if you want to immediately wow a small business client:
1. set up a gmail read only connector
2. have ai scan their email and flag spend that looks off
did this as a first win for a client and it caught a few spend increases employees didn't tell them about + got a 'wait, how did it know that?'
easy way to show your value in minutes.
if you're a marketing leader being told to hire a marketing engineer, read this.
hire them. but at a medium or large company, the role looks different than what you see on social.
one person can't get into every system, and your specialists won't love being told to change how they run their channel.
so before they start, get the basics in place for everyone:
- skills: finding indiv use cases, building assistants
- access: ai platform, connectors, clear rules
- support: role-based training, time to experiment
- responsibility: part of everyone's job, culture of sharing
the marketing engineer will have greater impact because the rest of the team is capable and included enough to buy-in and help design new systems.
lots of talk about solo marketing engineers. not much on the other 20 people on the marketing team...
made a guide for teams that want ai skills to spread past a few people. link below.
bad culture makes ai seem like it's the bad thing.
if ai feels like it's failing on your team, check these first:
- are people sharing what works or guarding it?
- do they get time to experiment or see it as more work?
- do people feel safe suggesting ideas for ai to help?
- is it safe to say "i tried this and it didn't work"?
how i get way more out of podcasts, youtube videos, and articles with ai:
1. pull the transcript
- podcasts: happyscribe
- youtube: grab it right from video page
- articles: obsidian web clipper chrome extension
2. send it to my agent (it knows my business & goals)
3. ask for the best ideas and ranked actions to take
4. work through or schedule them while they're fresh
the more ai can make for you, the more prioritization matters.
there's always more to do. more content, more channels, more tools and review, and things still take time. you can work all day and it still feels like it's not enough.
i find setting goals in a hierarchy helps a lot (quarterly, monthly, weekly, daily).
tolstoy said it best: "a man on a thousand-mile walk has to forget his goal and say to himself every morning, 'today i'm going to cover twenty-five miles and then rest up and sleep.'"
i'm redesigning the learner experience on my site for teams, managers, and people learning on their own.
if you've taken an online ai course, what made you stop halfway?
how i turn a framework i see on x into a deck made for my business in minutes:
1. keep a claude code setup that knows your business
2. save a deck style you like (mine is an html presentation i've refined over time)
3. drop in the post and say "adapt this to my business and create a deck in my style"
4. tell it what you want out of it:
- framework applied
- view of blind spots
- opportunities, ranked
- what you're already doing well
5. let a background agent run it while you work on something else
super, super cool way to make someone's free framework personal to you.
how i'd teach myself marketing engineering if my company wasn't helping:
1. pick a side project outside work
- a personal brand
- a research project related to your job
- a side project or idea you've had
2. install claude code
- coding tools get you the most from ai
- excellent tool for non-coders too
- place to learn technical skills for bigger workflows
3. create a folder and start building your context
- give info like background/goals ai needs to work well
- structure a dedicated folder where ai work happens
- sessions and capabilities compound as volume grows
4. connect data sources and tools
- have ai recommend tools to help, most have free tiers
- use a scraper like apify or firecrawl for research
- learn about connectors, apis, mcps, clis, safety, limits
5. create a reusable workflow to run
- build a skill for a task that repeats
- tune it until it's good enough to use
- if helpful, set it up to run on its own as a routine
made a free interactive course that walks through this:
want to run claude + obsidian but dont know where to start? this is for you.
walks through set up, structure, & processes to create an ai system that gets better the more you use it.
get the setup top ai users have, interactive & personalized.
free. link below β
i tracked ai marketing job postings for 2 months. here's what they're actually asking for:
- 86% mention governance or responsible ai
- 76% want the hire to build ai workflows
- 76% want them getting other people to use ai
- 67% ask to measure ai's impact
- 65% ask for agents
- 25% name the platform they want agents built on
- 20% mention ai content creation or coding
claude got named about 2x as often as chatgpt.
what ai marketing roles ask for:
- automate marketing workflows
- set ai guardrails and brand voice rules
- drive ai adoption across the team
- write the ai roadmap
- know the martech stack
- connect ai to live data
- build agents yourself
- train the team to run them
- run pilots
- write playbooks and prompt libraries
- pick and prioritize use cases
- evaluate ai tools and vendors
- build dashboards
- measure the impact
- show up for ai search (geo, aeo)
- know claude (chatgpt also mentioned)
if you want to become a marketing engineer, these are the 5 skills to build first:
1. know your company's data
- what platforms you use
- what each platform holds
- what you could gain if they talked to each other
2. how to connect that data to ai
- apis, clis, connectors, and mcps
- most major platforms have them
3. how to handle security and permissions
- how to safely store credentials
- read vs write permissions
- spend caps so you don't get a surprise bill
4. how to let ai write the code
- get comfortable in claude code or codex
- start by building a small tool that helps your work
5. how to maintain what you build
- bugs, old credentials, models changing outputs...
- know how to tune, fix, or retire what's slipped
if you want an ai role, the easiest one to get is probably at your current company.
i've talked to people who got pulled into ai roles, here's what they had in common:
- curious and already systems thinkers
- tinkered with ai in their own work first
- shared what worked with their team
- became known early as the ai person
when a role opened, they were the obvious fit. plus they already knew the business.
since almost nobody has years of experience with ai yet, internal hires have a leg up on outsiders.
if you're deep in ai work, take a few days somewhere with no cell service.
last week i adjusted my setup so i could keep stuff moving while i was off camping... but the joke was on me because there was no service the whole weekend.
i couldn't use any of it, which was definitely for the best.
cool to see how quick the hustle wears off when there's nothing to check or action. good reset.
how to claim your business in chatgpt in <10 mins to have more control over what it says about you:
1. search or add your business
2. add dns record to verify you own your website
3. add info like description, contact info, shipping, etc.
4. go to edit business info
5. verify identity with government id
once you're verified, you can:
- update info chatgpt can use to talk about you
- see how often & what topics you're mentioned in
- add an agent to chat with customers (us only for now)
- upload products with pricing or connect shopify
it's brand new, so expect more to roll out soon.
lots of talk about solo marketing engineers. not much on the other 20 people on the marketing team...
made a guide for teams that want ai skills to spread past a few people. link below.
CRACKED MARKETING 101
Watch AI agents ship 100+ static ads, make UGC videos for $4 that used to cost $500 and upload everything
Here's the ENTIRE system step by step (and ALL of the skills):
if you're setting up ai for a small business, leave ai out of the first conversation:
- ask how their work runs today, step by step
- ask where it slows down or breaks
- write down the friction points in their words
only pick a first ai set up and take them through the below after you have an understanding of their business.
if you're helping a small business get started with ai, do these 5 things with them first:
1. turn off training on their data
2. teach them about usage limits
3. set up a read-only connector
4. walk them through connector permissions
5. set up one reusable thing for their work as a quick win
once they have a better understanding of permissions, they can graduate to read/write tool connections.
the spend check below is a good quick win.