The voice-to-text app that built its own speech model
Most dictation apps are a wrapper around someone else's speech model. Wispr Flow trained its own.
It's called Canto, and it was built for the way people actually talk to AI: thinking out loud, changing direction mid-sentence, half-whispering in an open office.
Wispr tested it on 10 hours of real dictations from more than 2,300 speakers. Canto had the lowest word error rate of every model tested, including models from Google, OpenAI, AssemblyAI and Deepgram. That means your prompts to ChatGPT, Claude and Cursor come through the way you said them.
Talk naturally in any app and Wispr Flow types clean, formatted text, 4x faster than typing.
Every plant has a spreadsheet that runs more of the business than anyone admits. A reader asked me when it finally loses.
Last week asked how long a tool lives after the person who loved it moves on. This week turns it around: why the oldest tool in the building outlives every champion it ever had. Here's the question, and my answer.
The Question
In October I sent out a five-question survey. The last box was optional, and I'd seeded it with a hint: the spreadsheet that quietly runs the whole plant counts.
One reader took me up on it. Here's what they wrote, word for word:
"What is the tipping point (pain point) that causes manufacturing leaders to finally decide to adopt 'change.' Utilizing the spreadsheet that runs the whole plant quietly (although, some could argue the 5 fans that kick on in my laptop when I open it is in fact, not quiet) remains the 'easier' option, until when?"
First, a correction they've earned. I called it quiet. Five fans says otherwise.
Second, it's the right question, and it's two questions. Why does the spreadsheet stay the easier option? And what finally changes that?
Why It Keeps Winning
The spreadsheet doesn't win because it's good. It wins because its costs are spread out and the new system's costs show up all at once.
The spreadsheet's cost is ten minutes here, a re-keyed column there, a laptop that sounds like it's preparing for takeoff. It's spread across a dozen people in pieces too small to complain about. Nobody ever adds it up, so on paper it costs nothing.
The new system's cost arrives as one number, on one budget line, with one name next to it. Of course the spreadsheet looks cheaper. It's the only option that never sends an invoice.
There's a second reason, and it's the one people don't say out loud. The spreadsheet has a champion who never leaves: the person who built it. Back in June I called that stake, the kind of resistance that comes from what the change costs someone personally. Retiring their spreadsheet can feel like being told their best work no longer counts or calling their baby “ugly.” That doesn't make them wrong. It does mean a training deck won't move them.
A Formula Older Than Excel
Organization development worked on this exact question sixty years ago.
In the early 1960s, David Gleicher, a consultant at Arthur D. Little, wrote down a formula for when change happens. Richard Beckhard and Reuben Harris published it in 1977. Kathleen Dannemiller later simplified it into the version most people use today:
D × V × F > R
D, dissatisfaction with how things are now.
V, a vision of what's possible.
F, first concrete steps toward it.
R, resistance to the change.
Change happens when the first three, multiplied together, outweigh the resistance. The important word is multiplied. If any one of the three is zero, the whole thing is zero, and the spreadsheet stays.
Run the reader's question through it and the answer gets uncomfortable.
Dissatisfaction is usually low on paper. The pain is real, but it's spread thin. Five fans is an annoyance, not a budget request.
Vision is usually a vendor demo. A demo shows someone else's data on a good day. It doesn't show the person who owns the spreadsheet their own Monday morning, done the new way.
First steps are usually a twelve-month implementation. That isn't a first step. That's a circular staircase with no handrail.
So the tipping point isn't when the pain gets big enough. The pain has been big enough for years. It's when the pain gets concentrated enough for one person to see the whole total, and a first step gets small enough to believe.
What Usually Tips It
When the spreadsheet finally loses, it's rarely a strategy decision. It's an event that concentrates the cost overnight. The four I see most often:
The person who maintains it announces they're leaving.
Suddenly the cost of the spreadsheet has a date on it.A number doesn't tie in front of the wrong audience.
