Your Operating Standard
A written record of how your work gets done, with every line sourced and dated, that your people and your AI tools both read before they act.
Applied AI implementation · West Michigan
The output gets better. Not just faster.
We build AI into the work your people already do, one real process at a time, starting on site with the people who do it and keeping everything in accounts your company owns. The work comes out more thorough and more consistent, each new process builds on what the last one wrote down, and your best people get their attention back for the judgment calls and conversations that decide outcomes.
Conversations are always free.
Our deepest work so far has been inside West Michigan manufacturers, where the knowledge work is quotes, prints, quality paperwork and what senior people know that was never written down. The same method applies to non-profits, reps and distributors, and other organizations that run on knowledge work.
- Sources: where each fact comes from, and when it was last checked.
- Rules: how your company does the work.
- Past jobs: what you quoted or built before, and why.
- Deepest work
- Manufacturing
- So far, as of fall 2026
- Where
- West Michigan
- On site, remote where it fits
- Ownership
- Your accounts, your passwords
- Swap the AI; the Standard stays.
- First conversation
- Always free
What we do
What we build inside an operation
An engagement starts from one real process and uses only the pieces that process needs. Every engagement is scoped independently.
Everything we buildDocument work
The system reads prints, PDFs, spreadsheets, ERP exports and emails, checks what’s in them and drafts the working document. Every value traces to its source, and anything unclear goes to a person.
Feasibility first
When we aren’t sure something will work, we test it on your real documents first. The test ends in a written go, change course or stop, and if the answer is stop, you’ve paid for the test and nothing more.
Tools you own
Internal tools, reports and dashboards that read from the records you already keep. Tools your own people already built move into company accounts with instructions to rebuild them, so your business can maintain and change them.
Connecting what you run
The system works alongside your ERP, email and files. It reads through access you’ve approved and writes back only where that’s safe and someone has signed off, and the ERP stays the ERP.
Team training
We train your people in the room, then help each person set up at their own desk. The goal is that someone on your team can change the system without calling us.
Manufacturing
Standard work for your knowledge work.
Standard work gave the floor one current method that everyone follows and that gets revised on purpose. We bring the same discipline to the paper around the machine time, the documents that cross between the office and the floor.
- 01
RFQ to quote
The system pulls the RFQ, its prints and your past jobs into a draft quote. Your estimator makes the judgment calls and approves it before it goes out.
- 02
Print to traveler
Before a job is released, the system checks the print revision, the BOM and the traveler against each other, and anything that doesn’t match goes to a person.
- 03
Work instructions
Setup knowledge that lives with a few senior people becomes draft instructions the next person can follow, and those senior people check that it’s right.
- 04
Reports from the floor
The ERP holds the data but often not the view people need. The system builds that view from ERP exports and leaves the ERP in charge.
Who we serve
Beyond the shop floor
Most of what we do applies wherever the work runs on documents, records and the people who know how it’s done.
All focus areasNon-profits
Grant reports, donor letters and board packets can be drafted from your own records and your funders’ rules, with a person reading and sending every one.
Reps and distributors
The system matches incoming leads against your line card, routes each one to the line that fits and drafts a reply for the rep to review and send.
Other organizations
The method applies to service businesses, professional firms, schools and other organizations whose work runs on documents and records. Tell us what yours looks like, and we’ll tell you whether we can help.
How we work
How an engagement runs
It starts with a conversation and ends when your people can run the system without us. The full routing on How we work adds the steps skipped here, with who’s in the room and what gets decided at each one.
The full engagement, step by stepOP 010
A conversation
You getAn honest read, free
Tell us about one process that keeps stalling, and we’ll tell you whether AI will help. Sometimes the answer is not yet, and we’ll say what would change it.
OP 020
Working sessions with your people
You getAn agreed first process
When the first conversation shows a real path, this is usually the next step, and it’s typically free too. We sit with the people who do the work, at their desks or at a whiteboard, and agree on which process goes first and what a better result looks like.
OP 030
A scoped first project
You getA written scope you approve
We write the project down in plain language, including how we’ll test that it’s done, and you approve it before work starts. When the outcome is uncertain, the first scope is a small paid feasibility test that ends in go, change course or stop.
OP 050
Foundation
You getA Standard you own
We set up your Operating Standard in your accounts, under your logins, with passwords in your password manager, and two of your own people can get into all of it without us.
OP 060
Build and train
You getWorking systems your team knows
We build one process at a time, on real work, with the people who do it, and they decide where a person checks the result. Your team is trained as the system comes together, and a short written update, reviewed by Holden, keeps you current.
OP 080
Handoff
You getThe code, instructions, settings and accounts
We’re done when your people can run it, fix it and change it without us, and we test that with them. After that, Cadence, our ongoing support, is optional, and you can cancel it any month.
The Operating Standard
One written record your people and your AI both work from
Every system we build reads your Operating Standard before it does anything, so the output fits how your company works. Four rules keep it from going stale.
- Every line names its source.
- Every line carries the date someone last checked it.
- Updating it is part of doing the work.
- When the record and reality disagree, reality wins and the record is fixed the same day.
It lives in plain files, in your accounts, and it doesn’t belong to any AI vendor. Swap the AI; the Standard stays.
What we believe
What you can hold us to
- 01
You own what we build
Your Standard, your code, your accounts and your passwords all sit under your logins, so if we disappeared tomorrow, nothing would be locked.
- 02
The model reads and drafts. Code decides. People approve.
The model reads the messy input and writes the first draft, ordinary code handles the exact math, lookups and rules, and a person approves anything that goes out or can’t be undone. Your people decide where those checkpoints go, and removing one takes an owner’s decision.
- 03
Done means proven
It’s done when it has run on its own schedule, we’ve checked the result where it landed, and your people have accepted it. A good demo doesn’t count.
No job is done because an AI says it’s done.
Who you’ll work with
Holden Richardson
Founder and principal
Richardson Applied AI is Holden Richardson’s practice, and he’s the person you work with, from the first conversation through handoff.
His father was a hands-on manufacturing operations consultant who moved machines, changed processes and standardized the work on the floor. He re-engineered the physical operation; Richardson Applied AI re-engineers the information operation. Holden runs his own businesses on the same method he installs.
- Based
- West Michigan
- Working sessions
- On site, with your people
The gap
Having the tools is the easy part.
Plenty of organizations use AI now. Far fewer have written down how it should be used or measured whether it helps. Understanding the job is the slow part, and it’s the part we do, in your building, with the people who know it.
23%
of US small businesses surveyed have a documented AI policy, while 61% say they are actively using AI.
Source: Pax8 SMB AI Pulse, Q2 2026 (402 leaders at companies with 5 to 499 employees), July 13, 2026
18%
of professional-services respondents knew their organization tracks the return on its AI tools, as organization-wide AI use nearly doubled to 40%.
Source: Thomson Reuters, 2026 AI in Professional Services Report, February 9, 2026
State of AI
A monthly read on AI, with the sources attached
State of AI is our monthly report for people who run companies. Every claim is dated and linked to its source.
All resourcesOctober 2026
Covering September through the 27th: lower model prices, agent risks, applications a business can own, and practical knowledge people can use and share.
Video
Your smart factory is failing at the interfaces, not the model.
The video argues that plant AI pilots stall where the MES, the ERP and the people on the floor disagree about what a part, a batch or a shift is, and that owning that layer yourself matters more than a smarter model.
Tell us which process keeps stalling.
Bring one real example, like a quote, a month-end report or a compliance file, and we’ll tell you whether AI will help and what we’d try first.
Conversations are always free.