Industrial / AI pilots
How industrial AI, OT/IT, and plant-floor software teams convert paid pilots on a named line or site into production POs and MSAs.
Find: a lab demo won't convert on its own
Manufacturers, process plants, and discrete OEMs will hand you an innovation-center slot, a lab PoC on offline data, and a "let's explore AI" thread without much thought. None of that makes a pilot. A convertible pilot has a named plant, line, cell, or SKU family, a compelling event, written criteria ops and finance can underwrite, and a path to a PO or MSA if the proof lands.
Keep unpaid co-development, university or lab research, and science-project PoCs separate from customer pilots. Lab demos can build evidence later, and they seldom close deals this quarter. If one SOW mixes unpaid co-dev with a paid commercial path, name which clock you are on.
Hunt for compelling events: a planned outage or turnaround window, the CAPEX budget cycle, a quality escape or scrap spike, a throughput bottleneck, a labor shortage on a cell, a spike in energy cost per unit, or a corporate digital / Industry 4.0 mandate with a dated target.
| Signal | Treat as | Why |
|---|---|---|
| Lab demo / innovation-center PoC, no plant access | Nurture / lab | No PO clock |
| Named plant / line / SKU + deadline + buyer | Pilot candidate | Can end in PO / MSA |
| Free multi-quarter co-dev or university study | Design partner / research (separate) | Not this playbook |
Qualify: OT/IT, plant access, and the committee from day one
Industrial buyers expect OT access work to start up front. In the first week of serious interest, start OT security, the network segmentation and remote-access review, MES / SCADA / historian / ERP integration, safety and change control (MOC), IT vendor risk, and the SOC2 questionnaire. Run them before or in parallel with the plant-floor proof so you aren't starting after operators have said they like the dashboard.
Write down the committee. The plant champion (plant manager, VP Ops, or continuous-improvement lead) is not the economic buyer. The check-writer is often the COO, the plant GM, corporate ops, or a CAPEX committee. Plant enthusiasm without finance and OT/IT clearance dies after the demo.
Work backwards from your target ACV and the MSA / PO process. If production requires remote access through a DMZ, a historian tag list, or CAPEX committee sign-off, start those while the pilot runs instead of after the innovation-center tour.
| Role | Industrial quirk | Fail mode |
|---|---|---|
| Plant champion | Plant manager / VP Ops / CI lead | Single-thread; no CAPEX; no PO |
| Economic buyer | COO / plant GM / corporate ops / CAPEX committee | Loved demo, no budget line |
| OT / controls / automation | Network segmentation, remote access, PLC/SCADA touch | Block after lab proof |
| IT / InfoSec / OT security | Vendor remote access, SOC2, questionnaire | Access revoked mid-pilot |
| Safety / quality / MOC | Change control, LOTO, validation | Hold after criteria hit |
| Procurement | MSA, PO, insurance, site T&Cs | Quarter slip after OEE proof |
Setup: criteria ops and finance can underwrite
Write criteria a plant manager and a CFO or plant controller can both sign, and instrument them inside the pilot window on a bounded scope (one line, one cell, one SKU family): OEE, scrap rate, unplanned downtime hours, cycle time, first-pass yield, energy per unit, or changeover time.
"Operators liked the dashboard" and "AI accuracy looked good in the lab" won't pass as success criteria. Price the pilot when you can, and put the phase-two commercial terms (sites, lines, ACV intent, MSA / PO path) in writing so the pilot can't turn into a permanent unpaid PoC.
| Weak | Strong |
|---|---|
| Operators like the UI | OEE 62% → ≥68% on named Line 3 in pilot window |
| Should cut scrap | Scrap rate X% → ≤Y% on tagged SKU family |
| Less downtime | Unplanned downtime 14h/mo → ≤8h/mo on pilot cell |
| Faster cycles | Median cycle time 48s → ≤42s on named SKU |
| Feels smarter | OT access + MES/historian tags live before commercial ask |
Run: one line or site, midpoint scorecard, parallel close work
Land a quick win in weeks 1–2 that ops leadership and finance can see without living in the UI. A one-page memo with one underwritable metric beats another product tour.
