Are you thinking about who your next client is?

The data points somewhere most agencies are ignoring.

Reply below. Curious what you are seeing on the ground.

Most agencies are chasing tech firms. SaaS companies. Growth-stage startups.

Those clients know marketing. They have in-house teams. It is competitive.

Local service businesses need clients and have no marketing knowledge. That is a better pipeline to build.

The roles least exposed to AI are growing.

Local services. Trades. Physical businesses that cannot be digitised.

Electricians, plumbers, local contractors. Fragmented but increasing marketing budgets. Almost no agencies are targeting them well. That gap is the opportunity.

The workers getting squeezed are not in their 30s -50s.

The drop in job starts is specifically the 22-25 age group.

Older workers are mostly untouched for now. This is a generational filter, not a mass displacement event.

Here is what makes this significant.

No mass layoffs. No unemployment spike for existing workers.

AI is not firing people. It is just not hiring new ones.

That is a completely different problem. And almost no one is talking about it.

Entry level role closing soon, but new opportunities are emerging.

Since 2022, job starts for workers aged 22-25 in AI-exposed occupations have dropped 14.3%.

This is from Anthropic's own research. Real Claude usage data. Matched to O*NET occupation data. Published March 2026.

Stop waiting for platforms.

They're selling the dream. You can build the reality.

Build once. Own it forever.

Follow for more scrappy marketing automation + the tools I'm testing each week.

The DIY stack options:

Anonymous traffic ID reveal: 5x5, Retention .com, RB2B
AI: Claude, OpenAI, Google AI Studio
Workflow automation: Zapier, Make, n8n
Email platforms: you name it

Most of these have free tiers or cost less than one bloated SaaS subscription.

The builder advantage:

While everyone waits for platforms to catch up, you can stitch together existing tools and ship personalization TODAY.

Build it now. Own the edge. No permission needed.

n8n (workflow automation, free and self-hosted)
+ Claude API (about $20/month)
+ Behavioral webhooks from your email tool
= Audience-of-one approximation you control.

Here's the gap: Your competitive advantage.

While everyone waits for Mailchimp or HubSpot to add "true AI," you can stitch together existing tools and deliver personalization TODAY.

What platforms actually deliver: AI assistants.

Subject line generators. Basic scoring. That's it.

The gap between possibility and practical tooling is massive. Even email experts still recommend pre-AI tactics like abandoned cart sequences and urgency promotions.

What Mailchimp, Sendgrid, the email platforms promise: True AI email.

That means:
- Predictive personalization for each person
- Contextual timing based on behavior
- Offer optimization per individual
- Scaled to every single subscriber

These platforms can't deliver this yet.

Email is 55 years old (first one: October 1971).

Social media, Discord, WhatsApp haven't killed it.

Gmail's Promotions tab in 2013 was supposed to destroy email marketing. Didn't happen.

Why? Email is owned, measurable, and directly tied to buyer behavior.

Platforms won't deliver.

Email platforms promise AI personalization but won't ship it for 2-3 years. Smart marketers are building it NOW with workflow tools and LLM APIs.

How to build audience of one email automation yourself →

You're not busy.

You're scared.

Revenue work has emotional resistance. Busywork has none.

Track one week. Categorize every task: revenue-generating, foundation-building, or busywork.

Face your ratio. Then protect those revenue hours ruthlessly.

What's your ratio?

The fix that actually worked:

Block the first 4 hours of every day for revenue and foundation work only.

Protect that time like your business depends on it. Because it does.

Email, analytics, social scrolling? Confine to a 1-hour afternoon window.

Try the 2-hour test.

If you could only work 2 hours per day, what would you do?

It's never redesigning a landing page.
It's never checking your dashboard for the 5th time.
It's never scrolling competitor feeds.

The tasks you'd choose are the ones you're avoiding right now.

The data breakdown from 30 days of tracking:

12% revenue-generating:
- Setting outbound campaign
- Fixing leaking funnel step
- Test ad creatives

80% busywork:
- Checking analytics
- Check email inbox 10 times
- "Researching" competitors
- Redesigning pages not blocking funnel

Most marketers don't have a time problem.

They have a courage problem.

Revenue work feels exposed. Outbound emails =potential rejection. Meeting calls =hard conversations. New campaign launch = potential failure.

Busywork feels safe. Nobody judges your analytics refresh count.

A courage problem.

A marketer i talked to tracked 240 hours of work. 80% was busywork. 12% revenue. 8% foundation building.

But the real insight isn't about time management.

Thread: 🧵

Try this today.

Pick one task you hate. Unbundle it completely. Write out every sub-step. Brief it using the 5-part framework.

See what AI produces.

You'll either get what you want first time, or see exactly where your thinking is foggy.

Both outcomes are wins.

I used to blame Claude when outputs failed. Then I realized: garbage in, garbage out.

Unclear briefs = unclear thinking. Now I apply this to every task: competitor research, report generation, email templates, audience analysis.

Outputs got way more consistent.

There's a framework for this.

5-part brief structure (credit: Write with AI):
- Introduction: role + task + output format
- Instructions: exact steps, in order
- Context: rules, tone, examples
- Examples: patterns to recognize
- Delivery: format + constraints

The ones stuck trading hours for output? They can't articulate their own logic.

You can't delegate to AI what you can't explain clearly.

Until you unbundle these, AI fumbles. Clarity forces you to see what you've been automating in your head.

This is the meta-skill for automation-first marketers.

The ones who scale without hiring? They can break any task into clear, unambiguous steps.

AI can't execute competitor research of "your standard". It can execute those seven steps "you defined".

Same with reporting.

"Create weekly report" isn't one task. It's:
- Pull metrics from 3 platforms
- Calculate WoW change
- Flag anomalies
- Format in template
- Add context

The deeper issue: your tasks are bundled.

"Competitor research" sounds like one job. It's actually seven:
1. Define research criteria
2. Scrape competitor pages
3. Extract positioning
4. Compare features
5. Identify gaps
6. Organize in spreadsheet
7. Summarize findings

Good brief:
- Audience: B2B SaaS founders (not agencies)
- Tone: attach brand TOV
- Story arc: Problem → how we fixed it → outcome
- CTA: Download case study
- Length: Under 300 words

One brief breaks. One works.

Example.

Bad brief: "Write a LinkedIn post about our case study."