Content audit
Reads through the site and flags content that is thin, inconsistent or off-brand, as a list for review.
- InputsA crawl of the site
- I keepThe brief, what counts as a problem, and which fixes get made
- StatusIn use
Agents I build
I build narrow, dedicated agents, one job each, and run them across my day job and my two businesses. I decide what each agent is for, write the rules it works to, and review what it produces. Claude Code does the building.
How I work
-
I start with the people
Before an agent exists I work out who the product is for and what they are stuck on. Research agents gather the market, the competitors and the language real customers use. I read it, argue with it, and decide who the end user is. -
One lane per agent
I break the work into narrow jobs and build a subagent for each one: market research, finding the right end users, drafting, fact checking, reviewing. An agent with one job does it well. An agent with ten does none of them well. -
I write the rules each one works to
Every agent gets a written brief from me: what good looks like, which sources to trust, what it may never do, and when to stop and ask. The design systems, voice guides and motion rules on this site are the same habit. -
I test it, break it and rewrite it
I run the agent, read everything it produces, find where it goes wrong and rewrite the brief. Then I run it again. This is where most of my hours go, and it is why the output sounds like me. -
I review what ships
Agents plan my day, research and draft content, and write the code for Shoptimal, Sayforth and this site. Nothing publishes, ships or changes until I have read it, tested it and said yes.
The agents do the volume. I curate the direction.
The agents
Each row is one job. It lists what the agent reads, what I keep for myself, and whether it is running, in use or still being built.
At Victra
Web Experience and AI Manager at Victra, a Verizon Authorized Retailer. I manage the web experience and lead how the team uses AI, building automations that speed up workflows and improve the end product.
SEO audit
Checks pages against the SEO rules I wrote and reports what is missing or wrong.
- InputsA crawl of the site and my SEO rules
- I keepThe rules, the review, and the call on what changes
- StatusIn use
Outdated promotions audit
Finds promotional content that has expired and is still on the site.
- InputsA crawl of the site
- I keepThe definition of expired, and what comes down
- StatusIn use
Verizon consumer landing pages
Given a reference page, it scrapes the page, reformats it in the consumer brand’s design, verbiage and branding, and stores the imagery locally so the team can add it to the WordPress asset library.
- InputsA reference page
- I keepThe brand rules, the review, and the final call
- StatusIn use
Verizon business landing pages
The same job for the business brand, which has a different look, feel and branding from consumer.
- InputsA reference page
- I keepThe business brand rules, the review, and the final call
- StatusIn use
Apple, Samsung and Google landing pages
One agent for each device maker, each building to that maker’s own design, verbiage and branding.
- InputsA reference page
- I keepEach maker’s rules, the review, and the final call
- StatusIn use
Page-level SEO recommendations
Reviews the search setup of a single page and drafts the changes for the team to make.
- InputsA page and my SEO rules
- I keepThe rules, and the call on what is changed
- StatusIn use
Design UI and UX review
Reviews a page or a design against the design standards and writes up what to change.
- InputsA page or a design, and the design standards
- I keepThe standards, and the final call on every change
- StatusIn use
Analytics and reporting
Reads the project tracker, internal databases and Google Analytics, and assembles the reports the team asks for.
- InputsAsana, internal databases and Google Analytics
- I keepWhat gets measured, how it is read, and what the report says
- StatusIn use
Shoptimal
My Shopify app, live on the Shopify App Store. Agents I built run its content pipeline and my daily planning, and coding agents build the product to my briefs.
Topic research
Researches what to write about next, for the blog and for the social channels.
- InputsPublic search and topic research
- I keepThe brief, and which topics are worth writing
- StatusRunning
Blog writing and publishing
Writes the post and prepares it to publish. Nothing goes live until I have read it.
- InputsA researched topic and my voice guide
- I keepThe voice, the review, and the yes before anything publishes
- StatusRunning
Social post drafting
Drafts Instagram, LinkedIn and Threads posts for me to choose from and approve.
