Lab Notes
Working notes on marketing operations, AI and how teams actually get work done. Observations, experiments and the occasional rant, written by Katie, founder of SpotSeven Intelligence.
-
No. 16
The Screen Is Blank and Pink: Why Insurance Ads Are Losing Their Humanity
As insurers automate agents out of the business, their ads face a creative problem: how do you sell a cheap digital transaction when your brand was built on human trust? Allstate, Progressive, State Farm and Farmers, through a production lens.
-
No. 15
How does a human compete in an increasingly AI and automated world?
Where people still have the edge over AI, and five ways job seekers and employees alike can position themselves to stand out.
-
No. 14Experiment
How to Create an Automated Agent in Claude for a Repeatable Task
I turned my twice-a-week hunt for new opportunities into a Claude agent that runs on a schedule and hands me a short digest. Seven steps, four lab findings, and six other ways to use the same framework.
-
No. 13
How I evaluate a project management platform (and why it matters less than you think)
Eight steps for choosing a project management platform without getting dazzled by the demo, plus my free scorecard: 71 criteria, weighted scoring, no email required.
-
No. 12Experiment
Building a Website with Claude
I built the SpotSeven Intelligence website with Claude in a few days, start to finish, without writing a line of code myself. Seven steps, three lab findings, and what I'd tell anyone thinking about trying it.
-
No. 11Flattening, part 3 of 3
The genie is out of the bottle. What to do after you flatten.
You already flattened, and reversing it isn't realistic. Seven places to start if decisions are stalling and your best people are quietly overloaded.
-
No. 10Flattening, part 2 of 3
"AI can absorb that layer" is not a flattening strategy
AI absorbing a management layer isn't automatic. It takes planning, investment and time. Six things to do before you touch the org chart.
-
No. 09Flattening, part 1 of 3
Can fewer managers really get the same amount of work done?
Flattening promises fewer layers and faster decisions. Months later, things fall through the cracks more, not less. Six signs your org is feeling it.
-
No. 08
There are two kinds of operations people
Every project management platform promises to be the single source of truth. Every one slowly becomes the place truth goes to die. What the evaluations never measure.
-
No. 07
Broad enough to collaborate, deep enough to deliver
A friend with decades in the industry wrapped a project and said "I learned a lot." That's not a confession, it's the actual skill. On the T-shaped professional, and what to look for when you're hiring.
-
No. 06
Every AI integration is a bet, not a foundation
OpenAI shut down Sora six months after launch. The real AI risk isn't losing your job, it's losing the tool you built your workflow around.
-
No. 05
That's not a pricing strategy, that's a hostage negotiation
Spotify quietly downgraded a plan we were already paying for, then offered to sell it back. A lesson in how not to communicate a pricing change.
-
No. 04
The most interesting part of an 81,000-person study was the interviewer
Anthropic used Claude to interview 81,000 people in 70 languages. The study is interesting, but the method is the real use case.
-
No. 03
Nano what? Why ops people should care about Google's Nano Banana
Google's AI image tool lives inside Google Ads and can finally render legible text. The bigger story is the workflow: AI on both ends, with human judgment in the middle.
-
No. 02
For International Women's Day: the leader who taught me how to lead
One name keeps coming back to me on International Women's Day. What Jill taught me about leading, and a note to every young woman finding her footing.
-
No. 01
Your team isn't out of ideas. They're out of headspace.
The people who can see exactly what needs fixing are usually too buried to fix it. One question that surfaces what they already know, and why no AI tool can do it for you.
Nothing here yet. See all entries