How I’m using AI (not how you think)

Read Time: 5 minutes

Welcome to BE THE EXPERT! In today’s issue:

• How I am using (and not using) AI
• Why systemizing is important
• AI can “think” now???

How I’m using AI (not how you think)

This issue was NOT written using AI (and never will be).

However, I do use AI for a lot of other things in my expert business, so I thought it would be helpful to show how and why.

Right now, I put all the AI tools and services into two buckets:

• Things that may accelerate or improve the ROI on what you’re already doing
• Things that may enable net new solutions or services

AI to accelerate or improve the ROI on what you’re already doing

This bucket assumes you’ve already at least notionally systemized your business, as I describe in ​BTE12​.

You’ve defined systems, processes, and procedures for your business’s key functions, which are executed by you (and/or your team) and supported by relevant tools and technology.

Right now, those are (hopefully) generating a return on investment, and you are achieving some level of productivity.

As you consider the tsunami of AI news, tools, and services we’re currently experiencing, it’s important to have a quick method for deciding whether to explore a given option.

It’s also important to have a process for avoiding FOMO and changing tools every two weeks when a new one comes out.

How do we do that?

I take this category (improving what we’re already doing) and break it down one level further:

• Things that improve a team member’s or role’s personal productivity
• Things that improve team productivity
• Things that improve business productivity

As a business owner, I’m always trying to see if I can get more out of each individual without burning people out, whether I can get more out of our team working together, and whether I can shift more work to lower-cost resources over time and open up capacity for new work.

Here are a few specific examples of how we use AI to do that and how you can, too.

Personal Productivity

Every three months or so I try to look at each role and team member and the processes I have them executing.

Then we chat about what the most inefficient or time consuming parts are to see if we can improve them.

In our done-for-you content engine service, two examples are finding and selecting good clips from content for repurposing and then writing captions for the hundreds or thousands of social media posts we create for clients.

Those are obvious candidates now for AI. I picked this example though as AI does only ONE of those two well (so far).

With the right prompts AI is great for generating the captions.

Right now, we’re using AI (in this case inside of Descript) to generate the captions for all the social posts we create since it does a good job.

That has dramatically improved the productivity and ROI on our content repurposing process both internally and for clients.

That improved one of the two slowest areas.

But all the tools (Opus, Descript, etc.) are poor at picking good highlight segments.

We still use humans to find and select clips, as the tools aren’t there yet.

That said, every week, there are new ones or improvements.

I’m excited to try the new OpenAI models that “think” and reason better to see if we can improve and/or automate the selection of clips.

So, we have one example of where AI helped dramatically improve one of the more expensive and time-consuming parts of one of our key processes.

Then we have an example of how AI is not there yet, but where we are keeping a close eye on it to pounce as soon as there is something that meets our needs.

Team Productivity

The next area to consider is productivity across an entire team.

With each team member or role enabled by relevant AI tools and services, are there areas spanning the team that could be improved or accelerated?

A simple example here is AI for recording, transcribing, and summarizing meetings. We use Fathom for this in our Zoom meetings.

We get summaries and action items that go into our Expert Business Operating System in Notion where we manage and track all our work and deliverables.

A second example is Cassidy AI, where we build AI agents, assistants, and knowledge bases so that the team uses those and generates consistent output by using the same agents/prompts and our specific content.

In your business, think about the bottlenecks between different people and processes.

Those are areas where AI and automation may be able to help. Ask your team members if there are frequent cases where they are “waiting on XYZ” or “re-working on something due to miscommunication.”

In our case, by consolidating down to two critical tools, Descript and Notion, and leveraging the AI within them, we were able to streamline our repurposing process significantly.

Business Productivity

The final bucket for improving what you’re already doing is at the business-wide level.

If you offer an article writing service for businesses as an example, you may shift your entire business model from humans writing articles to AI writing articles with humans editing/improving them.

In our case, it is the combination of personal and team productivity that accumulated to a business-wide improvement.

How?

We were able to take out an entire role in our process because of the improvements mentioned.

We used to have three roles for repurposing: Someone to find and select highlights, someone to create the assets (highlights, quotes, etc.), and someone to draft all the social posts, captions, etc.

With AI, we were able to take that caption step, have role 2 handle it, and not need role 3.

Even better, role 2 does not need to be fluent in English or understand what makes for a good caption or social post since AI is effective at that part.

Role 1 needs those skills (for now) to pick and edit the highlights, but as soon as the AI models and tools get there, we’ll be able to simplify that part too.

Those examples show how personal and team improvements with AI can add up to a significant business improvement.

In our case, we’re able to fulfill one of our key services with only two-thirds of the staff and with a less expensive resource for one of the roles.

Better still, the quality of the results to our clients has also improved.

Once you’ve AI-enabled what you’re already doing, there may also be opportunities for net new products or services.

AI that enables net new solutions or services

The other main bucket of use cases for AI is net new products or services you may decide to offer.

You may go through the steps outlined above and create something great for your existing business internally and then decide to offer it as a service.

Pretty much every service my agency offers starts as something we build internally and then package up as DIY / DWY / DFY services.

With the content and AI tools we use becoming easier and easier for even non-technical people to use, we have been able to take our done-for-you content engine services and turn them into done-with-you and do-it-yourself options as well.

With the AI workflow platforms we use (our ACCELERATE platform) we can offer those as part of the platform in our SaaS business model.

In your niche, you may be able to find or create AI tools, services, coaching, or other elements to sell or affiliate with.

Those would be net new solutions you could add to your business model.

You may have ideas that previously would have been too costly or complicated to deliver but that, with AI’s help, might now have a good enough ROI to deliver.

Take Action

Consider setting aside a few hours at least once a quarter to review the steps I outlined in this issue.

• If you haven’t systemized your business, do that first.
• Then, if you have team members, talk to them about bottlenecks, slowest processes, etc.
• Consider AI options at the individual, team, and business level that would improve productivity and ROI
• Consider ideas for net new products or services that AI may open up in your business or niche

Saturday Spotlight 💡

Content Spotlight:​

In this video, I break down the exact steps you need to take to analyze, design, document, automate, and delegate your way to a fully systemized operation.

Tool Spotlight

​​Cassidy AI​​– Build AI automations powered by your content and data

Cassidy is one of my favorite AI tools and one we use daily in my team.

With Cassidy, it is one click to change between AI models when one comes out that is cheaper, better, or both.

It also enables us to create AI assistants and workflows integrated with our entire knowledgebase and content library in Notion.

News Spotlight

​Introducing OpenAI o1-preview​ – OpenAI

We’ve developed a new series of AI models designed to spend more time thinking before they respond. They can reason through complex tasks and solve harder problems than previous models in science, coding, and math. Today, we are releasing the first of this series in ChatGPT and our API. This is a preview and we expect regular updates and improvements.

This one is VERY interesting. Remember above where I said the current AI tools and models aren’t good at selecting great highlights from videos?

Well right before I published this issue, I tried the new OpenAI model on one of my video transcripts and it picked about 2/3 of the SAME highlight segments I did and the other ones it selected were pretty good.

As it processed, you could see that it was “thinking” and not just spitting out the first ones it could find.

It also stuck to my instructions and was able to understand timestamps in transcript files AND comply with the length I specified.

This reinforces the importance of keeping up with the advancements!