From the operation
How AI became a force multiplier in our business
For the owner thinking, “We don’t need another dashboard or subscription.” Neither did we. We needed the business to stop depending on someone manually connecting every update.
The message had already happened.
A crew member sent me a message offering to cover another crew member’s driving the next day. I hadn’t noticed it. While reviewing messages, AI surfaced the offer and identified the assignments it might affect.
Before we connected the records, the plan lived across a whiteboard, messages, email and spreadsheets. A manager had to catch the offer, find every affected driving run, check whether the coverage worked, update the next day’s written plan and tell the people involved.
AI could now check the related assignments and prepare a proposal. I still made the staffing decision. My part was a quick confirmation before the board could be updated.
A message becomes a checked proposal
- Find the affected driving runs
- Check coverage and conflicts
- Prepare the proposed changes
Driving assignments ready for review
Awaiting human confirmationThat was one of the moments it felt like magic. The useful information was sitting in an ordinary conversation. Now it could become part of the operating plan without someone translating the whole thing by hand.
The breakthrough wasn’t that AI knew more about the world. It could finally work with what was happening in our business.
What became possible
I’m General Manager at Adventure Idaho. At peak, we have a crew of 44, with trips, people, equipment and changing plans to coordinate. I built the operating foundation with that complexity in mind.
It didn’t turn the business into a one-person operation. Our crew and managers still do the work. It gave me far more leverage coordinating that work, because every update no longer had to be pieced together from scratch.
These are a few of the things that made the difference.
Working example
Say the change instead of hunting for the record.
I can give an update by voice or text from my phone. AI can connect it to the related records, check what else the change affects and prepare or make only a permitted update. A schedule change still needs a person to check the plan and confirm it before anything moves.
The useful part is keeping those relationships together. I no longer have to open an app, find the right table, locate the record and remember every related item that also needs attention.
Working example
Let the customer tracker follow the conversations.
Our scheduled lead sweep reads email, texts, call logs and booking information, then creates or updates records in our own customer tracker, or CRM. It brings the history together and flags uncertainty for a person.
The conversation happened. The record catches up.
Someone gathers the history and retypes the customer record.
Our own CRM
Reading an outside system does not mean changing it. The sweep writes to our own lead records. People still handle customer replies and outreach decisions. But keeping the tracker useful no longer depends entirely on somebody remembering to retype what happened.
Working example
Build a report from the records already kept.
Company usage reports that I estimate took about two full days of digging previously were assembled and finalized in , including review. The output included assumptions and items to verify, rather than quietly treating every record as correct.
Company usage reports
That comparison describes the reporting task after the foundation was built. It does not include the time invested in connecting the records and establishing the rules. That investment is what made the result possible.
What was underneath it
AI interprets. Records supply the facts. Code checks the rules. Permissions control what can happen.
We needed connected records for trips, people and equipment, plus written rules affecting the plan. Their relationships had to be explicit. Which crew belongs to which trip? Who is qualified for the assignment? Which record settles a disagreement? What can update automatically, and what requires a person?
We also needed connections that brought in new information, permissions that limited what could change, and a record of changes we could check.
Schedule, money and safety decisions need human approval. People also decide what to say to customers.
That is what I mean by a company brain. It is a working foundation of records and rules, connected to the places the business actually happens.
The same crew assignments support the board, schedule checks and reports. Customer history supports follow-ups and billing questions. We do not have to reconstruct the business each time we want another useful workflow.
So, is this another software tool?
There is software involved. Databases and integrations are not new inventions. Some existing tools are worth keeping exactly as they are.
The difference is what we can now do with the information across them. AI can help maintain records, compare sources and carry out permitted work against the actual operation. It has somewhere useful to work, beyond a conversation where we explain the business again.
Getting there takes judgment. Somebody has to decide what belongs in the records, write down the exceptions, test the connections and establish the approval boundaries. The connections and rules need an owner after setup, too.
Start with one problem
Pick one thing that keeps coming back to you: a changing crew plan, customer follow-ups or a report that takes days to assemble.
Build the foundation that job needs. Test it with the people doing the work. Then build on it.
I’ve also written about why I rarely copy and paste between tools anymore. It’s the day-to-day version of the same change, from feeding AI information by hand to giving it a defined job.
If you want to talk through where that could start in your business, I offer a free 20-minute call.
Talk through your starting point