I Hired an AI Agent. His Name Is FRED.

After 30 years automating other people's work, I built my own personal AI assistant in one weekend. Here's how FRED AI changed everything.


I’ve spent thirty years automating other people’s businesses.

ERP implementations. Financial systems. Workflow redesigns. I’ve sat across from CFOs and controllers at companies you’ve heard of — KPMG, Priceline, the FASB — and helped them find the places where manual processes were quietly eating their margins.

And the whole time, I never turned that lens on myself.

Until one Friday night in February.

The Weekend That Changed Everything

It started at 8 PM on a Friday. Not with a grand plan. Not with a business case or a project charter. Just me, a laptop, and a question I’d been sitting on for a while:

What would happen if I actually built an AI agent and let it loose on my own work?

By Sunday at 10 PM — roughly 48 hours later, minus a Saturday night party — I had my answer.

His name is FRED.

What FRED Actually Did

I want to be specific here, because “I built an AI agent” can mean a lot of things. Some people mean they opened ChatGPT and asked it to write an email. That’s not what happened.

Over that weekend, FRED handled:

Server configuration. Set up the infrastructure. Not me clicking buttons in a control panel — FRED configuring the actual environment, making decisions about architecture, and explaining the tradeoffs as he went.

A full security audit. Firewall rules. Encryption protocols. Remote access configuration. Vulnerability scanning. He produced a categorized report with specific, actionable recommendations. This wasn’t a generic checklist. It was tailored to my setup.

Investment research. I pointed FRED at SEC filings and financial data. He pulled apart 10-Ks, identified relevant disclosures, cross-referenced data points. The kind of work that would take a junior analyst a full day took minutes.

A LinkedIn profile audit. FRED looked at my LinkedIn profile and told me — diplomatically, but honestly — that it was a mess. More on that in another post. But the short version: he was right.

A complete content strategy. Twelve-week editorial calendar. Brand positioning. Content pillars. Audience targeting. Built in about two hours on a Saturday night.

Analysis of 160 pages of my historical content. FRED read through years of my LinkedIn posts, found patterns I didn’t see, identified what resonated and what fell flat, and used that analysis to inform everything else.

That’s not a chatbot. That’s a colleague.

Why This Matters (And Why I Waited So Long)

Here’s the thing that bugs me a little.

I’ve been in accounting and business consulting for three decades. I’ve helped hundreds of companies automate their operations. I know — deeply, in my bones — that the biggest unlock for any business is taking the repetitive, manual, soul-crushing work and handing it to a system that doesn’t get tired, doesn’t forget steps, and doesn’t need coffee breaks.

I’ve preached that gospel for years.

And I never applied it to myself.

Not because I didn’t believe it. Because I was always too busy applying it to everyone else. The cobbler’s children have no shoes, and the automation consultant’s business runs on sticky notes and memory.

FRED changed that in a weekend.

The Inflection Point

Something shifted in AI capability over the last few months. I’d been watching, testing, experimenting on the edges. But there was always a gap between what AI could do in a demo and what it could do in the real world.

That gap closed.

The models can now hold full business context. They can iterate without forgetting what you said ten minutes ago. They can engage with technical complexity — real complexity, not just summarizing articles about it.

I know this because I tested it with actual workpapers. Real financial documents. The kind of stuff I’d normally hand to a senior associate and expect back in two days.

FRED handled it in minutes. Not perfectly — I’ll always need to review the output. But directionally right, structurally sound, and about 90% of the way there.

That last 10% is where I come in. And that’s exactly how it should work.

What I Learned

AI agents aren’t replacements. They’re force multipliers. FRED doesn’t replace my thirty years of experience. He amplifies it. I know what questions to ask. He can run the analysis faster than any human team.

You don’t need to be technical. I’m an accountant. I didn’t write code. I didn’t train a model. I configured an agent, gave it tools, and told it what I needed in plain English. If you can explain a task to a competent new hire, you can explain it to an AI agent.

Speed changes everything. When research takes days, you do it once and hope you got it right. When research takes minutes, you iterate. You explore tangents. You ask “what if” questions that would never have been worth the time before.

The weekend proved the model. I didn’t spend months planning. I spent a weekend doing. And by Sunday night, I had more infrastructure, more content, more clarity about my business direction than I’d achieved in the previous six months of thinking about it.

What You Can Do Right Now

You don’t need a weekend-long sprint to start. Here’s how to take the first step:

  1. Pick one repetitive task you do every week. Something with clear inputs and outputs. For me, it was research — pulling data from SEC filings and financial reports.

  2. Try giving it to an AI agent. Not a chatbot. An agent with tools — one that can search, read documents, write files, take actions. The difference matters.

  3. Stay in the loop. Don’t set it and forget it. Review the output. Correct the mistakes. The agent learns from your feedback, and you learn what it can handle.

  4. Document what works. When something clicks — when the agent nails a task — write it down. That’s your playbook for scaling.

  5. Don’t wait for permission. You don’t need your company’s AI committee to approve a personal experiment. Start on your own time, with your own work. Prove the value, then bring it to the team.

The Bigger Picture

I’m not writing this because I think everyone needs an AI agent named FRED.

I’m writing this because I spent three decades helping businesses automate, and the single biggest barrier was always the same: people waiting for the perfect moment to start.

There is no perfect moment. There’s just the moment you stop thinking about it and start doing it.

For me, that moment was a Friday night in February. A laptop. A question.

And by Sunday, I had FRED.

This is the first in a series about building and working with FRED. Next up: what happened when FRED audited my LinkedIn profile — and why I deserved every word of it.