AI tooling · Content strategy
The AI content auditor I built for my team
One of four tools I built so my strategists spend their hours on judgment, not busywork. Each halved our time to deliver for its task focus.
My thinking
My philosophy is that tools should make people more efficient. You could dig a hole in the dirt with your hands but a shovel is way better. In areas where there’s nuance, expertise accumulated over years of experience, or just intuitive understanding still required, AI tools should be an enhancement rather than a replacement. My goal with these tools was to automate the time-consuming parts of the tasks, which is where AI is genuinely good, so that the person using the tool could focus their energy on things like insights, recommendations, and next steps.
What it does
My goal here was to improve the content audit process. You point the tool at a site and give it a framework: the funnel stages that become the columns, the content categories that become the rows, the jobs content can do, and a scoring rubric in plain language. It crawls the whole site, discovering pages through the sitemap and robots file, then an LLM reads each page and places it: which category, which funnel stage, what it’s doing, and how strong it is against your rubric.
Click any screenshot to read it full size.
You define the framework in plain language. The funnel stages become matrix columns; the categories become rows.
Point it at any site. Here it’s auditing my own.
It crawls in real time, sitemap discovery, page fetch, then classification.
Each section is classified by an LLM rather than keyword-matched, which is what makes the scoring judgment-aware instead of mechanical.
Deliverable
The result is a live matrix. Every cell is a place content should exist, and it tells you whether what’s there is strong, thin, stale, or missing entirely. The gaps light up. Instead of a strategist spending days building this by hand, they open it already built, audited, and organized based on their framework. Their task begins at exploring a curated dataset so they can apply their own expertise and provide real insight.
| Awareness | Consideration | Decision | |
|---|---|---|---|
| Product / Services | GapNothing here yet | StrongDesign & strategy | GapNothing here yet |
| About / Leadership | StrongWho we are | ThinHow we work | StrongContact |
| Case studies | ThinTeasers, no proof | StrongFull write-ups | StrongOutcomes |
| Blog / Editorial | StrongRegular posts | StaleOld references | GapNothing here yet |
| Legal | StrongPrivacy, terms | GapNothing here yet | GapNothing here yet |
Illustration of the output. Every gap and weak spot is visible at a glance, filterable, and exportable.
Running it live
Here it is pointed at this very site. Part of me hates seeing how threadbare my own site is. The perfectionist in me wants to see green across the board. But parisi-hq.com is my small, personal site. I already know where my content gaps and weaknesses are. I’m glad my tool agrees with me independently. That’s a good sign that I can trust its output.
The results view: filter by category or status, inspect any cell, export.
The outcome
This is one of four tools I’ve built for my content team, alongside a brand-health analyzer, a content-strategy wizard, and a content-gap dashboard. Each one takes a slow, manual job off the team’s plate. Each one has more than halved our time to deliver, and, more to the point, they moved my strategists’ hours out of collection and into judgment, which was my main goal anyway.
My role
I designed and built these myself, directing and testing the code rather than writing every line by hand, the same AI-assisted way I build everything now. The thinking behind them, what to measure, how to score it, what a strategist actually needs on a Monday morning, leverages my own years of knowledge as a content strategist, as well as the insight, input, and QA testing from the rest of my content strategy team, who all helped hone this tool into something viable rather than a nebulous positivity echo chamber.