Operation Monorail

Produced by DALL-E

”Bart: True or false: You can get mono from riding the monorail.

Homer: False. No wait, true!”

Marge vs The Monorail, The Simpsons

Previously at the Volcano Base, there was mainly fear as I started to document the car-crash that my career has turned into. It’s been a long week since then, but I’m feeling more optimistic.

Mission Briefing

Operation Monorail

Last week, I’d started thinking about what I could possibly do to make money. I knew I had skills and experience that should be valuable to someone out there. Because I’d been focused on “applied artificial intelligence” for the last few years, I thought maybe I could create agents for people (AI systems that can use tools to achieve an objective).

But I had everything backwards. It was panic causing me to ignore all the advice I’ve given to clients over the years: don’t start with the technology!

I spent the rest of the week trying to formulate something simpler.

Jobs-To-Be-Done

I don’t want to create a big business. I want something small. Cozy, even. While I can do a lot by myself, there are aspects of running your own thing that are a real drag - mainly anything to do with admin.

So I started thinking about how I could run my micro-business by myself, but with the assistance of a set of AI agents. A synthetic team. It’s totally plausible to automate the hell out of whatever I want. If there’s a project that requires different skills to mine, I know a lot of people that can help.

In order to achieve this way of working, I’d need a way of keeping my costs to an absolute minimum because, well, I’m broke and have no income at all.

I’d need to pay particular attention to standardising operations, so that automation can do its thing, and free me to do the non-standard things. I’m fast at learning new things, so I wasn’t worried about that part of the equation.

Hmm, but I’d need some sort of “hub” or “command centre” to manage the automated processes… STOP! I was off thinking about the technology again.

In order to define a relevant offer for small businesses, I needed to know more about their pains and problems. Ideally, I wanted to know their jobs-to-be-done, and only then see if I had anything relevant for them.

Without time or a research budget, I decided to use tech to help me out. I lurk in Reddit’s smallbusinessuk sub, and they’re always talking about their challenges. So I scraped their top concerns and analysed them for theme and specific JTBD. Method for that is in this video if you want to try it. Took 30 mins and cost $4.

Yes, there are more “robust” ways to do this, but I had neither time nor money.

Scratch your own itch

What I discovered was that their JTBD were very similar to mine:

  • reduce operational costs
  • increase profit margin
  • generate leads
  • ensure customer satisfaction
  • maintain standard operating procedures
  • make better use of digital tools
  • address skill shortages
  • and a whole lot of queries about taxation that I’m not going to be able to help small businesses with

This was a lightbulb moment for me. If I were to hit these companies with an offer to create AI agents for them, even if they were interested, they wouldn’t be able to get the benefits. They needed something more foundational first. I was in danger of trying to sell monorails to people that needed a bicycle.

Hypothetical offer

So I put together a hypothetical service offer:

  1. Establish an operations “hub” (get all the business basics in one place, ready for automation)
  2. Automation (eradicate admin and busywork)
  3. AI staff (for those ready for the next level)

I needed all these for Volcano Base too, so I finally allowed myself to go deep on the technical practicalities.

An operations “hub” would need CRM, lead capture and management, some marketing tech (like for newsletters, transactional comms, social posts), a way of managing communications across channels, appointment booking, estimates, contracts with doc signing, invoices, surveys and forms, ideally a way of hosting short training assets or a knowledge base, a way of supporting customers…

That was going to cost a fortune and a hell of a lot of systems integration! I’d need subscriptions for things like calendly, mailchimp, google workspace, typeform, docusign, hubspot, and then zapier to tie it all together. And that’s before even thinking about monthly costs for OpenAI and Anthropic.

But I found a way

I found a way to do it all for a fraction of the cost, and that’s what I’m going to be offering 1-5 person UK businesses.

After their operations hub is set up, I’ll help them automate their processes and even create AI staff for them if they want.

It’s still hypothetical, so I’m talking to some small businesses about doing this for them for free to test it out. It’ll be important for me to understand the typical objections, setup practicalities etc. But I’m feeling optimistic.

Classified Intel

Some interesting stuff I discovered on my adventures.

Understanding AI Agents (Without The Hype)

AI agents are entities that can perceive their environment and take actions to manipulate it. While they can enhance business processes, they require careful design and should not be seen as a quick fix for all tasks.

Apify

Apify is the largest ecosystem where developers build, deploy, and publish data extraction and web automation tools. They call them Actors.

ChatGPT Video Feed

The ChatGPT mobile app can now interpret what it sees through a real-time camera feed.

> END OF LINE
Operation Monorail
Older post

Fear is the mind killer

Starting to document my crash-and-burn.