Cycle Time and Downtime cardsCamera clip from before a filler stop, played in front of the Factbird Live viewFactbird view of the AI Visual Counter counting cookies on a cooling conveyor, with one cookie marked red by a quality check

Connected data

Bring visual data into the production view you already use

Factbird shows visual signals right next to your downtime, speed loss, performance, and quality check data.

Line 2▼LIVEREPORTSREGISTER STOPSBATCHESANALYTICSOEE
30M1H4H8H12H24H1W11/13/2026, 12:00 PM → 11/13/2026, 4:00 PMACTIONS▼
pcs/min050100150200
Thursday, 13 Nov, 2026 · 02:34 PMLine 254.12 pcs/minBatch 1186Expected speed: 100 pcs/minValidated speed: 130 pcs/minCamera: Station 4 starved for 3 min 40 s
12:0013:0014:0015:0016:00
Date
Expected speedValidated speedShow stops on chartShow annotations
ANNOTATIONCAMERA
Worth investigating: Station 4 was starved for 3 min 40 s right before the dip. The PLC reported the machine as running.
Which station held Line 2 back on the day shift?AI AssistantFactbird MCP
Ask your line a questionGet answers in the Factbird AI Assistant, or in Claude, ChatGPT, and Copilot via MCP

Layers that work together, all around your line

Cameras, edge processing, AI, and the machine data you already have overlap in one platform, so each layer adds context to the others.

1

Hardware

Factbird® VIEW and Factbird® EDGE

Cameras film the stations and areas you choose. Factbird® EDGE processes on site and blurs faces before any video leaves the device.

2

AI capability

Visual Intelligence

Turns footage into cycle times, station states, material flow, and product counts.

3

Platform

Production Insights

Shows it all together in your dashboards, by shift, batch, and product.

Production Insights
Camera and machine data in one view
Manual stationsMaterial flowStopsProduct countsProduct QualityShift and batchPeople

What you get with Visual Intelligence

Count irregular products

Count odd-shaped, overlapping, or hard-to-measure products with the AI Visual Counter.

Measure manual or machine cycle times

Measure how long a manual or machine cycle takes by station, person, or robot.

Understand station utilization

See how much of the shift each station is in use and how work moves between stations.

Get insights into machine states

Know when your machines are running, adding value, or waiting.

Detect starved or blocked processes

Detect when a station waits for material or is held up downstream.

Connect video footage to downtime

Review video around detected events to validate what happened around downtime and losses.

Perform product quality checks

Spot visible defects and deviations and do regular quality checks of your products.

View operator interactions

See whether a station was staffed when a cycle ran long or the line slowed down.

Trust and privacy

Designed for trust

Built to measure processes, not people. Here’s how Factbird protects the people on camera.

Processed on the edge

Detections run on the Factbird® EDGE device at the line.

Blur individuals on the device

People can be blurred, or blacked out entirely, before any footage leaves the device. You choose what your site requires.

Permission-based video access

Decide who can review footage and who only sees the resulting data.

Person at a filling line blurred by the Factbird EDGE device, and a card showing who can see video and who sees data

Getting started

No big rollout. You start with one station, see the signals in Factbird, and expand from there.

1

Tell us about the station

A manual process, a recurring speed loss, or a station you can't measure today. We look at it with you.

2

We plan the application together

Capture method, camera placement, and the privacy and works council requirements that apply at your site.

3

Signals in Factbird, then expand

The first application goes live in your Factbird view. Add the next station when you see the value.

Visual Intelligence FAQ

Can't find what you're looking for? Browse the guides in the Support Hub, or ask your Factbird contact.

What is Factbird Visual Intelligence?

Factbird Visual Intelligence is an AI capability in the Factbird platform that turns camera footage into production data. It captures what happens on and around your machines, such as manual work, material flow and product counts, and shows it next to your machine data.

What can Factbird Visual Intelligence measure on a production line?

Depending on the application, it produces signals for product counts, cycle time, station utilization, machine state, and starved or blocked conditions, plus video from the moments around an event.

How do you measure manual cycle time automatically?

A camera watches a defined area around the station. Factbird Visual Intelligence detects when a cycle starts and ends, processes the detections on the Factbird EDGE device, and sends the cycle time to Factbird, where you can review it by shift, batch, product, line or site. No stopwatch study needed.

How does it help with bottleneck analysis?

It shows where product stops moving and whether a station is starved (waiting for material) or blocked (held up downstream). Seeing how often that happens, and on which shifts, shows where the real constraint on the line is.

Can it count products without a sensor?

Yes. The camera can count irregular, overlapping, or hard-to-sense products, and you see the counts in Factbird next to your other production data.

Does it show why a line stopped?

It shows the camera clip from the moments before a stop, next to the stop in your machine data. It does not assign the root cause automatically, but it gives your team the context to investigate, and AI highlights events worth looking at first.

Do we have to replace our sensors or PLC connections?

No. Factbird Visual Intelligence sits next to the sensors and PLC connections you already have. You add a camera where context is missing and keep what already works.

How is Factbird Visual Intelligence different from Factbird® VIEW?

Factbird® VIEW gives you video for stops: click a stop and see what happened. Factbird Visual Intelligence uses AI to turn what the camera sees into quantitative signals, such as counts, cycle times and starved or blocked states, for dashboards, analytics and AI tools.

How is the privacy of people on camera protected?

Detections are processed on the edge device at the line. People can be blurred before footage leaves the edge device, and access to video is controlled through Factbird permissions.

Where can I use the data?

In the Factbird views you already use, such as downtime and speed loss, filtered by shift, batch, product or site. Supported signals are also available to Factbird analytics, the AI Assistant and Factbird MCP, so you can ask Claude, ChatGPT, or Copilot in plain language.

Get started

See whether Factbird Visual Intelligence fits your line

Tell us about a manual process, a recurring speed loss, or a station you can't measure today, and we'll help you work out whether a camera is the right capture method for it.