Lunch & Learn  ·  Part 2
AI in Action —
The AURA Way

From understanding AI to using it — consistently, inside your business.
With a live demo built inside Claude Cowork.

Part 2 of the AI for Local Businesses series. Part 1 covered the fundamentals — what AI is, how LLMs were built, and why it matters now. Today we go further: how to think in systems, what Claude Cowork is, and a live demo of LaymanCRM — a real skill built to solve a real problem local business owners face every week.

Today's roadmap
The AURA Way.

Everything we cover today maps to one of these four steps. By the end, you will have seen every one in action.

A 01
Activate
Connect AI to your real business data.
Today → We connect Claude Cowork to your contacts and documents.
U 02
Use
Ask questions every day. See what it can do.
Today → We run the LaymanCRM demo. You see a real skill in action.
R 03
Rely
Find what saves time. Make it repeatable.
Today → We map the demo to your business. You find your card pile.
A 04
Automate
Set it running. Let AI handle the busywork.
Today → You design your first skill using three simple questions.

AURA is the adoption framework at the heart of The AI Guy's approach to local businesses. It is not a product — it is a progression. Today every section of this session lives inside one of these four steps, so the framework becomes real through experience rather than description.

Quick recap from Part 1
Three things
to hold onto.
1
AI is pattern recognition at scale.
It does not think. It finds and applies patterns — across text, images, audio, and video — with a precision no human can match at volume.
2
You can now talk to computers in plain language.
No code. No technical knowledge. Describe what you need, and it responds. This is what opened AI to every business owner.
3
The challenge was never the tool — it was having a system.
Most businesses use AI like a search engine. Today changes that.

These three ideas from Part 1 form the foundation for everything in Part 2. The third point is the most important for today — it is the bridge from understanding AI to using it in a way that creates real, consistent business value.

The mental model shift
Prompts don't
compound.
Systems do.
What most businesses do
Ask a question. Get an answer. Feel impressed. Start fresh tomorrow with a different question.

Nothing connects. Nothing accumulates. AI stays a novelty.

The difference between using AI as a novelty and using it as a business asset is systematic thinking. A system is simply a repeatable workflow: defined inputs, a reliable process, and a predictable output. Once that exists, AI stops being something you try and starts being something you rely on.

Your operating environment
Claude Cowork is not
a chatbot you restart.
It already knows your business.
🧠
Persistent Context
Every conversation starts from a shared knowledge base — your files, your data, your business context. You never have to re-explain yourself.
📋
Instructions
You configure how it works — your tone, your priorities, your way of doing things. It behaves consistently because it has been told how to show up.
Skills
Saved, reusable workflows you can run on demand. The building block of every system you will build. More on this in a moment.

"Think of it as a team member who already has your onboarding done — and gets smarter every time you add to the project."

Claude Cowork (referred to as a Claude Project in Anthropic's tooling) is a persistent, configurable workspace. Unlike a standard chat session, a Cowork project retains context between conversations, follows configured instructions, and can run saved workflows called skills. This makes it suitable for systematic, repeatable business use rather than one-off queries.

The building block
A skill is a saved,
reusable workflow.
Input
📥
What goes in
A photo, email, voice memo, document, form — anything your business runs on.
Skill
What it does
Instructions you wrote in plain language. Extract, enrich, summarize, draft, classify.
Output
📤
What comes out
A CRM entry, email draft, task list, file update — a consistent, useful result.
What makes it a skill
You write it once. You run it every time the situation arises. Same input — same high-quality output. No re-explaining. No inconsistency.
What it is not
Not code. Not automation software. Not a technical project. Just clear instructions — written in plain language — telling Claude what to do and how.

A skill is the fundamental unit of an AI system inside Claude Cowork. It is a saved set of instructions that defines what to do when given a specific type of input. Because it is reusable and consistent, it can become a reliable part of how the business operates — rather than something that needs to be set up from scratch each time.

