AI value stream mapping from video: LeanSuite turns phone footage of each process step into a current-state value stream map with measured cycle times, counted inventory, the lead-time ladder and process cycle efficiency, then branches it into future-state scenarios you can compare and export.
By clicking "Schedule Demo," you agree to the use of your data in accordance with LeanSuite's Privacy Notice, including for marketing purposes.
Watch
· 4:48
Most value stream maps get drawn in a conference room, from what people remember. The cycle times are guesses. The process has already changed by the time the map goes on the wall. Then nobody redraws it, and the next workshop gets built on top of it. In this episode Rosa and Gurpreet find out their map is wrong, five days before a workshop. They go to the…
Press play to load the player from YouTube, or watch it on YouTube.
What it answers
Phone video of each step and the inventory between them goes in. A current-state map, built from what was filmed, comes out.
Cycle time, changeover and uptime for each step, read from the filmed cycles instead of typed from memory.
Inventory between steps is counted and converted to days, then shown in the lead-time ladder.
Process cycle efficiency is calculated from the mapped times, with no separate spreadsheet.
Branch the current map, add kaizen bursts and compare the two side by side. Export to PNG, PDF or VSDX.
The problem
Maps drawn from memory go out of date fast, and their times are guessed.
Drawn VSMs reflect what people remember, not what happens.
By the time the map is done, the process has changed.
Cycle and lead times are guessed, not measured.
How it works
Film each process step, let the AI build the map, then re-shoot after each improvement to keep it current.
Film each area of the flow, such as cutting, welding, painting, assembly and shipping.
The AI builds the flow with process boxes, cycle times, inventory and lead time.
Upload fresh video after each improvement and the value stream map updates.
From video to result
The AI builds the first draft of the map. You decide what is right.
You
Record each process step and the inventory between steps on a phone.
You
Upload the video to LeanSuite, pick value stream mapping and tag the line.
The AI
Boxes, lead-time ladder and process cycle efficiency, each backed by timestamped video.
You
Correct any box, certify the study, then branch the map into a future state.
Features
Built for industrial engineers, CI leads and supervisors. No AI expertise required.
Feature 01
Each filmed step becomes a VSM box with cycle time, changeover time and uptime.
Automatic VSM box generation per filmed step
C/T, C/O, and uptime populated from video data
No manual box-drawing for filmed steps
Feature 02
The AI counts inventory between steps and converts it into days of WIP.
Automatic WIP-to-days conversion
Inventory measured directly between steps
Real observed data, not manual estimates
Feature 03
Branch the current-state map, add kaizen bursts for the future state, and compare the two side by side.
Branch current state into future-state scenarios
Kaizen bursts simulated in-line
Side-by-side comparison in one map
Feature 04
Several product families on one map, each with its own routing.
Multiple product families on one map
Independent routing per family
Shared vs. divergent steps visible at a glance
Feature 05
Process cycle efficiency is calculated from the mapped value-add and non-value-add time, with no separate spreadsheet.
Automatic PCE calculation from mapped data
Live value-add vs. non-value-add breakdown
No separate spreadsheet needed
Feature 06
Edit any box, then export to PNG, PDF or VSDX for Visio.
Direct in-map editing of boxes and data
Export to PNG, PDF, or VSDX
Visio-compatible output for downstream editing
From video
Walk the process once, film each step and the inventory between them, and the map is built from what you saw.
Record each process step and the inventory between steps on a phone.
Drop the video into LeanSuite, pick value stream mapping and tag the line.
The AI builds the map with process times, wait times, inventory and information flow.
Review the map, then take the AI-recommended future-state map as your starting point.
Why It Matters
Each box comes from filmed observation, a re-shoot keeps the data current, and future-state kaizen bursts sit in the same map.
Truth, Not Memory
Every box is backed by timestamped video.
Easy to Refresh
Re-shoot to refresh, no markers required.
Future-State Ready
Simulate the future state in the same tool.
Who it is for
The boxes and times are drafted for you, so your time goes on checking the map and on the future state.
A current-state map backed by video, and kaizen bursts in the same map for the improvements you choose.
Inventory between steps in days, and the lead-time ladder that shows where the waiting is.
Current and future state side by side, exported to PNG, PDF or VSDX.
Honest limits
Worked example
An illustrative line, not a customer: cutting, welding, painting and assembly, filmed on a phone, with the inventory between steps counted on the same walk. Customer demand is 600 pieces a day.
1. Step data from the video
In the example, cycle times of 45, 60, 30 and 90 seconds per piece, changeovers of 30, 45, 0 and 15 minutes, and uptime of 95, 90, 100 and 98 percent, each read from the filmed cycles rather than typed from memory.
2. Inventory in days
900, 1,800, 600 and 300 pieces waiting between steps. At 600 pieces a day that is 1.5, 3, 1 and 0.5 days, 6 days of stock in all.
3. The lead-time ladder
Processing time is 45 + 60 + 30 + 90 = 225 seconds, under 4 minutes per piece. Lead time is the 6 days of waiting plus those 4 minutes.
4. Process cycle efficiency
225 seconds of value-add work in roughly 518,400 seconds of lead time (6 days at 24 hours) works out to about 0.04 percent. The map computes it; no spreadsheet.
Result: under 4 minutes of work inside about 6 days of lead time, a process cycle efficiency near 0.04 percent. The waiting, not the work, sets the lead time, and the future-state branch starts with the 1,800 pieces queued before painting.
FAQs
Value stream mapping (VSM) is a lean method that draws every step, wait and inventory point from order to delivery, so teams see where time is lost and design a better future state.
Film each process step and the inventory between them, then upload. LeanSuite builds the boxes, lead-time ladder and PCE; you review the map and branch it into a future state.
No. You film the steps and upload the files, and the AI builds the map from that video. It does not send live alerts at the station.
You do. The AI drafts the boxes, each backed by timestamped video, and a person reviews, edits and certifies the study.
Footage is encrypted in transit (TLS 1.2+) and at rest (AES-256), processed in your own tenant and not used to train external models. LeanSuite has a SOC 2 Type 1 report.
Yes. Export to PNG, PDF or editable VSDX (Visio).
Lead time is computed from observed WIP between steps plus C/T at each process box.
Yes. The same engine works for back-office processes captured by screen recording.
Preferred but not required. Manual data entry is supported for steps that can't be filmed.
Film WIP locations or enter counts manually; both feed the lead-time ladder.
Yes. Branching, merging, rework loops and outside processes are supported.
Upload the video and the AI builds the current-state map for you to review and edit.
FREE RESOURCES
Not ready for software yet? Start with these free resources.
Book a demo and see the current-state map built from video of your own steps. Vision AI is priced separately.
