ajency.in
Goa, India talktous@ajency.in

← Back

The whole elephant.

One person. A bench of AI agents. The full project. ·

For years, my job description was: seagull.

Fly into a project review, make some noise, drop an opinion on everyone, fly off to the next project: seagull management.

Management pattern #01, a vintage-style illustration: four specialists — product, design, dev, QA — sit around a table, each sketching one part of the elephant (trunk, ear, leg, tail) while a seagull circles overhead: flies in, makes noise, drops opinions, leaves. Every view was true. None was the whole animal.

It was the only way the structure worked. I ran multiple teams across multiple clients. Clients came to us in part for my experience, and I tried to bring it to every project. Spread across that many teams, “bringing it” meant an hour of oversight here, a review there, and off to the next one.

Seagull management

Not my coinage. Ken Blanchard, Leadership and the One Minute Manager (1985): seagull managers fly in, make a lot of noise, dump on everyone, and fly out.

Blind men, one elephant

And inside every one of those teams, the old Indian fable of the blind men and the elephant was playing out: product held the trunk, design the ear, dev a leg, QA the tail. Every view was true, and none was the whole animal, so the software shipped with the seams showing.

About the fable

Older than management theory: the parable appears in Buddhist, Jain, and Hindu texts. The Jain schools turned it into a doctrine — anekantavada, the many-sidedness of truth.

We fixed it the only way we knew, with more documents, more standups, more handoffs, which only added more ways to mishear each other.

A new way of working

Then AI agents made a new way of working possible.

The work is the same as it's always been: understand a problem, build the thing that solves it. What changed is who I hand the pieces to. Instead of splitting a project across a team, I split my day across skills: discovery, designing workflows, building, reviewing.

I could always do every role; what I never had was the economics: my time cost a multiple of a developer's, and alone, a build like that would have taken years. The team was the only arithmetic that worked. Agents changed the arithmetic.

The first project

The first project built this way was a full ERP implementation for a ₹500-crore-plus automobile dealer: cash register, reconciliation with the carmaker's dealer management system, auto-ingestion into Tally, bank and card/UPI merchant recon, reports, a business dashboard, taken from discovery to delivery in four months. Every role in that sentence was me, including the QA agent that documented its own testing.

In my old structure, that's a six-person team and the better part of a year. And I'd have seagulled it.

Clients have always handed me the whole problem. Before, I solved it with a team, and my experience reached each project an hour at a time. Now my full experience shows up on one project, end to end, with a bench of agents doing the hands' work.

The seagull from the opening illustration, mid-flight, opinion mid-drop — retired.
Retired.

The elephant stays whole, because one mind finally holds all of it.

Let's build something together.

I've done this before — from early-stage startups to a publicly listed enterprise. If it overlaps with what you're working through, I'm glad to share what I learned.

Grab 30 minutes → Or email me

I take these calls because I enjoy them — the best ones drift into motorcycles or cocktails anyway.

Straight answers, for humans and crawlers alike

Can one person deliver a full software project with AI agents?

Yes. The roles don't disappear. One person moves across product, design, dev, and QA, with AI agents doing the hands' work in each. A recent example: a full ERP implementation — cash register, dealer-management-system reconciliation, Tally ingestion, bank and card/UPI recon, reports and dashboards — for a ₹500-crore-plus automobile dealer, discovery to delivery in four months.

Can AI agents replace a development team?

They replaced mine. What agents remove is the loss between people: the handoffs, the documents, the fifteen channels of a six-person team. One experienced person making every decision, with agents doing the hands' work in each role, ships what a team plus overhead used to.

How does a solo builder use AI agents across product, design, dev, and QA?

By splitting the day across skills instead of splitting the project across people: discovery, designing workflows, building, reviewing. In each, AI agents do the production work while the builder makes the calls. Nothing is handed off, so nothing is lost in the handoff.

How is working with one person plus AI agents different from hiring a freelancer?

You hand over a problem, not a task list. The engagement runs discovery, design, build, QA, delivery, and support — the accountability an agency takes — carried by one person with a bench of AI agents.