The Accidental AI Power User

TL;DR

• Every company has someone quietly using AI to move faster with spreadsheets, reports, macros, docs, and internal workflows.
• They may not call themselves a developer, but they are already doing lightweight software work.
• These people are exactly who companies should enable, not ignore.
• With the right AI training, they can become a serious productivity multiplier without needing to become full-time engineers.

Abstract

You have probably seen this person already.

They are not on the engineering team. They may work in operations, finance, sales, HR, customer support, analytics, logistics, or compliance. They are the person everyone goes to when the spreadsheet is broken, the report needs cleanup, the process needs a shortcut, or the team needs a weird internal workaround by Friday.

Then they start using AI through a chat box.

Suddenly they are writing better Excel formulas. They are generating macros. They are cleaning data faster. They are building little automations. They are asking the AI to explain scripts, rewrite reports, debug errors, and turn messy instructions into working steps.

They may not think of themselves as technical.

But they are already becoming something important: an accidental AI power user.

Table of Contents

• The Person Already Doing The Work
• Why This Is More Than A Productivity Hack
• They May Not Be Developers, But They Are Building
• Why They Eventually Hit A Ceiling
• What The Right Training Unlocks
• Why Companies Should Pay Attention
• Summary
• Next Steps

The Person Already Doing The Work

Inside many companies, there is someone who quietly keeps work moving.

They know the spreadsheet with twelve tabs and three hidden dependencies. They know which export from the enterprise system has to be cleaned before finance can use it. They know which report always breaks because one department names the same field three different ways.

Before AI, they were already improvising:

• nested Excel formulas
• pivot tables
• macros
• CSV cleanup
• copy-paste workflows
• reporting templates
• lightweight scripts
• internal docs
• process checklists

Now they have a chat box that can help them think through the work.

They ask:

• “Write an Excel formula that does this.”
• “Explain why this macro is failing.”
• “Turn these steps into a script.”
• “Clean this messy list.”
• “Help me compare these two exports.”
• “Write a formula that checks these conditions.”

And it works often enough that their output changes.

They become faster. They become more ambitious. They start solving problems they used to route to someone else.

That is not a small thing.

Why This Is More Than A Productivity Hack

It is easy to dismiss this as someone getting better at Excel.

That misses the larger shift.

This person is learning to translate business intent into executable logic.

They are not just asking for a formula. They are describing a rule.

They are not just asking for a macro. They are describing a workflow.

They are not just cleaning a report. They are formalizing a process that used to live in someone’s head.

That is software thinking, even if it does not look like software engineering.

AI gives non-developers a new interface into that kind of work. It lets them express intent in natural language and get back a formula, script, macro, checklist, or structured plan.

This is why English is becoming a new interface. The person does not need to know every syntax rule before they can start. They can describe what they want, inspect the result, and iterate.

They May Not Be Developers, But They Are Building

There is a category of work between “business user” and “software engineer.”

It includes people who:

• build spreadsheet models
• automate repetitive reporting
• connect tools with exports and imports
• write formulas
• adapt templates
• create internal processes
• document workflows
• build no-code or low-code tools
• ask technical questions better than their job title suggests

AI makes this group more powerful.

They can now ask for:

• better formulas
• explanations of unfamiliar code
• macro improvements
• simple scripts
• data validation checks
• report templates
• workflow diagrams
• automation ideas

This does not make them senior engineers. It makes them better operators.

And in many companies, better operators create enormous leverage.

They remove bottlenecks. They reduce busywork. They make reports cleaner. They help teams make fewer manual mistakes. They create small internal tools that save hours every week.

That is exactly the kind of person companies should enable.

Why They Eventually Hit A Ceiling

The chat box gets them started.

But without training, they will hit a ceiling.

They may not know:

• how to break a problem into smaller steps
• how to test whether a formula is correct
• how to spot hallucinated code
• how to protect sensitive company data
• when a macro is risky
• when a spreadsheet should become a proper internal tool
• how to document what they built
• how to ask AI for alternatives
• how to review generated scripts
• how to avoid turning one-off fixes into fragile infrastructure

That is the danger.

The same person who can save the company hours can also accidentally create a hidden dependency nobody understands.

The answer is not to shut them down.

The answer is to train them.

What The Right Training Unlocks

The right AI training does not try to turn every operations person into a software engineer.

It teaches them how to use AI tools with better judgment.

Good training helps them learn:

How To Describe The Problem Clearly

Better prompts start with better problem framing. What is the input? What is the output? What rule should be applied? What edge cases matter?

How To Ask For Stepwise Help

Instead of asking AI to do everything at once, they learn to ask for a plan, inspect it, then generate one piece at a time.

How To Test The Output

They learn to use sample rows, expected results, edge cases, and sanity checks.

How To Debug Safely

They learn to paste errors, ask for explanations, compare fixes, and avoid blindly accepting the first answer.

How To Protect Company Data

They learn what not to paste into a tool, how to anonymize examples, and when to use approved systems.

How To Know When To Escalate

They learn when a formula is fine, when a macro needs review, and when the work should go to engineering.

How To Document Their Work

They learn to ask AI to explain the formula, comment the macro, write a handoff note, or create a small internal guide.

This kind of training turns raw enthusiasm into useful leverage.

Why Companies Should Pay Attention

Companies spend a lot of time thinking about AI strategy at the executive level.

That matters.

But a huge amount of AI adoption will happen in the middle of the organization, where people are doing messy work with spreadsheets, reports, internal systems, and recurring processes.

That is where the accidental AI power user lives.

If you enable them, they can:

• reduce manual reporting work
• clean up data workflows
• build better internal templates
• automate repetitive tasks
• improve team documentation
• catch errors earlier
• prototype internal process improvements
• communicate better with engineering

If you ignore them, they will still use AI. They will just do it without shared standards, training, or review.

This is the same point I made in AI coding for internal software teams: the choice is not AI or no AI. The choice is trained adoption or unmanaged adoption.

Summary

The person using AI to write Excel formulas, debug macros, clean reports, and automate small workflows may not call themselves a developer.

But they are already doing the kind of work that makes a company faster.

With the right training, they can become much more than “the spreadsheet person.” They can become a productivity multiplier across the team.

You do not need to turn them into engineers. You need to teach them how to use AI with better judgment.

Next Steps

Look around your company for the person who already does this:

• fixes the spreadsheet
• cleans the report
• writes the macro
• figures out the workaround
• documents the messy process
• quietly uses AI to move faster

That person is not a risk to suppress. They are potential to train.

If you want to unlock that person, or a whole team of them, let’s chat. I run practical AI training sessions that help business teams use AI tools for formulas, macros, internal workflows, and lightweight automation without creating a mess your engineers have to clean up later.