How we work
One process at a time, with the people who do the work.
An engagement starts with a free conversation and is designed to end when your people can run what we built without us. In between, we work at the desks of the people who do the job and prove each process on real work before we start the next.
- Starts with
- A free conversation
- Built
- One process at a time
- Ends
- When your people run it
- After
- Optional Cadence, month to month
Step by step
How an engagement runs
There are nine steps, and each one ends with a decision. Not every engagement needs all nine. The feasibility test only happens when the outcome is uncertain, and Cadence, the ongoing support after the handoff, is your call.
OP 010
A free conversation
You getAn answer, including “not yet” and what would change it.
You tell us what’s going on in your own words, and we walk through one real process together: what arrives, who touches it, what comes out, and how often. Then you get an honest read on whether AI is worth your money here.
OP 020
Working sessions with the people who do the work
You getAn agreed first process, and a plain description of what a better result would look like.
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 around a whiteboard, and ask them to show us the last one they did. We write down what everyone just knows, what your systems already export, and where someone’s judgment decides the outcome. We also ask what should never be automated here, and we honor the answer.
OP 030
A scoped first project
You getA short written scope you approve before any work begins.
We write the first project down in plain language: the process, the result we’re aiming for, what’s excluded, how we’ll test that it’s done, and a limit you approve before work starts. We also agree with you on how to measure the process as it runs today, before anything changes, so no result gets claimed without that baseline. We recommend the smallest thing that will last, and that can be as small as a written procedure or a scheduled job inside the systems you already use.
OP 040
A feasibility test, when the outcome is uncertain
You getWritten findings that end in a recommendation to go, change course, or stop.
Some work can’t be judged from a conversation. When the outcome is uncertain, we propose a small, paid test on your real documents before a bigger commitment, checked against examples your team already understands, including the awkward ones.
OP 050
The foundation, in your company’s name
You getAccounts, passwords and an Operating Standard your company owns. Our access is limited to what the work needs, and it’s written down.
Before we build anything, we set up what it stands on, in your company’s name: accounts with at least two of your own people as owners, passwords in your password manager, and a separate login for each person and each automated job. Then we start your Operating Standard, the written record of how the work gets done. We bring in whoever handles IT for you on anything that touches your network, email or security.
OP 060
Build one process at a time
You getOne working process, measured against how it ran before, with its written instructions stored next to it.
We build with the people who do the work, on real jobs, and they decide where a person checks the result. Those review points stay unless you decide in writing to remove one. Working sessions happen on site, and some of the build work between them runs remotely, inside your company’s accounts. When a new way replaces an old one, the two run side by side until the results match, and the old way keeps running until your people accept the new one.
OP 070
Training your team
You getPeople who can use the system, and Champions who can change it.
We train your team inside your company, in the room first and then at each person’s desk. Each person gets set up on their own computer with their own login, and we check that the system answers a question we already know the answer to and says “not found” when the answer isn’t in the Standard. One or two people become Champions, the people on your team who own the system after we leave.
OP 080
Handoff, proven by what your people can do
You getEverything in your name: the code, the written instructions, the settings and the accounts.
The handoff is a test, and your people take it. They find the Standard and the live systems, trace one input through to its output, stop a run, fix a failure safely, make a reviewed change, and run everything without Holden’s computer or accounts. A good demo doesn’t count as a handoff.
OP 090
Cadence, ongoing support if you want it
You getAn agreed amount of time each month, with a limit you approve.
Cadence is our ongoing support. After the handoff you can keep it as a monthly arrangement for upkeep when models or document formats change, measurement reviews, more training and new work. It’s month to month. If your Champions have it handled and you stop, that’s the system working.
Limits
What we won’t do
- Lock you in. Everything is built in your company’s accounts, and support after the handoff is month to month.
- Pitch AI as a way to do the same work with fewer people. We’re after better output, and more of your people’s attention on the calls only they can make.
- Hand you a slide deck and leave. We build a working system inside your operation and write down how it works.
- Automate what isn’t worth it. A process that runs a few times a year rarely pays back the setup, and we’ll tell you so.
- Put an approval gate on everything. A person approves what can’t be undone, like a quote going out or a job being released, and the rest runs without anyone clicking “approve” all day.
- Take over your ERP’s job. The ERP stays the system of record. We read from it, and anything written back goes through a path you’ve approved.
Progress
How you’ll know where things stand
While the work is underway, you get a clear view of progress and a short written update on a regular schedule. Holden reads every update before it reaches you.
- Progress is reported item by item against the scope you approved.
- Every item is done, in progress, at risk, blocked or not started. We don’t make up a percent-complete or a deadline.
- Each update names the decisions we need from you and who is waiting on whom.
Pricing
How pricing works
Every engagement is scoped independently, so we don’t publish prices. Conversations are always free. Paid work, like an Opportunity Assessment, a feasibility test or a first project, is scoped in writing first: what the work is, what’s excluded, and a limit you approve before it starts.
Work stops at that limit unless you approve more. Cadence, the ongoing support after the handoff, works the same way, with an agreed amount of time each month that you can cancel any month.
FAQ
Common questions
Is it just Holden? What happens if he’s unavailable?
Yes. Richardson Applied AI is Holden’s practice, and the person in your first conversation is the person who scopes the work and builds it with your team. That’s part of why everything is built in your company’s accounts and written into your Operating Standard, and why the handoff is a test your people take: your Champions have to show they can run the system, fix a failure and make a change without his computer, accounts or help. If he’s unavailable partway through a project, what’s already built is in your accounts and written down, not locked up on his side.
How long does it take?
It depends on the process and on what’s already in place, so we don’t give a blanket timeline. The first project covers one process, and its scope says how we’ll test that it’s done. You see it working on real jobs before the next one starts, and our progress updates never carry a made-up deadline.
What do our people have to do?
Show us the work, and check it. The people who do a process walk us through it, decide where a person reviews the result, and test the new way on real jobs. One or two of them become Champions, who own the system after the handoff. Plan on their time in working sessions and training, and on real jobs while each process is built.
Do we need an IT department?
No. If you have one, or an outside IT provider, we bring them in early, and anything touching your network, email, logins or security goes through them. What you do need is a person inside who can own the system after we leave.
What happens to our data?
What we build runs in your company’s accounts, not ours. We choose the AI tools and their settings with you, including how each provider handles your data, and sensitive information is kept separate by who’s allowed to see it. Confidentiality terms, including what we keep on our side for the project, go in the written agreement.
Conversations are always free.
Step one is a conversation about one real process, and we’ll tell you whether it’s worth working on.