AI, 3 Chords, and the Truth - What I Got Right, and What Changed

By Rick Ataide, Blue Team Advisors
It's Tuesday night. You're having dinner and your phone buzzes. A customer wants to know when someone can come out. You could wait until tomorrow. You answer, check the calendar, text someone on your team, make yourself a note to put everything into your system tomorrow. You probably won't. Not because you're disorganized. Because you're running a business. Sound familiar? 

Maybe your version isn't a customer texting after hours. Maybe it's the proposal you're finishing at 10 PM. The salesperson who spends Friday afternoon updating the CRM instead of selling, or the question everyone asks Sarah because Sarah is the only person who knows how that particular thing works.

This is how businesses grow.
• Add people.
• Add software.
• Create workarounds.

Without designing it, we create ‘the way work gets done around here.’

Last December, I wrote an article called ‘AI, 3 Chords, and the Truth.’ I proposed that business leaders could think about AI through three simple ideas:

People. Outcomes. Data.
• Start with people.
• Focus on outcomes not technology.
• Understand information the business depends upon.

I still believe all three. Eight months later... I was asking too small a question.

Reimagine our Tuesday-night customer inquiry.

Why are you responding? What if AI helps write the response?

We give AI access to your calendar. Now it can suggest an appointment.

Why does someone have to copy the customer's information into your CRM tomorrow? Connect that too.

If the customer needs an estimate based on information your business already has, why does someone have to assemble it manually and another send the confirmation?

Keep pulling that thread and something interesting happens. We started by asking: How can AI help a person perform a task? We end up asking: Why does the work happen this way in the first place?

Then February Announcements Got My Attention

OpenAI said GPT-5.3-Codex was its first model “instrumental in creating itself,” helping with its own training, deployment and evaluation. The same day, Anthropic said, “We build Claude with Claude.” No, AI hadn't become an independent engineer. It was AI increasingly participating in building the next generation of AI. Days later, Coinbase introduced digital wallets for AI agents. Why would software need a wallet? Because there's a difference between recommending an action and having the authority to take it. An AI agent could select an approved shipping provider and pay the $185 charge without waiting for approval.

That's when the larger idea clicked for me. Most conversations begin with something like, "How can my employees use AI?" It's a reasonable starting point. But consider what happens when we take it further. Imagine a prospective customer contacts the business at 7:12 PM. Instead of leaving a voicemail, the customer gets an immediate response.

• The system determines what the customer needs.
• Asks appropriate questions.
• Schedules the appointment and sends confirmation.
• Updates the CRM.

By the time that employee starts work tomorrow morning, the administrative work surrounding that opportunity is complete. The employee hasn't been replaced. Something more interesting happened. The employee's job changed. Their greatest value isn't copying information between systems and sending calendar invitations. It’s their expertise, judgment, relationships and accountability.

Now multiply that idea across a business.
• Sales.
• Customer service.
• Accounting.

The interesting question isn't, “What tasks can AI automate?” It's, “How should this work happen now?”

Remember the Tuesday night event? The desired outcome wasn't “answer the text faster.” It was turn an interested prospect into a qualified, scheduled opportunity before that prospect went somewhere else.

That's what I mean by an outcome: the result the business needs, not the tasks people perform.

AI gives us an opportunity to question whether those tasks and who performs them still make sense. What outcome is this entire system supposed to produce? Ask how you would design it if you were starting today.

Find something important that is slower, harder, more expensive or more dependent on people than you think it should be. Then resist the temptation to immediately automate it. First ask: Why do we do it this way?

Many of us remember “The Jetsons.” It was someone's vision of the future. Flying cars. Robot housekeepers. Video calls. George Jetson still commuted to work and sat behind a desk. Think about that. The creators imagined technology transforming everything, yet they didn't reimagine work.

​Businesses are in danger of doing the same thing with AI, taking the company we already have, preserving every job, process, and approval... …and add AI. We'll have our flying cars, and we'll still be driving them to the same office. I have a name for that. The Jetsons Problem. That's where the conversation about AI and business gets much more interesting.