The graded approach

Yesterday I mentioned how the way people are selling AI based offers simply isn't good enough.

They're selling these complete systems that are pretty high level, to people just dipping their toes.

The result is people getting stuff they don't know how to use.

This is something I've noticed with a lot of the people I've been talking to and my own customers recently.

And it got me thinking that there has to be better way to do this. A way where people are not only getting the result, but are increasing in knowledge and confidence as well.

So, I wanted to break down my new offer stack with the reasoning behind things.

Step 1 - Building the foundation

Despite all the hype around AI, most people are still using a basic ChatGPT account and treating it like a chatbot - then getting pissed when it can't do the job.

The people doing interesting things though have a proper architecture around their AI systems.

They have a tiered file system that includes things like...

• Claude.md instruction file telling the AI how to act
• Brand context so it knows who you are and what you do
• ToV guidelines so, when it does create something, it sounds like you
• Skills files with specific abilities and skills trained into it's processes

... just as a start there as well.

Why most people don't get results formthe skills you sell them is they don't have the framework.

So, my thought is the first thing I need to do, is help people set up the right environment for the AI to thrive.

Which is why I put a free offer together about the foundational AI Architecture.

​You can check it out here, but it essentially builds out the architecture outlined above.

It creates the chassis that people can then add to.

Step 2 - Specific abilities

once the framework is ther,e you can start adding in skills/agents.

These are basically specific trainings that have the AI do a specific task with the SOPs you'd use as a real person.

Thigs like...

• Write a sales page
• Optimise an email sequence
• Do customer research

... all of these things have a step-by-step system you'd use.

You create an agent to do it, and within that framework, it can tailor it to your business.

When people have the framework, you can just sell them these skills.

I've done this with the Social AI Agent. It;s essentially a skill that will handle a lot of repurposing tasks for content for you.

And it plugs into the framework.

​You can check that out here.​

Step 2.5 - Customisation

Now, of course these skills might need a little tweaking to the process you use.

Could be using a template you know works, focusing on one specific cohort, using a specific design.

Whatever.

But, out of the box, it'll never be perfect.

This is the hardest bit to sell as it's user-by-user dependent.

This is where something like GM+ comes in. A community environment is a great wayfor you to get peer and expert help to customise the agents and really get them tuned in to exactly what you need.

​You can check out the new GM+ info here. ​

Step 3 - Full integration

This is the hardest step, and really hard to just hand off to someone without close assistance.

At this stage, we're not just using skills, we're improving on them with a few things like...

• Multi agent systems where one agent finishes, and hands work off ot the next agent in the system to continue the next step
• Integrations with services like YT, so you can either pull info or automaticaly publish there
• Full automation of the entire pipeline that uses AI at ONLY the correct areas

Take my own content engine as an example.

I can record a YT video and simply say to it, "run the engine".

This thing will then...

1. Pull the transcript from my YT
2. Identify short form videos to cut out
3. Create 70+ social posts optimised for different platforms, including
   1. Multi-angle posts for LinkedIn, FB, X
   2. Create multiple images Instagram and Linkedin
   3. Turn key messaging and teaching elements into carousels for insta and Linekdin
4. Automatically schedule these for publication
5. Publish them on the platforms on a pre-set cadence
6. Monitor for feedback and analytics to improve the next cycle

All of that done from one prompt.

I was thinking of selling these as standalone systems

But, after a few GM+ people tried them, I realised that these fully integrated systems are a lot to figure out and set up on your own.

These things need more of a white glove/hand holding approach to set up.

And I'm working on something I'm calling "Automated Growth" to help people set up a full ACCER Model AI based system in their business with these highly automated approaches.

That'll be coming soon, so keep an eye out for more info.

In the meantime, I'm wondering if you've either seen the same ifg your'e selling Ai. That most people are not yet comfrotable enough to get results.

Or if you've bought something that failed to get you the right results.

Speak soon,

Pete "3 tiers" Boyle