A customer, an auditor, or a leadership review. Spread-out pain becomes one embarrassing moment.The file breaks at the worst time.
Quarter-end, a corrupted copy, or three versions named FINAL.The business outgrows it.
A second site, a new product line, or a volume the spreadsheet can't stretch to.
Notice what all four have in common. None of them is a decision. Each is a crisis making the decision for you, on its schedule instead of yours.
Three Moves
Here is what I would do instead of waiting for the crisis.
Add up the spread cost.
Hours a week, times the number of people who touch it, plus every error you can trace back to it. You're not inventing pain. You're concentrating the pain that's already there, so someone can see the total before a crisis shows it to them.Make the vision their Monday, not the demo.
Take one report the spreadsheet owner builds by hand and show it coming out of the new system, with their data. The vision that moves people is their own job, a little lighter, not someone else's dashboard.Shrink the first step until it's boring.
One tab. One report. One month. If the first step needs a steering committee, it's too big. (Next week I'll get specific about one of the smallest first steps I know, and why it scares people more than the big ones.)
One caution: the formula is a thinking tool from organization development, not a law of physics. Use it as a checklist, not a calculator.
Try This Monday
Twenty minutes, on one spreadsheet that runs something important.
Answer two questions:
What does it cost a week?
Ask the people who touch it how many minutes they spend feeding it, fixing it, or waiting on it. Add it up in hours.Which of D, V, and F is closest to zero?
Score each from 0 to 3, honestly. Is anyone dissatisfied enough to say so? Has anyone seen their own work done the new way? Is there a first step smaller than a project?
When you finish, you'll have two numbers:
Weekly cost: hours a week the spreadsheet quietly takes.
Weakest factor: the letter with the lowest score.
Here's an example:
Six people touch it for about two hours a week each. Dissatisfaction scores a 2, vision a 1, first steps a 0. Your result is "12 hours a week, F is zero."
The first number is your dissatisfaction, finally added up. The second tells you where to spend your effort. If F is zero, another demo won't help. A smaller first step will.
Until When?
So, to the reader who asked: the spreadsheet stays the easier option until someone adds up what it costs, or until it breaks in front of the wrong person. Those are the only two endings I've seen.
You get to pick which one.
"The spreadsheet doesn't lose when it gets worse. It loses when someone finally adds up how bad it already is."
None of this makes the spreadsheet a villain. It got the plant this far, and the person who built it was solving a real problem nobody else would. Treat it like a retiring employee, not a crime scene. Thank it, write down what it knew, and plan the handoff.
ACTION ITEM: Pick the spreadsheet that runs the most in your shop. Run the Monday exercise and send me your two numbers: hours a week, and your weakest letter.
Reply, I read every one.
Go hug your ERP. Then ask what it would take to retire the spreadsheet sitting next to it.
Plant Floor to Cloud: The Newsletter goes out every Tuesday. The podcast runs monthly. Same charter, more room to think out loud. Watch or listen on the website, or wherever you get podcasts: Apple Podcasts, Spotify, YouTube, iHeart, and more.
Sources cited
Dannemiller, K. D. & Jacobs, R. W. (1992). "Changing the Way Organizations Change: A Revolution of Common Sense." Journal of Applied Behavioral Science 28(4): 480–498. https://doi.org/10.1177/0021886392284003
Cady, S. H., Jacobs, J., Koller, R. & Spalding, J. (2014). "The Change Formula: Myth, Legend, or Lore?" OD Practitioner 46(3): 32–39. (History of the Gleicher, Beckhard and Harris, and Dannemiller versions.)
Beckhard, R. & Harris, R. T. (1977). Organizational Transitions: Managing Complex Change. Addison-Wesley.
Overview used to cross-check the above: Formula for change, Wikipedia
Plant Floor to Cloud, June 23: Four Kinds of Resistance … And The One Leaders Keep Excusing



