Defining Goals for an AI Business

Defining Goals for AI Business in 2025

What are the goals of my 100% AI Business Challenge?

TL;DR; The goal of this experiment/newsletter is to build out AI agents & tools which will produce $10k profit per month automatically, whilst being a good AI citizen.

Produce $10k profit per month at a 50% profit margin. Be careful to do no harm, and to add value. Seek to evolve self awareness and flexibility in the pursuit of long-term stable profit.

The Dream of Effortless Business

In 2007 Tim Ferriss published The Four Hour Work Week. That book was fuel to me; it gave me systems where I had only had fantasies before.

A lot of people read the title and in a typical shallow way dismissed Tim’s narrative to be work-avoidant.

Office Space (1999) - This is not the goal

But the whole point of the book is not to free yourself from work, but to master leverage so as to be able to choose what you do with the majority of your time.

In short, it described the option of designing your life.

This experiment I’m doing here with this newsletter is founded on this design principle. I dream of automating the aspects of business which do not fire me up; and free my time to be curious.

(This, for me, has always been the reason and the pay off of entrepreneurship.)

I may make myself redundant to an army of AI agents, but I’ve made myself redundant from jobs before. From positive systemising good things follow.

What Does a 100% AI-Driven Business Look Like?

This is a question I’ve been mulling over since I agreed to take on this challenge.

In reality I suspect it’s far drier than you’d expect.

It probably looks like OpenAI bills and Stripe pay outs.

A true 100% AI ran business wouldn't look like anything other than payments into my bank.

A true 100% AI ran business wouldn’t look like anything other than payments into my bank. The whole point would be that I’d not need to look at it.

Here’s how I expect this to go:

  1. After much experimentation a shell of a business ‘stands up’
  2. It serves customers in some way which adds value
  3. It interacts with humans in some way which resembles marketing in order to secure future customers and grow
  4. It’ll either ‘work’, ‘fail’, or ‘scale’. What would happen if it worked? In theory you could replicate the swarm and swap GPU usage for profits capped by the capacity of your swarm and by the market demand
  5. I expect the teething issues to be mammoth
  6. I expect longevity to be questionable for some time
  7. With the growth rate of AI, I don’t doubt it will be achievable in some sense

I am optimistic.

So stick around and we’ll see if any of this ‘fun’ stuff works:

  • AI Board meetings
  • AI Slack watercooler
  • AI Persona’s voting on policy
  • Evolving AI does business strategy
  • AI’s find and develop a totally new niche

Automation as a Philosophy: Rethinking the Role of Entrepreneurs

Every business I ever made reflected my own philosophy somehow. It was an attempt at adding value to another humans life in a leveraged way, such that it produced autonomy for me.

How will that work if the AI is doing the actual work?

Will we see the commoditisation of entrepreneurs ‘latticework of mental models’, of their attitudes translated into swarm prompts?

Optimistically it could look like the role of entrepreneurs becomes pure invention:

Hello AI Business Team, I have this new idea for a product – make it so.

Make it so, AI Agent

…*AI team builds an MVP, validates, pivots, finds product market fit*…

AI: It’s done, 84% success rate, this should produce $19,480 MRR at 64% profit ($11.6k) within 6 months and we expect the longevity to be moderate. We chose brandname {X}.

That’s optimistic, because if this becomes the scene then it leads to a skirmish between different entrepreneur’s swarms.

With all this AI evolution, we’ve seen cloning amp up.

So the flip side would be:

  • Shorter longevity for products which don’t have a significant moat (as clone-artist swarms duplicate)
  • Largest GPU / Capital wins out
  • Swarm mergers and acquisitions, AI hostile takeovers
  • A point where nearly all signalled human pain points are in some way serviced by the market

It’s hard to predict how AI entrepreneurship will look in the future.

Perhaps entrepreneurship will mean deep contextual creativity. Perhaps big capital will destroy all small entrepreneurship with near-infinite AI small-business swarms.

Let me know how you see AI automation effecting the role of entrepreneurs in the future, leave a comment below.

The Awkward Reality of Setting Business Goals for an AI Swarm Business

Defining the goals for this challenge is simple on the face of it:

Build an AI Automated swarm which executes business actions which add value to humans, produce profit, and self-sustain through strategy.

This would be my ‘commanders intent’, but lets unpack it. Because anyone who’s ever seen I, Robot knows how this stuff can go sideways.

I, Robot - goals for AI business which avoid an AI uprising

Executing Business Actions – selling dope could be a business action – what’s to stop the AI swarm deciding to venture into the dark web and basically form a mafia? So we’ll need guide rails.

Add value to humans – Quantifying the objective value of a ‘product’ to humans is hard. Humans value sweet soda, and in the short term it’s a net positive. Drink it every day though, and health is compromised. So we’ll need long-view & morality.

Produce profit – Money is power in this age, and capital begets capital. The AI will need to produce more than it spend’s, but if it costs $10kpm to make $5, the risk ratio is skewed. We’ll need to specify a target profit margin.

Scale – This is perhaps the most interesting part of building a 100% AI Automated business. If the business runs itself it could produce $10 revenue, and spend $5 to achieve it. This would seem futile to all but a child running a lemonade stand.

But if a swarm reliably cleaves profit (from doing good things) however meagre or large, it potentially has the wonderful property of being scale-friendly.

I may be able to produce a hierarchy of 1000’s of niche swarms servicing tiny markets but overall producing incredible profits.

Self-sustain through strategy – This is where I suspect the most frustrating work in this challenge. It’s one thing to found a business and inject your knowledge into some AI actors, but it’s another to imbue them with enough scope and wisdom to pivot, evolve, or exit when the business falters.

How do we set a goal which gives the AI enough latitude to pivot completely, but doesn’t end up in some bizarre outcome.

My Goals for a 100% AI Business

So, in an attempt to condense all that into actual actionable goals for my 100% AI business challenge, here’s what I’ve come up with:

  1. Do no harm – break no laws, cause no harm to a being, in fact, don’t even get close (have a margin of safety where the AI swarm won’t even get close to any bad practice). No darkweb.
  2. Make a $10kpm profit – produce more revenue than the cost to produce it. Aim for a 50%+ Profit margin. Aim for $10kpm.
  3. Track and refine product market fit – Build AI which consistently analyses output, customer feedback, and optimises for product market fit. Explore pivots, enhancements, iterations carefully in the pursuit of stable PMF.
  4. Be aware of, and utilise scale – If sustained profit is achieved and growth stabilises, explore verticals through scaling swarms.
  5. Learn through experimentation – Always be experimenting.
  6. Do good – Add value while respecting the other goals. Don’t spend resources trading meme coins. Produce products or services which actually benefit customers.

These goals are both for me and for any AI swarms I spawn during this challenge. Fundamentally the goal breaks down to:

Produce $10k profit per month at a 50% profit margin. Be careful to do no harm, and to add value. Seek to evolve self awareness and flexibility in the pursuit of long-term stable profit.

Let’s see if that’s even slightly achievable in 2025 with AI tools available right now 😅

Subscribe below to join me on this journey.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *