I am terrible at fiddly work. My wife mocks me for it.
She says: “You might be big print, but I look at the little print.”
She’s right. I tend to move things at speed using a blunt instrument, but she adds the small, crucial adjustment that makes things actually work.
(If you haven’t seen the South Park episode where Randy builds an AI business and his wife has to clean up the mess, you really must.)
Details slip past me. I have ADHD so I’m always chasing the interesting stuff. I enjoy hosting dinners and events, but arrangements for them are dull.
For every City CIO Club and Board Table dinner, someone has to reach out to the attendees, ask them which of the menu choices they would like, collect the answers and send them to the restaurant.
You can do this with forms and spreadsheets, but the menu changes every time and there are always new email addresses to add, so it never quite becomes routine.
Each dinner costs me two to three hours of WhatsApping CIOs, sending emails, chasing the slow ones, checking the wines and place cards and emailing the club.
This week the job was pressing.
The Walbrook Club wanted the menu choices yesterday because the summer holidays had delayed the process itself.
At that point there is a choice.
I can do the work, or I can build a system that does the work.
The entrepreneur builds the system, but only after going through the whole process manually first. That rule comes from years of getting it wrong as a business owner. If you automate something you have never done yourself, you get a fast way to produce the wrong thing.
So I showed Claude how I had been doing it. I gave it the templates I send to attendees and told it where the finished list needed to go. It came back with a process flow that was wrong and a set of emails that read oddly, so I rewrote them and gave it a keyword to wait for before anything sends.
Nothing goes out until I have approved it.
My AI virtual assistant Marvina now has an email address. It writes to the attendees, asks for their choices, watches the replies, puts the names, titles and dishes into the spreadsheet and shares it with the Walbrook.
Building it took about 90 minutes, mostly because of permissions I had to grant along the way. That job has left my sausage fingers for good.
While Marvina was doing the dinner, I was with a second Claude building the new executive education page on TES and the partner page. Over the weekend I revamped The Proposition website as well, and I think it looks great.
A third Claude was helping me with interviews. I have done very few for TES and it is time I did more.
Journalists use platforms to send requests for interviews, comments, insights and product reviews. There are several around the world. You set up a profile, post the request and a flood of pitches arrives from PR companies and people who want to answer.
I had Marvin build a system that uses all of the platforms at once. It sends the enquiry, collects the responses, manages them, recommends which I should take and then sets up the interviews with the ones I choose. That gave me 20 interviews with excellent CEOs set up in about two hours. Next time it will take ten minutes.
A fourth Claude was working alongside that one, helping me identify the CEOs I most want to talk to and who would make a great conversation.
You get the idea.
Make a list of the jobs you hate and give AI some time to build systems around you. You can really leverage them well if you put some good thought into it.
But you have to firstly be prepared to build rather than do, and secondly, stick with it until you’re happy. That is the discipline and it can also be dull.
Once you get the first run done well, ask it to suggest improvements.
In partnership with Got rizz? Do you have the charisma to sell the CEO? The Proposition helps sales teams win deals and markets by changing the story and enhancing their C-level game. Take the customer drift test |
Subscribe
