JOAN
Menu

Manufacturing

Keep production running, and keep your people’s know-how on record.

Modern systems and AI for manufacturers, with your people making the calls.

Start a conversation See the numbers

Manufacturing

Operators get the latest approved work instruction.

They ask in plain words and see the revision and who approved it. Printed copies that are out of date get flagged.

Illustration: an operator asks for a bolt torque and gets the value from the approved work instruction, with its revision and approver, and a warning that the station’s printed copy is out of date.

Illustration. Names and details are invented.

What we hear

Where the work gets stuck

Four problems we hear from owners, plant managers and finance leads of mid-sized manufacturers.

  • Unplanned stops.

    A machine fails, the line waits and orders slip.

  • Experienced staff retire, and their know-how goes with them.

    Experienced people retire, and new ones are hard to find and slow to train.

  • Plans that do not agree.

    Sales, the schedule and materials live in different systems and spreadsheets, and quotes take too long.

  • Material prices change faster than quotes.

    Material prices change, and margin disappears between the quote and the invoice.

The insight

What an hour of downtime costs

An hour of stopped production costs from tens of thousands of dollars to millions. The meter starts the moment the line stops.

  • A consumer-goods plant

    $36,000

    for one hour of unplanned downtime, about $10 a second

  • A smaller manufacturer, at the top end

    up to $150,000

    for one hour of unplanned downtime, about $42 a second

  • A large car plant

    $2.3 million

    for one hour of unplanned downtime, more than $600 a second

A large plant loses about 27 hours a month to unplanned downtime, down from 39 in 2019.

Source: Siemens, The True Cost of Downtime, 2024. Siemens sells maintenance technology. Its survey is mostly of large manufacturers, with data from April 2019 to March 2023; $150,000 is its top-end estimate for smaller firms. Per-second figures are our arithmetic (per hour ÷ 3,600).

Read this chart as a list

An hour of unplanned downtime costs about $36,000 at a consumer-goods plant, up to $150,000 at a smaller manufacturer and $2.3 million at a large car plant. A large plant loses about 27 hours a month, down from 39 in 2019 (survey data from April 2019 to March 2023). Source: Siemens, 2024.

The numbers

What is at stake

Every figure is from the named public source and year. Select a source to read it.

Stage by stage

Where plant work slips

Choose a stage to see what slips today, where AI can help and what your people decide.

Stage 1 Quote the job

What slips today
Drawings, past jobs and today’s material prices sit in different places, so quotes are slow.
Where AI can help
Pull the drawings, past jobs, current prices and capacity into a draft quote.
Your people decide
The price, and whether to take the job.

83%

of US manufacturers named rising raw material costs as a challenge in 2026. National Association of Manufacturers, 2026

Stage 2 Plan production

What slips today
Orders, stock and capacity live in different systems, so the schedule is argued over.
Where AI can help
Bring orders, stock and capacity into one weekly view.
Your people decide
What runs, and when.

What it looks like

The Monday production meeting starts from one set of facts, not three.

Stage 3 Run the line

What slips today
A machine fails, the line waits and orders slip.
Where AI can help
Watch machine readings and maintenance logs for the signs that come before a failure.
Your people decide
When to stop a line for repair.

27 hours

a month lost to unplanned downtime at a large plant. Siemens, The True Cost of Downtime, 2024

Stage 4 Check quality

What slips today
Quality notes sit in spreadsheets, so repeat defects are spotted late.
Where AI can help
Spot defects that keep coming back and draft the quality report.
Your people decide
Whether a batch ships, and the safety call.

What it looks like

A defect that keeps coming back is flagged at the next meeting, not found at month end.

Stage 5 Keep the know-how

What slips today
Experienced people retire, and new ones are slow to train.
Where AI can help
Gather work instructions and fixes into one guide anyone can search.
Your people decide
What goes into the guide, and who signs it off.

1.9 million

of up to 3.8 million US manufacturing jobs to fill by 2033 could go unfilled if workforce challenges are not addressed. Deloitte and The Manufacturing Institute, 2024

Where we would start

Three places to start

Each one uses information you already have. Your people make the decisions.

  • An early warning on equipment.

    Machine readings and maintenance logs are watched for the signs that come before a failure, and your maintenance lead gets a short list to check. The aim is more repairs in planned windows, not mid-shift.

  • Know-how at the workstation.

    Work instructions and the fixes your best people know are gathered into one guide anyone can search. A new operator asks a question and gets your firm’s answer, with where it came from.

  • Faster, steadier quotes.

    When a request for a quote arrives, the drawings, past jobs, current material prices and capacity are pulled together into a draft. Your estimator checks it, so it can go out sooner and with fewer surprises.

Quality log, Planning system, Maintenance sheet, Quotes joined into one set of factsSeparate tools (Quality log, Planning system, Maintenance sheet, Quotes) are drawn as dashed tiles, each joined by a line to one solid block: One set of facts. Above it sits a plate: Your people decide.Quality logPlanning systemMaintenance sheetQuotes One set of facts Your people decide
  • Your tools today: Quality log, Planning system, Maintenance sheet and Quotes
  • One set of facts, owned by your firm
  • Your people decide
Quality log, planning system, maintenance sheet and quotes, joined into one set of facts. Your people decide.

The difference

What changes with Joan

  • Designed for fewer surprise stops, with more repairs on your schedule.
  • What your best people know is written down and kept in the building.
  • Quotes built on today’s costs, so they can go out sooner.
  • Your data, methods and records stay with your firm, so you can change software or AI later without starting over.
How we build

Client work

Already at work in Manufacturing.

  • Floura & Co.Plain-English questions about stock, assembly and production runs, answered from Floura’s live inventory and production system, read-only and logged.Read the Floura & Co. story
See all our work

Security and compliance

What your compliance officer will ask

Three plain answers. Joan builds the controls; your firm decides how they meet its own obligations.

  • Are our supplier prices and terms safe?

    We keep them in your own systems wherever we can, seen only by the people who need them. Where it matters, we connect read-only, so the AI can read records but cannot change them.

  • Does our know-how stay ours?

    We keep your process knowledge, drawings and methods in your own accounts wherever we can. Joan does not train or fine-tune AI models on client data.

  • Can we tell who looked up what?

    Yes. Each question is logged for audit, with who asked and when. Where the AI only reads, it cannot change a record.

Joan does not train or fine-tune AI models on client data.

Security and data sovereignty

Keep reading

Useful for every firm

All insights

Next step

Tell us where the work gets stuck.

Tell us what slows your plant or your office down. Please leave out customer and pricing details.

Start a conversation