- InputsThe blog and the topic research
- I keepThe voice, the choice of post, and the yes
- StatusRunning
Daily planning, Chief of Staff
Reads the state of the business each morning, picks the day’s three tasks, and hands real work to specialist agents that stop and ask before anything changes.
- InputsThe state of the business
- I keepThe yes before anything changes
- StatusRunning
Product audits
Audit agents review the product and its pages for problems before I do.
- InputsThe built product and its pages
- I keepThe review, and what gets fixed
- StatusIn use
Coding agents
Build the app to my written briefs. I decide what it is and how it behaves, and I review every change before it ships.
- InputsMy briefs
- I keepThe design, the rules, and every review
- StatusIn use
Sayforth
My iPhone speaking app, built and in testing.
Market and competitor research
Researches the market, competing apps, and how onboarding is done elsewhere.
- InputsPublic sources and competing apps
- I keepWhich findings I act on
- StatusIn use
UX and QA audit
Reviews the built app screen by screen, with screenshots and a motion review, and reports what is off.
- InputsScreenshots of the built app, and the motion rules I wrote
- I keepThe rules, and what gets fixed
- StatusIn use
Blog drafting
Drafts blog posts for me to review and rewrite.
- InputsA topic and my brief
- I keepThe voice, the review, and the yes
- StatusIn use
Coding agents
Build the app to numbered implementation contracts and approved mockups.
- InputsNumbered implementation contracts and approved mockups
- I keepThe mockups I approve, the contracts, and every review
- StatusIn use
This site
I designed and directed this site. Research, audit, SEO and coding agents built it, governed by a brand standards reference and a voice guide.
Site research
Gathers the background and examples I ask for before I design a section.
- InputsThe questions I set
- I keepWhat I take from it, and what I leave
- StatusIn use
Site audit
Reads the built pages against the brand standards and the voice guide, and lists what breaks them.
- InputsThe pages, the brand standards and the voice guide
- I keepThe standards, and what gets fixed
- StatusIn use
SEO audit
Checks every page against a written SEO checklist.
- InputsThe pages and the checklist
- I keepThe checklist, and the call on each change
- StatusIn use
Coding agents
Build the pages to my design direction.
- InputsMy design direction, the brand standards and the voice guide
- I keepThe design, the rules, and every review
- StatusIn use
Personal
A menu-bar app on my Mac that I use every day. It is for me alone.
Menu-bar app
The glass holds the day’s three tasks and shows how much of the day is left. The bar answers quick questions from the current numbers and sends real work to a team of specialist agents.
- InputsThe state of the business
- I keepThe yes before anything changes
- StatusRunning
Voice dictation
I hold the Fn key and talk, and the words land in whatever app is in front.
- InputsMy voice
- I keepWhat I say, and when I send it
- StatusRunning
One agent, end to end
Take a reference page and a brand. The agent rebuilds the page to that brand’s standards: consumer and business look and read differently, and so does each device maker. It collects the imagery for the team and hands me a page to review. It does not publish anything.
- A reference page
- The brand to build for
- That brand’s written rules
- Reads the page
- Rebuilds it to the brand
- Collects the imagery
- A page for me to review
- Imagery stored locally
- Questions it stopped on
-
It reads the reference page
I give it a page and name the brand to build for: Verizon consumer, Verizon business, or the device maker. -
It pulls the copy and the imagery
The words and the pictures come off the reference page and are sorted so they can be rebuilt. -
It rebuilds the page to the brand’s standards
The brand’s design, its verbiage and its branding, not the reference page’s. Each brand has its own agent and its own rules. -
It stores the imagery locally
The team can then add it to the WordPress asset library themselves. -
It hands me a page to review
I read it, correct it, and decide whether it goes any further.
What I decided
- The brand rules. Each brand has its own written rules for look, feel, wording and branding. The agent works to the rules of the brand I named and no other, and never mixes one brand’s styling into another’s.
- When it stops and asks. When the reference page uses a layout or a component the brand rules do not cover, when copy or imagery is missing or unclear, and when two rules disagree.
- What it may never do. Publish a page, change anything on the live site, upload to the asset library, or write copy, offers or product details that are not on the reference page.