The 91-Year Old Vibe Coder Paving the Way

Hi there 👋 

For our American pals reading this, Happy 4th!

This week you’ll hear from:

• CEO’s facing hard truths when trying to integrate AI and automation

• A 91-year-old vibe coder using Claude + Replit

• And a client onboarding flow that just…runs itself

Each story shows: The biggest AI wins don’t start with tools, they start with better thinking. 

Let’s dig in.

For automation and AI to be well integrated and provide value, teams have to face the truth about their process and how work gets done.

A hospital tried to use tech to improve patient triage. But, like all teams integrating AI well, they started with mapping out their process. And just as we often see with our own clients, the real insight wasn't about software, it was about the mess beneath it.

This moment says it all:

Here's what happens: When companies try to plug in AI, they're really starting a deep audit of how their business actually works. I wrote about this recently: how the teams getting real ROI aren't starting with custom GPTs or fancy tech. They're first fixing their data and processes, often led by operations folks who understand how information and work actually flows.

And what do they find? Work happening through spreadsheets, silo’d IM threads, and tribal knowledge. Processes that are overly manual and bloated, or there’s almost no process at all.

The research backs this up. When people map these processes, it forces leadership to rethink how their organization actually works. What they thought was structured turns out to be held together by workarounds and institutional knowledge.

The pattern that works: Define how you actually do things → evolve those processes to work with AI and automation → roll it out so teams are aligned and getting the value.

Because AI isn't just a new tool, it's a forcing function. It exposes all the organizational scar tissue. But that's where the real opportunity starts; and where the people who see it most clearly become your biggest champions for change.

📚 Read the full study →

The challenge:
A Professional Services team needed a smoother way to onboard clients, without relying on spreadsheets, manual follow-ups, or everyone trying to play hero by reinventing the wheel each time they bring on a new client.

What we built:
An onboarding system that moves from “Closed Won” to welcome email, without anyone chasing steps.

Here’s how it runs:
→ Airtable classifies the clients engagement type
→ DocuSign sends the right docs from pre-defined templates
→ Front generates a custom welcome pack
→ Airtable tracks onboarding status
→ A human gives one last review
→ Front sends a polished welcome package for the client to have smoother (and 50% faster!) onboarding

Why it matters:
Now, onboarding just…happens.
No one’s creating new versions of the same emails from memory. The system runs whether someone’s at their desk or not. And clients get a better experience.

Does your client or customer onboarding needs a lot of manual hand-holding?
If it still does, we should talk.

John Blackman built a full app with no coding background.

What’s new:
This 91-year-old retiree used free tools (Claude + Replit) to build a complex system for running his church’s community events. Total spend: ~$350.

Why it matters:
You don’t need to “know tech” to use it. You just need a real problem, and the willingness to ask dumb questions until something works.

💡 What stood out:

• He wrote a product brief, not code

• Used AI like a teammate, not a tool

• Built admin portals, volunteer systems, QR-coded reports, from scratch

• Solved a real problem that mattered to real people

What we love:
He calls it “vibe coding.” We call it the future.
This is how more business tools will be built: by people with ideas and a willingness to learn. As a retiree, I’m sure John has more time than most to learn but this is an area we’re focusing on a lot lately as it’s the future of how tools will be built for teams.

📺 Give it a watch, it's inspiring →

The problem:
Everyone says they “use AI.” But just like Excel, there’s a vast difference between basic usage and advanced skill. Some people make checklists in Excel. Others use Pivot Tables and Macros to build models.

Why it matters:
Fluency gaps slow teams down. And without role-specific expectations, companies can’t train, track, or scale what’s working. With AI, it’s the same: leaders need to define what “good” looks like, or risk getting left behind.

⚙️ Build Your AI Fluency Ladder

Start here, a simple framework to shape your team’s AI expectations 👇

📊 Zapier is Setting a Good Example

Zapier has shared their internal AI fluency framework that breaks AI skill levels into four stages, customized for different departments:

• Unacceptable: No curiosity, no use

• Capable: Exploring tools like ChatGPT

• Adoptive: Embedding AI in workflows

• Transformative: Rethinking strategy and delivery with AI

How we’re using this: We’re applying a fluency model across team training and client work, helping teams set clear, role-specific expectations and track how AI moves the needle.

👉 See the original post by Zapier’s CEO on LinkedIn

Box dropped a banger of a blueprint for what it takes to be an AI-first company.

Why it matters:
Most companies are stuck in "AI-as-a-side-project" mode. Box outlines how to shift from tinkering to transformation, and gives tactical advice for each team.

💡 5 insights I enjoyed:

• Rethink, then automate. Don’t just speed up legacy workflows. Ask: What should this look like now that AI exists?

• AI needs new roles. Managers evolve into orchestrators of AI agents, less task assignment, more oversight and outcome design.

• Your data is gold… or garbage. AI thrives on high-quality, well-permissioned, unstructured content. Without it? No insight, just hallucination.

• Culture beats code. Encourage every team (not just IT) to build their own workflows. Some of the best AI ideas will come from HR or Legal, not just DevOps.

• Governance ≠ checkbox. Treat AI governance like product strategy, not compliance theatre.

📊 Where AI Delivers ROI
Tasks that are both repeatable and require judgment? That’s where the magic is.
Think onboarding, contract review, customer requests.

📎 Read Box’s full guide to Being an AI-First Company →

Until next time 👋 

PS.: If you’re planning your fall AI roadmap, we’d love to help you get started.

Let’s chat.