How I’m using AI right now

OK, so the robots are among us and the cat is quite unmistakably out of the bag. We’re constantly being told that while we won’t be replaced by AI, we will be replaced by people using AI - but what exactly does that mean?

Among all the LinkedIn posts about people who’ve automated their entire business and now make $7.6m every week while they sleep, it’s hard to know where to start. I’m by no means the most knowledgeable person when it comes to the use of AI, but I have tried a lot of things and am now in a pattern where I have several tools that work for the way I work - I thought it would be helpful if I shared them with you.

Here’s how I’m using AI right now:

Content:

By far the biggest time-saver for me is in content creation, editing and ideation. I’ve not got ChatGPT writing all my posts yet, mainly because I enjoy writing for its own benefits - clarity of thought, preservation of tone of voice and authentic connection chief among them - but I do use it as a kind of “research assistant” for my blogs. Usually I’ll come up with a topic, write out a bullet-pointed rough outline of what I want to say in my notebook, take a picture and upload it to a specially trained GPT that then expands the outline and fills it in with research, statistics and examples. I’ll then take the output and use it as a guide while I write the final post. (Ironically, this was one of the few posts I haven’t used that process for…).

I like having some striking imagery for these posts too (see above!), so I use Midjourney to create that - I find that asking it to create specific pictures “in the style of” artists I like is very effective. And yes, I know there are creative attribution questions around that, but we can come back to those in a future blog…

When I’ve shot videos, my team and I will edit them using Descript, which creates a transcript that you can use to edit the video by just highlighting text and cutting or deleting it, then we typically run them through Munch to get short, shareable clips, complete with subtitles, for our social posts. We’ll often use Capcut too to make them “pop” a little more.

Business Planning and Advisory:

At the end of 2024 I’d been playing around with creating custom GPTs to help with specific tasks, and the idea struck me that I could create an entire board of carefully trained AI advisors, give them the knowledge and tools of the people I admired most in each field, and consult with them regularly. I have some general ones (like an investment advisor) that I use for personal tasks, but I’ve put together a board of GPTs for our businesses that I “meet” with once a month to review our progress - a Chair, CFO, COO, CTO, Chief Marketing Officer and Chief People Officer, as well as a trained business and financial analyst.

These “advisors” have been incredibly helpful in identifying blind spots, scoping alternative perspectives and challenging my assumptions around various topics. I don’t ask them what to do, but use them as a sounding board and a way to get feedback on ideas and initiatives. As a side note, the investment advisor has really helped to sharpen my personal portfolio and although it’s hard to compare performance over the last six months to any previous period due to the ups and downs of the market, I have a much clearer picture of how my investments are balanced and correlated.

Admin:

A lot of people are deploying AI to cut their administrative tasks down, but this is probably where I’ve applied it least. We brought Fathom call recording into my company Singular last year, and that’s been great from a coaching perspective as well as in regards to clear, actionable notes/summaries following calls, and we’ll often take the transcripts from those calls and run them through ChatGPT to get insight on specific questions our clients might want the answer to.

Tool development:

More recently, I’ve been getting very into Replit as a way to develop applications that make life easier, like a reporting tool for our business, a proposal generator and a couple of other simple things. Using it in conjunction with our CTO GPT in ChatGPT has generated some really impressive results, although we haven’t yet built anything that’s externally available yet.

We’re also deep into the process of building and training Veriti, the custom LLM at the heart of our Versapia platform, and have been using her internally to support some of the consulting work we’re doing, which has saved huge amounts of time and generated some insights we may not otherwise have uncovered.

Things that haven’t worked:

We’ve run plenty of experiments to end up with a suite of tools that provide us with real value, and there have been lots of applications we’ve tried that really haven’t given us much for the time invested. We’ve found Copilot’s not very useful, and despite a company-wide experiment with Motion for time management, most of us abandoned it quickly after a period of initial optimism.

The key for us has been finding tools that either do things better than we can do them ourselves, or that slot into our existing processes. Adapting the way we work in order to incorporate AI has been less effective.

In any case, these are the things I’ve found most useful. What’s worked well for you? Let me know in the comments.

(If you know someone who could do with some pointers on how to use AI in their business, or someone who’s further along than me, send them this blog post and encourage them to sign up to Versapiens)