22nd Annual Northeast Lean Conference · Recap

Standard Work 2.0™ | Northeast LEAN Conference Recap

The full slide deck from my session at the 22nd Annual Northeast Lean Conference: how to bring AI onto a regulated shop floor with the same discipline Lean brought to the physical one, so automation removes the Hidden Factory instead of building a second one.

Download Full Slide Deck (PDF)

PDF · 13 slides · 0.8 MB · free, no form

Session
Standard Work 2.0™: Engineering the Human-in-the-Loop AI Strategy
Conference
22nd Annual Northeast Lean Conference (GBMP)
Date
September 30, 2026
Location
Springfield, MA

Session recording

Watch the talk

The full session, start to finish, including the Friday Morning Action Plan and the room's questions.

Recording in post-production

The session video publishes here shortly. Until then, the full deck is a free download above, and every slide is summarized below.

Inside the deck

The talk, slide by slide

Each slide carries one idea. Together they map Lean's oldest disciplines onto the newest source of waste on the floor: unstructured AI.

  1. Slide 1: Engineering the human-in-the-loop AI strategy

    01 · Operating system for regulated production

    Engineering the human-in-the-loop AI strategy

    Translating Lean principles to eliminate the Hidden Factory of digital waste.

  2. Slide 2: A subscription is not a strategy

    02 · Methodology over software

    A subscription is not a strategy

    Unstructured AI adoption accelerates waste. Governance gets hardwired into the workflow before any LLM tooling arrives.

  3. Slide 3: The Hidden Factory in regulated production

    03 · Administrative tax

    The Hidden Factory in regulated production

    15 to 20% of factory capacity erodes into duplicate entry, offline spreadsheets, and manual re-keying.

  4. Slide 4: Six signals of digital waste, mapped to Lean

    04 · Diagnostic framework

    Six signals of digital waste, mapped to Lean

    Friction to motion, handoff loss to transportation, workarounds to overprocessing, redundancy to rework, overproduction to inventory, misalignment to unused talent.

  5. Slide 5: The Paperclip Maximizer in factory operations

    05 · Uncontrolled AI risk

    The Paperclip Maximizer in factory operations

    Un-gated AI floods the office with unreviewed documents. Pull-based staging gates hold digital work-in-progress at zero.

  6. Slide 6: Selective amplification vs. digital noise

    06 · Operational amplification

    Selective amplification vs. digital noise

    The Hearing Aid Model: AI amplifies the signal, and the operator's tacit knowledge interprets it.

  7. Slide 7: FADE vs. RISE: the 3C Protocol and Human API

    07 · System architecture

    FADE vs. RISE: the 3C Protocol and Human API

    Machines take the routine 80%. People keep the 20% that appreciates: reasoning, imagination, connection, and explaining meaning.

  8. Slide 8: Digital twins and system state

    08 · Data governance

    Digital twins and system state

    One-click process visibility, R&D separated from production, and AI configuration under document control.

  9. Slide 9: Industrial safety PL, applied to AI prompts

    09 · Safety standards

    Industrial safety PL, applied to AI prompts

    Prompt drift gets the same discipline as machine guarding: boundary constraints, a halt circuit, dual-run verification, weekly drift audits.

  10. Slide 10: Active monitoring and CUI/ITAR data enclaves

    10 · Boundary containment

    Active monitoring and CUI/ITAR data enclaves

    A primary model, an AI Whistleblower monitor, a Blue Hold circuit breaker, and an agentic firewall at the ITAR and CUI border.

  11. Slide 11: The Digital Calibrator and Friday Morning Gemba

    11 · Actionable roadmap

    The Digital Calibrator and Friday Morning Gemba

    Four hours a week of one manufacturing engineer keeps prompt templates calibrated and model drift audited, like CNC tooling.

  12. Slide 12: Foundational research

    12 · References

    Foundational research

    MIT Sloan, Autor, Kremer, Shannon, Bostrom, Feigenbaum, and the Standard Work 2.0™ whitepaper.

Download Full Slide Deck (PDF)

PDF · 13 slides · 0.8 MB · free, no form

The Friday Morning Action Plan

Three moves before any AI touches the floor

The talk closed on one page of homework. It is where every engagement starts.

  1. 01

    Map the Hidden Factory

    Walk the Gemba with one question: where does the same data get typed twice?

  2. 02

    Define the 80/20 boundary

    Decide which work the machine runs and where a human gate stays mandatory.

  3. 03

    Build the digital safety circuit

    Stand up 3C staging templates (Capture, Communicate, Check) before deploying AI.

Put the plan to work on your floor

A shop-floor diagnostic starts at step one. We find where your Hidden Factory is eating capacity, where the 80/20 line belongs, and which gates go in before AI goes live.

Prefer to self-assess first? Run the six-question operational diagnostic

Schedule a Shop-Floor Diagnostic

Pick a time that works. Thirty minutes, and you leave with a first read on your floor.

Strategic conversation, 30 minutes

October 2026

Only dates with openings can be selected.