No login, no sales call, no sample data standing in for yours. Ten tools you can run yourself right now, plus three production tools that ship as part of an engagement — what each one takes in, what it gives back, and why it matters.
There are a lot of tools here, and you don't need all of them. They answer five questions in sequence, each one feeding the next, and you can stop after any step with something useful in hand.
Start with the Lean AI Readiness Scorecard. Rate eighteen signals and you get a score, a profile across six dimensions, and the three gaps most likely to stall an AI initiative. If the score is low, the best next move is usually fixing the process, not buying technology, and you will know that before you spend anything.
Pick one process, not the whole business. If you can describe how it runs today, paste that into the AI Opportunity Audit and it returns a value stream with the waste named. If you would rather build the picture yourself, use the Value Stream Mapper: enter the steps, cycle times and inventory, and it works out lead time and value-added ratio.
Take the steps you found and run them through the Judgment Lane Classifier, one at a time. Each step lands in one of three lanes: plain software, an AI agent, or a person in the loop. "Not AI" is an accepted answer. Once you have scored several workflows, the Lean × AI Opportunity Map puts them all on one chart, so you can see what to fix first, what to standardize, and where AI is worth trying.
Put your top one or two candidates into the Automation ROI Calculator. It uses published wage data for the US, Singapore and India, so the payback figure holds up when finance asks where it came from.
The Skill File Generator turns an existing SOP into a structured file an AI agent can be trained on. The Agent Audit Trail shows what accountable AI looks like: every decision logged, nothing released without sign-off. The Daily Operating Rhythm shows what each executive role should check before 9am, and is worth a look once you are running something and need to watch it.
If you only have time for two: do the Scorecard, then run your single most painful process step through the Lane Classifier.
When you want this on your own data: the client tools below take over. Process Miner reads your real ERP event log, Microbot Miner ranks automation candidates by return, and North Star Compiler sets the strategy above both. The five-day assessment is the usual way in.
Rate eighteen signals across six dimensions: Lean process maturity, data foundation, infrastructure, organizational readiness, use-case viability and governance. You get an overall score, a profile by dimension, and the three gaps most likely to stall an AI initiative.
Score every workflow in your operation — Order Processing, Procurement, Incoming QA, Warehouse Ops, whatever you run — on Lean maturity and AI opportunity, and see all of them on one quadrant map.
Score a real task on judgment, risk, volume and exceptions. It routes the step to one of three lanes — deterministic software (rules, no model needed), AI agent (real judgment, applied narrowly, inside guardrails), or human in the loop (decision stays with a person) — and says so plainly when the answer is "not AI."
Payback modelled on published wage data, not invented numbers — US, Singapore, India. Pick your region; the honest answer changes with it.
Describe how a process runs today. Get back a value stream with the waste named, every step routed to a lane, and the risks stated as plainly as the opportunities.
Upload or paste an SOP and get back a structured Markdown "skill file" — the same AI-ready instruction format we use to train agents on a client's real procedures, with the Lean waste named and the one human checkpoint called out.
Assemble the end-to-end value stream in the conventional VSM format — process boxes, data boxes, inventory triangles, an auto-computed lead-time ladder — the way you'd place it on a wall with sticky notes.
Watch an agent check five supplier certificates against a batch spec — every read, conversion and decision logged, nothing released without a person's sign-off.
Pick a seat — COO, CFO, CHRO, CIO, or CMO — and see the numbers that role should check before 9am, plotted against target on gauges and bars. A few numbers are editable, so you can drop in your own rough figure and watch the status flip.
Two agents from the lab: one finds people already engaging with your world on LinkedIn and scores them into an outreach pipeline; the other turns sales calls and case studies into on-brand LinkedIn drafts, routed through a human approval queue.
These decide whether AI-generated output is good enough to ship. Every result is drafted by one AI, then scored by a second, independent one against a fixed rubric — and sent back for revision until it clears the bar. Provisioned as part of an engagement.
Turns a leadership questionnaire into a boardroom-ready Hoshin Kanri X-Matrix — strategic objectives cascaded into department action plans, with a rules engine that flags any correlation without a real supporting link before the Auditor reviews it once.
Reads your actual ERP event log and reconstructs the production flow that really ran — not the one on the whiteboard — then overlays Takt time, cycle time, VA/NVA and bottlenecks onto it.
Takes the Process Miner output, or a process inventory on its own, and ranks automation candidates by return — frequency, handling time, error rate, systems touched.
Start with the five-day assessment, or talk to us directly about what you're trying to fix.
Book the assessment →