Perspectives · November 17, 2026
Get your operation ready, not your team
Most companies try to bring AI in through their people. The bigger opportunity is in the clumsy operating process those people are stuck inside.
By Equipo Baguals

Ask a traditional company how it plans to adopt AI and the answer usually sounds like a training program. Licenses for everyone. A few workshops on writing prompts. Maybe a champion in each department who shows the rest how it’s done. Six months later, a handful of people use the tools every day, most don’t, and the operation looks exactly like it did before.
That outcome isn’t a failure of the people. It’s a failure of where we aimed.
The team already has a job
The people who keep a company running are busy. They answer customers, chase suppliers, type invoices into the ERP, reconcile spreadsheets and confirm deliveries over the phone. Their day is full before anyone asks them to learn a new skill.
When AI adoption depends on each of them changing how they work, it competes with everything else on their plate, and it usually loses. Not because they resist change, but because the incentive is weak: a faster way to write an email doesn’t remove the email, the spreadsheet or the phone call that comes after it.
Training has its place. Curious people should get good tools. But training alone turns AI into a personal productivity hack, and personal hacks don’t add up to a different company.
The real bottleneck is the process
Look closely at almost any operation and you’ll find the same pattern. A piece of information arrives in one place, usually an email, a PDF or a WhatsApp message. Someone reads it, types it somewhere else, checks it against a third system and then tells a fourth person it’s done.
Every one of those handoffs is a wait. Every one is a chance for a typo, a missed message or a case that falls through the cracks. And every one depends on a person remembering to do it.
That is what we mean when we say the operation is clumsy. It isn’t that people are slow. It’s that the work is shaped around moving information by hand between places that don’t talk to each other. No course fixes that, because the problem isn’t in anyone’s skills. It’s in the shape of the work.
What a ready operation looks like
Getting a company ready for AI means reshaping that work so agents can do the repetitive parts, while people keep the parts that need judgment.
In practice, that looks like a few concrete things:
- Rules live in the data, not in someone’s head. If a price depends on the route and the weight, that rule sits in a table the agent reads, not in a note on someone’s desk.
- Every action is on the record. When an agent answers a customer or updates a system, there is a log that shows what it did and why.
- Sensitive decisions wait for a person. Moving money, changing customer data or committing to a price needs a human approval by default.
- Agents show up where work already happens. Customers keep writing on WhatsApp and email. The agent meets them there; nobody has to learn a new app.
None of this requires your team to become AI experts. It requires someone to sit with them, understand how the work really flows, including the exceptions nobody wrote down, and build the agent around that.
What changes for the people
When the operation is ready, the team’s job shifts. People stop retyping and chasing. They start supervising: reviewing the cases the agent flags as doubtful, approving the actions that need approval and handling the exceptions that need a human conversation.
That is a real change, but it’s a gradual one, and it happens inside the work they already do. Nobody has to change careers overnight.
There is also something the company gains that it usually loses today. Every case an agent handles leaves a trace: which rule applied, which exception came up, how it was resolved. Over time, that becomes a living memory of how the company works, one that belongs to the company and doesn’t walk out the door when someone changes jobs.
Where to start
Don’t start with a company-wide AI plan. Start with one workflow that eats a lot of time and can be measured: document entry, a common customer question, a round of follow-up calls. Agree on how you’ll judge it before anyone builds anything. Then get it into production and see what happens.
The company doesn’t need its people to become AI experts; it needs an operation ready to work with agents. That is a much more achievable goal, and it’s the one that actually changes how the company runs.