Hold a formal midpoint against the written scorecard. Adjust only against the criteria, and don't expand to a second line or plant without saying so.
Keep OT access, network segmentation, MES / SCADA / historian / ERP integration, safety / MOC, IT vendor risk, and procurement moving each week. Industrial pilots miss the planned outage, CAPEX window, or fiscal close when the team finds OT remote-access or tag-list gaps in the final week.
| Window | Proof work | Close work |
|---|---|---|
| Weeks 1–2 | OT/IT access, first live line/cell run, baseline tags | Quick-win memo to EB + plant sponsor |
| Midpoint | Scorecard vs written criteria | Confirm / narrow lines or sites; invite EB + OT/IT |
| Final week | Evidence pack + integration / safety artifacts | Ask for PO / MSA schedule |
Close: cut lines, sites, SKUs, or scope instead of price
Restate the criteria, the evidence, and the commercial path from the agreement, then ask for the PO or MSA schedule. "Ops loved the demo" without a signature is a stall.
If budget or risk appetite wobbles, cut to one line, one cell, or one SKU family. Don't teach procurement that a successful plant pilot earns a list-price haircut.
Name the industrial stall mode before you improvise.
| Stall | Restart |
|---|---|
| OT access blocked | Daily owner; EB sees the blocker on the outage / CAPEX date |
| IT vs OT fight | Freeze to agreed remote-access path; escalate COO / plant GM |
| Plant champion only | Multi-thread COO/CAPEX + procurement this week with underwritable ROI |
| Safety / MOC hold | Freeze scope; clear change-control before expanding tags or actuators |
| Scope sprawl across plants | Freeze to one line / one site; re-score |
| Loved demo, no CAPEX | ROI + reduce lines/sites, not list price |
| Procurement takeover | Loop champion; tie delay to compelling event |
Expand: land one line or site, then expand without pilot pricing
Treat the pilot scope (one line, one cell, one SKU family) as a proof sample. Expand from the phase-two price and the MSA / PO terms in the agreement.
Add lines, sites, or SKU families after production is live. A published win (with permission) helps you sell the next plant. If you keep "pilot pricing" going across a manufacturing network, each expansion starts at a discount.
Worked vignette (fictional teaching example)
Use this as a pattern drill. It is not a claimed case study, and the numbers are illustrative.
Operator checklist
Use the Industrial AI pilot scorecard template for the blank worksheet, a filled example, the OT/IT/integration parallel-path checklist, and a copy-ready AI prompt.
FAQ
- Is a lab demo or unpaid university PoC the same as a pilot?
- No. Lab demos and unpaid co-dev can build evidence later, but they run on a different clock. A convertible customer pilot is paid or contractually scoped, with underwritable ops and finance criteria, named stakeholders including OT/IT and procurement, and an end date tied to a PO or MSA. Keep unpaid science projects out of it.
- When should OT/IT integration and plant-floor access start?
- In the first week of serious interest, often before or in parallel with the plant-floor proof. Industrial buyers won't clear production on a lab demo alone while OT remote access, segmentation, or the historian tag list is still open.
- Should we discount after a strong industrial pilot?
- Cut to one line, one cell, or one SKU family instead. A list-price discount after OEE or scrap proof teaches procurement to expect a haircut on each win.
FounderNexus
If the pilot worked and the PO is stuck, bring it to FounderNexus. Operators who have closed this motion will pressure-test the ask instead of cheering you on.
Join FounderNexusClaims and patterns on this page are paraphrased from practitioner session notes and the Startup Bible (Pilots that convert). See Sources.
Continue the journey
- IndustriesIndustry and company-size overlays on the same six-phase pilot spine: who signs, what counts as proof, and how deals stall.
- Industrial AI pilot scorecardBlank worksheet, OT/IT/integration parallel-path checklist, filled fictional example, and AI prompt for scoring an industrial / AI plant-floor pilot toward MSA or PO.
- Security and procurement in parallelStart security and procurement in week one.
- Success criteriaWrite three to five measurable success criteria before kickoff.
- Ask for the POEnd the pilot by asking for the purchase order.
- Land and expandExpand after a paid pilot without anchoring at pilot price.