The problem this solves
Competitor sites are the most honest document your rivals publish and the least preserved. Pricing changes ship quietly on a Tuesday, a homepage repositions over a weekend, a feature disappears from the comparison table, and within a month nobody can prove any of it happened. The public web archives crawl too sparsely and too randomly to serve as evidence, so the institutional memory of your market lives in screenshots someone may have taken and probably cannot find.
The cost shows up at decision time. Sales asks when the competitor dropped their entry tier and gets a shrug. The pricing committee debates whether a rival's premium positioning is new or was always there. Positioning work starts from what people remember about competitor messaging, and memory is confidently wrong exactly when it matters. Every one of these conversations wants the same missing artifact: a dated capture of the page as it stood.
This mission builds that artifact on a schedule. Each run captures the pages that matter as both pixel-true screenshots and paginated PDFs, files them in a dated folder structure, and writes a short note on what changed since the previous capture. The archive compounds: each month it runs, the before-and-after record of your market gets one month deeper.
How the mission runs
- Visit each competitor page as a buyer would. The rendered browser loads each tracked page fresh, scripts executed and content fully assembled, so the capture reflects what a visitor sees rather than what the raw HTML contains. Home and pricing pages are the default roster; any URL worth remembering can join it.
- Capture twice: screenshot and PDF. Each page is preserved in two forms: a full-page screenshot that is pixel-true to the rendered experience, and a paginated A4 PDF that prints, attaches, and drops into a deck cleanly. The pair covers both uses an archive gets: visual proof and portable document.
- File into the dated archive. Captures land in Drive under a per-competitor, per-date folder structure, named consistently so the archive stays navigable at any depth: competitor, page, date, format. Nothing depends on anyone remembering to file anything.
- Diff against the previous run. The mission reads each page's content alongside its predecessor from the last run and writes a change note: pricing moved, a tier renamed, a headline swapped, a section removed. Runs with no changes say so explicitly, which keeps the note trustworthy.
- Summarize the interesting changes. When something moved, the run closes with a short brief on what changed and what it might mean, with the before and after captures linked side by side. The evidence and the interpretation travel together.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
A Drive archive that deepens every run: dated screenshot and PDF captures of every tracked competitor page, plus a change log noting what moved between runs and a brief when something meaningful shifted. Twelve months in, you own a private wayback machine for your market with better coverage of your competitors than the public one.
Make it yours
- Add product, careers, and customers pages to the roster: hiring pages telegraph strategy and logo walls quietly document churn.
- Capture the phone rendering as well as desktop for competitors whose buyers skew mobile, since repositioning sometimes ships to one layout before the other.
- Pipe the change brief to a Slack channel so pricing and positioning moves surface the week they ship rather than the quarter someone notices.
Frequently asked questions
Why not just rely on the public web archive?
Public crawls are sparse, unevenly timed, and frequently miss rendered content, cookie-walled layouts, and the exact pages you care about. A scheduled capture of your specific competitor set, in two formats with a change note, is a different class of record: complete for your market, dated on your cadence.
Is capturing competitor pages acceptable practice?
The mission reads publicly served pages the way any visitor's browser does, respecting the same access every buyer has. It never logs in, never probes, and never touches anything nonpublic. Preserving what a company chose to publish is standard competitive practice.
How does the change detection avoid crying wolf?
The diff works on the page's readable content rather than its raw markup, so rotating testimonials and script noise do not register as changes. The note flags substantive movement: prices, tiers, claims, sections. Runs where nothing moved are recorded as exactly that.