The connection
AURA lives
inside Cowork.
A Activate Connect your data to a Cowork project. Your files, contacts, documents — whatever your business runs on.
U Use Run skills against your data every day. Ask real questions about your real business. Build trust in what it can do.
R Rely Your skills become standard operating procedure. The ones that save time stop being optional and start being expected.
A Automate Skills trigger on new input without you initiating. The system runs. You focus on the work only you can do.

Cowork is not just a tool. It is where AURA lives — the environment in which each step of the framework becomes something you can actually do.

Claude Cowork provides the technical foundation for the AURA framework. Each step of AURA corresponds to a specific way of working inside a Cowork project. Activate means connecting data sources. Use means running skills daily. Rely means embedding skills in standard workflow. Automate means configuring skills to trigger without manual initiation.

The problem we are solving
You just left
a great event.
Now what?

You had real conversations. Right here at 25N. Five business cards in your pocket. Three LinkedIn connections pending. A few names you remember vaguely and one person you genuinely want to follow up with.

In 48 hours, the context is gone. Who they were, what you talked about, why it mattered.

What usually happens
Cards sit in a drawer. LinkedIn connections go uncontacted. You meant to follow up. You just never did. The opportunity quietly expired.
What we are about to build
A system that turns a card and a LinkedIn screenshot into a complete contact record — and drafts a personalized follow-up email in your voice. In seconds.

The networking follow-up problem is nearly universal for local business owners. The challenge is not motivation — it is friction. Without a system, even well-intentioned follow-ups get delayed until they feel irrelevant. LaymanCRM eliminates the friction by making the process instant, automatic, and personal.

LaymanCRM  ·  The workflow
Five steps.
One system.
📷
Upload card
Photo of the business card from the event
🖼️
Add LinkedIn
Screenshot of their LinkedIn profile
Skill enriches
Cowork skill combines both inputs into a full contact record
💾
Saves to CRM
Enriched contact written to your local JSON CRM
✉️
Drafts outreach
Brand voice skill writes a personalized follow-up email
💡

Two inputs. Thirty seconds. A complete contact record — and a personalized email ready to send. This runs entirely inside Claude Cowork.

LaymanCRM is a Claude Cowork project built to solve the post-networking follow-up problem. It uses two skills: an enrichment skill that processes a business card photo and LinkedIn screenshot to populate a structured JSON contact record, and a brand voice skill that drafts a personalized outreach email based on the enriched data. No external CRM. No monthly fee. Your data stays local.

LaymanCRM  ·  What the skill processes
Two inputs.
One enriched contact.
What comes out
NameSarah Chen
CompanyChen Financial Group
RolePrincipal
Emailsarah@chenfinancial.com
LinkedInlinkedin.com/in/sarahchen
Met at25N Networking Event, May 2026
ContextDiscussed AI for client intake; exploring tools to reduce onboarding time
Talking pointsEstate planning practice; team of 4; manually re-entering intake data across 3 systems

The enrichment skill inside LaymanCRM reads both inputs simultaneously — card photo and LinkedIn screenshot — and produces a structured JSON contact entry. It extracts factual fields from the card, supplements them with context from LinkedIn, and adds the event context. The output is richer than anything you would have time to type manually, and it is immediately ready for outreach.

LaymanCRM  ·  The outreach email
It sounds
like you.
🎙️

The brand voice skill knows how Raj writes — his rhythm, his directness, his warmth. It produced this from the enriched CRM entry in seconds. It is not a template. It is a trained voice that sounds like a real person, because it learned from one.

The brand voice skill is a separate Cowork skill configured with writing samples and tone guidelines. It reads the enriched contact record, pulls the relevant context, and produces a personalized email that reflects the actual conversation from the event. The result feels handwritten — because the pattern it follows was built from real writing, not generic prompts.

