The AI Box
Plugs into any industrial hardware you've already got.
Turns your operational data into whatever structure your business actually needs. Programmable by you, modular and transferrable.
Live in weeks · No new hardware · Read-only by design
Clarity · Facility knowledge base
Status: Live
Hardware required0
System accessRead-only
Time to first insight2 weeks
Works withYour existing stack
✓ERP connected — read-only
✓Historian connected — read-only
✓Shift notes ingested — 14 new entries
→Knowledge base compounding…
The core loop
Point at a handoff. Get a data structure back.
Before
Business Step A
Business Step B
The handoff between them isn’t tracked anywhere.
You click
Business Step A
Business Step B
Enrich this data!
High impact
CauseSequence timing drift
Duration3 sec microstoppage
Frequency120×/day
Compounding cost, annualized
36.5 hrs · $365K
3 sec × 120×/day × 365 days
Built for
50–500 people
Industries
Discrete, formulation & metals
Deployment
Live in NYC
Founded by
Manufacturers
The problem
Your plant already knows the answer. Nobody wrote it down.
20–40 years, walking out the door
The retirement cliff is here.
Your most experienced machinists and operators are retiring after decades on the floor. Training a replacement takes 6–9 months of shadowing — and everything that never got written down leaves with them. The valve sequence. The workaround. The reason Line 3 runs differently from Line 4.
$10K–$50K per hour
Every investigation starts from zero. Again.
When a line trips or a batch runs off-spec, the ERP knows what was ordered, the instrumentation knows what the line was doing, and the shipping system knows what went out. None of them talk to each other — so root cause takes a room full of tired people three days, and the same problem comes back next changeover.
$500K+ integrator quotes
Digitization has been priced out of reach.
Getting data out of a legacy floor has meant a six-figure systems-integration project or point solutions that need new hardware on every asset. So the data stays locked in the machines, and the plant keeps running on memory instead.
It’s not a data problem. Your plant is drowning in data. It’s a connection problem — and connection is cheap now.
The product
Three things you get on day one. Everything else is an add-on.
Clarity connects to what you already run — ERP, historian, MES, SCADA, even a 30-year-old machine with nothing but a PDF manual — and turns it into three standard outputs every deployment ships with. No systems-integration project. No new hardware.
Standard output 1
Downtime & Reliability
Every stoppage traced to a cause, not just logged as a number.
- MTBF and MTTR by machine and line
- Pareto of your top 5 downtime root causes
- The full causal chain behind any stoppage — the order, the machine signal, and the shift note that explains it, connected automatically instead of reconstructed by hand
Standard output 2
Throughput & Flow
A live process map of the line, with the slow handoff flagged for you — see it below.
- Real-time units/hr at every handoff between stations
- Cycle-time variance by cell and station, measured against takt
- Pinch points flagged automatically, before they become a stoppage
Standard output 3
People & Knowledge
Who knows what — and what happens the day they’re out.
- Time-to-competency by role
- Single-point-of-knowledge risk: the processes only one person can run
- Shift and resource performance, with coverage gaps flagged before they bite
These three ship with every deployment. If your floor needs something specific — SPC drift alerts, changeover coaching, a custom compliance log — it’s scoped as an add-on during onboarding, not a separate sales process.
Throughput & Flow — live
This pace already exists in your historian and machine network. Most OEM control panels never surface it, and building a dashboard to see it usually means a systems-integration project that takes months. Clarity reads it straight from data you already have.
Captures every read point
Clarity
All 3 read points · one view
Recommended: fix the Business Step B → External Dependency bottleneck first
Scroll to see the full line →
Stage 1
Business Step A
Wherever your process starts
Read point
2,050/hr
Never tracked before
Stage 2
Business Step B
Where the work happens
Read point
1,840/hr
Never tracked before
−10% · Bottleneck
Stage 3
External Dependency
Outside your direct control
Read point
1,790/hr
Never tracked before
Stage 4
Business Step C
Where it lands
OEM path
Months · $15K–$20K per machine · still no line-rate data
Clarity path
Live in weeks · zero new hardware · every handoff visible
Skip this, and the Pareto of what’s actually hurting your OEE lives in someone’s head. The pinch point costing you 10% of throughput goes unnoticed until it’s a stoppage. And the one person who knows the workaround retires before anyone else learns it.
How it works
Three steps. Zero hardware.
Step 1 — Connect
Connect
Point Clarity at what you already have: ERP, historian, MES, machine network, documents, spreadsheets. Secure, read-only connections. Cloud by default, on-prem if your environment requires it.
Works with SAP · Oracle · Ignition · OSIsoft PI · Custom MES · CSV
Step 2 — Capture
Capture
Machine telemetry and operator knowledge land in the same place, timestamped and cross-referenced. The knowledge base starts compounding from day one.
