how to take over a "boring" local market fast with AI generated meta ad campaigns
whether you're a consultant/agency or an operator...
1) positioning
2) structuring your offer
3) building campaigns (without touching ads manager)
plus, how I've solved my design frustrations with coding agents
AI turned building into the easy part. so the real trap now is false productivity: endless dopamine, and a project that never ships.
you get a win every few prompts, so you keep going. two weeks later you're still deep in it and can't even say what problem you started with.
the way out is setting your gates before you start:
1. write a phased PRD where "done" is an actual gate. you don't start the next thing until you've hit it.
2. ship the MVP and let the market tell you if it's real, instead of building out every case first.
3. pick a launch date and ship whatever you've got when it comes.
building will always feel like progress. shipping is the only thing that proves it.
clip's from one of our live sessions in The Boring Marketing Community π
everyone in AI is fighting over the same 10,000 customers.
meanwhile there's a plumber in your city doing $2M a year who's never heard the word "agent."
that's the whole opportunity. the AI niche is crowded, disloyal, and churns the day a new tool drops. the boring trades one town over have demand that shows up every single day...and competitors who can't spell SEO.
you don't have to invent a market. you point the skills you already have at one that's been sitting there the whole time.
two ways to spend your AI skills π
this came from a hot seat inside The Boring Marketing Community, where we go deep on using AI to grow real local businesses. join the room π
https://t.co/sfHR3C5whH
5 mechanics = one new $25k/mo market.
my diesel-repair business never had a demand problem. i could generate calls all day. what i couldn't do was find enough mechanics to take them.
so i ran ads to recruit the workers, not the customers.
get 5 in a market and i flip it on. avg job ~$1k, ~60% margin after labor.
then i run the same playbook into the next city.
i could get calls all day. finding mechanics to take them is the actual business.
I used Astra + Blender to build a launch film for a luxury brand that doesnβt exist.
One watch. An entire world built around it.
Imagine showing up to your next client pitch with something like this.
minimum viable campaign might become one of the most useful frameworks for small marketing teams using AI.
mvc = build the smallest real campaign that can tell you whether the bigger campaign is worth building.
AI has made campaign execution incredibly cheap.
you can go from an idea to a landing page, 10 creatives, 5 emails, nurture sequence, sales collateral and automation in hours.
but, faster execution doesn't tell you whether the underlying bet is any good.
that's where mvc comes in.
instead of building everything, start by asking:
what's the minimum we need to put in front of the market to get a trustworthy signal?
the framework is simple:
1. define the bet: what do we believe will happen?
2. define the audience: who specifically needs to respond for this to matter?
3. define the signal: what behavior would tell us there's real potential?
4. build the minimum: one audience. one message. one offer. one channel. only the assets necessary to make the test credible.
5. set the threshold: decide what strong, mixed and weak evidence look like before seeing the results.
6. decide: strong signal β expand
mixed signal β change + retest
weak signal β kill
AI can help throughout the cycle:
research the hypothesis, find existing evidence, build the test, analyze response quality and capture what you learned.
then when the market gives you a reason to believe, use all that AI execution leverage to build the bigger campaign.
this is the shortcut to being able to make more bets, learn cheaply and put your resources behind the ones that earn the right to scale.
if you want AI to help a local business get found, capture demand, convert leads and find growth opportunities, start with
building a local business kit.
the idea is to make sure the important parts of the business exist digitally and are accessible:
- be found β Google Business Profile, website, service + location pages
- build trust β reviews, testimonials, before/after proof, FAQs
- capture demand β calls, forms, bookings, CRM, lead sources
- know the customer β call transcripts, estimates, lost reasons, customer history
- learn + grow β revenue, margins, conversions, channel + location performance
once these pieces are connected, AI has actual business context to work with.
now you can build things like:
a. local opportunity radar to find where to expand
b. demand capture engine to turn interest into bookings
c. revenue leak scanner to find money being left behind
d. customer reactivation to bring old customers back
e. follow-up engine to recover missed leads
f. customer intelligence to spot patterns across calls, jobs and revenue
and every new lead, call, job and customer gives the system more to learn from.
before thinking about which AI tool to add, make sure AI has enough of the business to work with.
six months from now, will your AI know why you killed that campaign, changed the offer, or moved the budget?
your next marketing decision should start with everything you've already learned. AI can make that possible.
every meaningful marketing decision leaves behind useful context:
a. what triggered it
b. what evidence you had.
c. what you decided.
d. why you chose it.
e. what you expected to happen.
f. what actually happened.
the problem is, most of that reasoning disappears into Slack threads, meetings, dashboards, or someone's head.
start capturing it in a simple DECISIONS.md.
trigger β evidence β decision β why β expected result β actual result β status β next step
then give AI access to it alongside your marketing context.
the next time you're deciding whether to change an offer, retry an audience, move budget, or run an experiment, it can look at the decisions that came before it.
every decision makes the next one better informed.