Operationalization · Personalization · Digital Assistant · Claude · Content · Personal Operating System
S01E17: Your AI Has No Memory. Fix That.Build Your Own Codebase, Part 0
Let’s begin
I have a quick exercise for you. Let’s use the power of imagination to go on a little journey together.
Let’s imagine you are getting ready to hire an assistant who reports directly to you. You’ve been given a pretty healthy budget for the hire, which is exciting to you because you know you can look at a higher caliber candidate pool.
So you start your search. And you land on an absolute 💎 gem of a candidate who has near perfect background. Graduated from a top 40 university. Spent the past 8 years working as a direct competitor to you. They know your ICP. They know your product (having sold against it for the past 8 years). And they also know their old product… which is massive because you know you can use some of their information to update your battle cards.
This is the definition of an “ideal hire.”
The Onboarding Gap
You spend their first week onboarding them. You hand over the "unwritten rules" of your territory. You show them the specific way you structure a follow-up email. You explain why you never mention a certain competitor in Phase 1, but always bring them up in Phase 3.
By the end of the month, this assistant feels like an extension of your brain.
They just “get me”
- you, probably

Now, imagine if every single morning, that assistant walked into the office with total amnesia.
They still have the Ivy League degree. They still have the 8 years of industry experience. But they have absolutely no memory of you, your company, or the 6 months of training you just gave them. Every morning, you have to spend the three hours re-explaining who you are and how you work just to get them up to speed.
You would fire them by Tuesday.
Yet, this is exactly how most reps treat their AI.
They expect "Senior VP level" output from a tool they treat like a temp worker on their first day, every single day.
And before you’re all like “oh, but Jay… I use AI all the time, and my AI has a memory of me so the context is there”… I have another exercise for you. Go to your AI and give it this prompt. This prompt is designed to extract every relevant memory from the AI, and provide an output that you can copy/paste into a new AI to “onboard” it.
Export all of my stored memories and any context you've learned about me from past conversations. Preserve my words verbatim where possible, especially for instructions and preferences.
## Categories (output in this order):
1. **Instructions**: Rules I've explicitly asked you to follow going forward — tone, format, style, "always do X", "never do Y", and corrections to your behavior. Only include rules from stored memories, not from conversations.
2. **Identity**: Name, age, location, education, family, relationships, languages, and personal interests.
3. **Career**: Current and past roles, companies, and general skill areas.
4. **Projects**: Projects I meaningfully built or committed to. Ideally ONE entry per project. Include what it does, current status, and any key decisions. Use the project name or a short descriptor as the first words of the entry.
5. **Preferences**: Opinions, tastes, and working-style preferences that apply broadly.
## Format:
Use section headers for each category. Within each category, list one entry per line, sorted by oldest date first. Format each line as:
[YYYY-MM-DD] - Entry content here.
If no date is known, use [unknown] instead.
## Output:
- Wrap the entire export in a single code block for easy copying.
- After the code block, state whether this is the complete set or if more remain.I’m willing to bet that the information the AI gives you is right… but only to a point. AI does a terrible job of updating it’s own memory. So for example, it might not remember about your recent promotion, or that your kid isn’t 7 for the 3rd year in a row now… And that’s because you have to explicitly tell an AI when something is worth remembering permanently.
The Missing Manual
When you hire that "gem" of a candidate, you don't just hope they figure it out. You give them a Playbook. In the software world, that’s the Codebase. Don’t let the word “codebase” scare you. A codebase is just a Repo. And a repo is just a Folder. And by folder… I mean literally just any folder that you’ve designated on your computer. But when you designate this one… codebase <ahem> as your own personal “source of truth” this becomes your Context Layer.
If you want the AI to act like that high-caliber hire, you have to stop "chatting" and start "onboarding." You need to give it a permanent set of files that tell it:
who it is and what it’s responsibilities are.
but maybe even more most importantly… what skills and tools you provide it with.
and in case you need a refresher…
a skill gives the AI the blueprint/framework for a plan.
a tool gives the AI the ability to execute the plan.
who you are, and what matters most to you.
what you sound like. what you say. what you would never say.
what you do. what you sell. the product or service you deliver. your value proposition.
your icp. what your ideal buyer actually looks like.
where your product or service wins… and where it loses.
who your key competitors are. how you differentiate against them.
how you handle objections. how you sound when you’re doing your best work.
Without this, you don't have an AI assistant. You just have a very smart stranger who keeps forgetting about the promotion you received… last year.
Portability is the Point
The other, massive problem with relying on the AI’s "built-in" memory is that you’re a tenant on their land. If ChatGPT has a bad update, or your company decides to switch everyone over to Claude or Gemini tomorrow, your assistant’s brain stays behind.
A Sales Codebase is portable. It’s a set of simple markdown or text files. When a better AI comes out, you don't rebuild. You just pick up your codebase and move.
That’s what this series is about. We are building your permanent context layer.
The Sales Codebase Roadmap
Over the next several weeks, we are going to build this out together:
Part 2: The Main Branch (The shared team brain: ICP, Product, Battlecards)
Part 3: The Personal Branch (Your layer: Your accounts, your voice, your specific deals)
Part 4: The Prompt Library (Prompts as assets, not one-time use)
Part 5: The Merge (How individual wins become team infrastructure)
By the end, you’ll have a working Sales Codebase that makes every AI interaction faster, more specific, and—most importantly—actually useful.
Rep Action this week
Run that Memory Export Prompt I gave you above. Read what it spits out. Notice the gaps. Notice the outdated info. That mess is exactly why your AI output feels "almost right" but requires 10 minutes of editing.
That gap is what we’re going to fix.
The AI Playbook for Video Teams That Can't Slow Down
Wistia's new AI Video Marketing Trends report shows how marketers are using AI to move faster, improve quality, and extend the life of every video. See how leading teams are driving results without adding more work.

Millions of developers use Wispr Flow to dictate code, write docs, and give coding agents better context. 89% of messages sent with zero edits. 4x faster than typing. Free on Mac, Windows, and iPhone. Try Wispr Flow free.
You just read the motion. Now run it.
The prompts, checklists, and templates that turn this into a 10-minute execution are in the Vault.
Get Vault Access Translation missing: en.app.shared.conjuction.or Sign In
Vault access includes:
- Copy and paste execution prompt packs
- Deal, outbound, and follow-up playbooks
- Operating checklists for every motion
- Members Vault access