First insights in under 2 weeks
Step 3 — Ask & act
Ask & act
Your team queries it from the floor the moment they need an answer. Alerts surface emerging patterns before they escalate.
Built for gloves-on, loud-floor useBeyond storage
It doesn’t just store what you know. It finds what you don’t.
Knowledge gaps · This week
Flagged from shift handover notes · 4 mentions this month
Staff report the Line 3 changeover process is undocumented — it exists only in one operator’s memory.
Start documenting this process
Est. cycle time improvement
15%
Flagged from rework logs
Rework is 3× higher on the Mixing → Filling handoff than any other transition — no documented procedure exists for it.
Build a procedure for this handoff
Highest-rework transition on the floor
3×
Clarity watches handover notes, rework logs, and machine data for the places your documentation runs out — then tells you which gap is costing the most.
Results
One real deployment. No hypotheticals.
Custom-formulation manufacturer, New York City — first Clarity deployment
Senior operators carried years of undocumented process knowledge — the valve sequence, the workaround, the reason one line ran differently from another. New hires learned it the slow way: by shadowing someone for months.
We connected their existing systems read-only, mapped the line end to end, and established real baseline cycle times for the first time. Everything the team already knew went into the knowledge base instead of staying in one person’s head.
Senior-staff training time
–50%
Every new hire starts from what the last one learned
Cycle time, key subprocess
–8%
≈ $35K saved per quarter on one line
Time to first baseline
2 weeks
New hardware installed
0
Security & data
Read-only. By architecture, not by promise.
Clarity never writes to your systems.
Every connection is read-only. It cannot modify your ERP, your historian, your PLCs, or anything else. Zero risk of data corruption — your systems run exactly as they did the day before.
One connection point, not a fleet of devices.
No new sensors, gateways, or device classes for IT to manage and patch. One integration to vet, with documented scope of exactly what Clarity touches.
Your data stays yours.
The knowledge base your facility builds is your asset — it is never shared, pooled, or used to serve other customers.
Deployed to fit your environment.
Cloud-hosted by default. On-prem or air-gapped deployment available for organizations that require full data sovereignty.
Get a free operations audit
A read-only look at your floor — we show you what Clarity surfaces, no commitment.
The team
Built by people who’ve run machines, not just modeled them.
Wil Barnes (left), Oscar Elizalde (right) — Brooklyn, NY
Wil Barnes & Oscar Elizalde — Co-founders
- We’re manufacturers ourselves — subtractive and additive, aluminum and steel, fabrication.
- 10+ years each across software development and manufacturing engineering. We speak both languages, and we’ve sat through the $500K integrator quote on the other side of the table.
- Brooklyn-based. For NY-metro plants, we’re on your floor same-day when it matters.
Sid Powell — Advisor, CEO, Maple Finance
Nick Hebert — Enterprise Sales Director
Pricing
Starts at $2,000/month. One prevented stoppage covers the year.
Where our first deployment started
Cloud
$2,000/mo
A single connected operation
Unlimited source connections, direct query access for your whole team.
Month-to-month · No long-term commitment
Enterprise / On-Prem
Custom
Full data sovereignty — on-prem or air-gapped
Deployed inside your environment, dedicated support.
Your infrastructure, your control
Discovery (1–2 days on-site)
→
Install & stabilize (wks 1–2)
→
Baseline report (wk 2)
→
Variance analysis (wk 4)
→
Pilot review & business case (wk 8)
Every phase produces a document you can put in front of leadership without editing it. If we prevent one stoppage, the pilot has paid for itself many times over.
Questions plant managers actually ask
Yes — that’s the point. If the machine has documentation, Clarity generates an integration driver from it. A 1987 lathe with a PDF manual is just another source. The legacy floors other vendors walk away from are where Clarity works best.
Read-only connections to the systems you choose — ERP, historian, MES, machine network, even shared drives and spreadsheets. You control scope. Nothing is written back, ever.
No. There’s no hardware install on your assets and no changes to your systems. Connection happens alongside normal operations; the pilot clock doesn’t even start until data is flowing cleanly.
No new hardware to connect what you already run — and where a pinch point has never been instrumented, we deploy it once, as a standard kit, not a project.
You do, outright. Every fault, fix, and pattern captured is your facility’s asset. It’s also why Clarity gets more valuable every year it runs — and why it’s not worth ripping out in year two.
Especially for you. Clarity was built for 50–500-person manufacturers running on paper logs, whiteboards, and one very experienced person’s memory — the operations that were never going to justify a $500K integration project.
First insights in about two weeks from connection. A full pilot review — baselines, variance analysis, quantified improvement opportunities — lands at eight weeks.
The knowledge to keep running is already in your building. It just needs somewhere to live.
Tell us about your operation — we’ll show you what Clarity finds on your floor