Naming what you just saw
That was
a system.
The mental model
Input Card photo + LinkedIn screenshot
Process Cowork enrichment + brand voice skills
Output Enriched CRM entry + personalized email
Mapped to AURA
A Activate Your contacts are now data
U Use Run it after every event
R Rely Post-event routine: 10 min, not zero
A Automate Runs whenever you add new inputs

The LaymanCRM demo illustrates every element of the AURA framework in a single workflow. It is not a complex system — it is two skills, two inputs, and two outputs. But because it is repeatable and reliable, it qualifies as a genuine system. That is what creates business value: not sophistication, but consistency.

Four more examples
The same pattern.
Different business.
🎙️
After a client call
In: Voice memo
Out: Meeting summary + action items in your task list
📋
New client intake
In: Photo of intake form
Out: Enriched client record, filed and ready
📧
New inquiry email
In: Incoming email
Out: Draft reply in your brand voice, ready to review
📄
Scanned invoice
In: Photo of invoice or receipt
Out: Categorized expense entry in your log

Every one of these follows the same structure. Same input. Same process. Consistent output. A real estate agent, a therapist, a caterer, an accountant — the pattern is the same. Only the subject matter changes.

These four examples illustrate that the skill pattern from LaymanCRM is universal. Any business that receives recurring inputs and produces recurring outputs can build a skill for it inside Claude Cowork. The goal is to identify one high-friction, high-frequency workflow and start there.

Your turn
What's your
card pile?

The thing that piles up in your business and rarely gets done — not because you don't want to, but because there is no system. That is your starting point.

📇 Business cards from events you haven't followed up on
📥 Client inquiries that didn't get a same-day response
🗒️ Meeting notes that never became action items
🧾 Invoices or receipts that still need to be logged
✉️ Follow-up emails you keep meaning to write
💬

What is yours? Take 30 seconds. You already know the answer.

This is the highest-value moment of the session. Every business owner has at least one workflow that accumulates and goes undone — a stack of friction that quietly costs them time and opportunity. The goal is for each person in the room to name theirs, because naming it is the first step toward building a skill for it.

Design your first skill
Three questions.
That is all it takes.
1
What is the input?
A photo, an email, a voice memo, a form, a document, a screenshot — whatever currently arrives and sits unprocessed.
2
What should the skill do with it?
Extract information, enrich a record, summarize content, draft a reply, classify and route — or some combination. Describe it in plain language.
3
Where does the output go?
A CRM entry, an email draft, a task list, a file, a notification. Where does this result need to land to be useful to you?

If you can answer these three questions, you can build the skill inside Claude Cowork — in plain language, today. No code. No technical background required.

These three questions are the design framework for any skill built inside Claude Cowork. They force clarity on what the workflow does and why it matters, before any configuration begins. Most business owners can answer all three in under two minutes once they have identified their card pile.

Your next move
Start one thing
this week.
1
Set up a Claude Cowork project
Free or low-cost to start. Create a project. Give it a name that matches your business or workflow. That is your AURA home base.
2
Go back to your card pile answer
The workflow you identified a few slides ago. Write down the three questions for it — input, process, output. Ten minutes of thinking, maximum.
3
Build the skill in plain language
Describe what the skill should do, using your three answers. Paste it into your Cowork project as a skill instruction. That is the whole build process.
4
Test it with one real input this week
One card. One email. One voice memo. See what comes out. Adjust the instructions if needed. That first run is worth more than any planning.

The most common mistake after a session like this is trying to build everything at once. One skill, one real input, one result this week is the target. The goal is not to automate the whole business — it is to create one moment where a system works and you trust it. Everything else follows from there.

Let us connect
Thank you.

If you want help setting up your first Cowork project — or thinking through which workflow to start with — I am happy to continue the conversation.

No pitch. No pressure. Just a practical conversation.

The local accountant who knew your name. The small firm that actually had time for you.

These businesses deserve the same tools the big companies are building for themselves.

That is why I am here.

The AI Guy  ·  Raj Sehgal
Chicagoland  ·  Proud member, Schaumburg Business Association

If this session raised questions about your own business and where a Cowork skill could help, reach out. The best next step is usually a short conversation — not a sales process. Part 3 of this series will go deeper into automating full workflows.