Skip to content
Curious Context Labs

case study · Marketing · Agent

Knowing about competitor price changes before a customer tells you

How a weekly digest of public competitor moves, with a link for every claim, can replace rumours and stale slide decks.

business
Manufacturing
build
Agent
timeline
2–3 weeks
read
3 min

01 the business

Specialist packaging supplier

55 staff, selling online and through distributors

Everyone agrees competitors should be watched, but nobody owns it. Price changes come to light when a customer mentions them on a sales call. The last competitor deep-dive is a year-old slide deck.

02 before

The problem

Everyone agrees they should keep an eye on competitors, but nobody owns it. Price changes and new products get noticed when a customer mentions them on a sales call. The occasional deep-dive goes into a slide deck and is stale a month later.

  • Competitor moves are heard second-hand, often weeks late.
  • Sales relies on outdated talking points.
  • Nobody knows which search terms competitors are winning.

03 the build

What the AI does

  1. Step 1: Checks the competitor websites, pricing pages and product listings you name for changes since the last visit.
  2. Step 2: Follows their blogs, press releases, job ads and public ad libraries for launches and new campaigns.
  3. Step 3: Reads new public reviews of competitors and groups the praise and complaints by theme.
  4. Step 4: Uses Semrush or Ahrefs data to show which search terms they are gaining or losing on.
  5. Step 5: Writes a weekly digest with links to every source and posts it to Slack or a Notion page.

paused: waiting for a person to approve

How it was put together

  • Listed eight competitors with their pricing pages, product listings, blogs and job ads.
  • Checked each source weekly for changes and stored a dated copy of what changed.
  • Grouped new public reviews of competitors by theme, such as delivery times or quality.
  • Wrote a weekly digest with a link to every source, posted to Slack and Notion.

04 human in the loop

Where a person stays in charge

The agent only reads public sources. A named owner in marketing reviews the digest before it is shared with sales or leadership, and decides whether any finding needs a response.

  • Only public sources are read, and each site's terms and robots rules are respected.
  • Every claim in the digest links to its source.
  • A named marketing owner reviews the digest before it goes to sales or leadership.

05 rollout

From kickoff to live in 2–3 weeks

  1. Week 1Agree competitors, sources and what counts as a change worth reporting.
  2. Week 2Run the first digests for the marketing owner only.
  3. Week 3Share weekly with sales and leadership.

06 after

What changes

The team hears about competitor moves within days, from a single trusted source. Sales gets current talking points, and pricing or product decisions start from facts rather than rumour.

worked example · your numbers will differ

Assume eight competitors with five sources each, checked by hand at 6 minutes per source every week.

8 × 5 × 6 minutes = 240 minutes, or 4 hours a week of checking that usually does not happen. The digest takes about 15 minutes to read.

What to measure

  • Days between a competitor change and the team hearing about it.
  • Digest items that led to an action, such as a price review.
  • Sales feedback on talking points.
  • Search terms gained or lost against named competitors.

07 evidence

What the research says

  1. About 75% of generative AI's potential value sits in four areas: customer operations, marketing and sales, software engineering, and R&D.

    McKinsey & Company, The economic potential of generative AI: the next productivity frontier (opens in a new tab) · 2023

  2. In a field experiment with 758 BCG consultants, those using GPT-4 on tasks within its abilities finished 12.2% more tasks, 25.1% faster and at over 40% higher quality, but did worse on a task outside its abilities.

    Harvard Business School / BCG, Navigating the Jagged Technological Frontier (Dell'Acqua et al.) (opens in a new tab) · 2023

08 lessons

Mistakes to avoid

  • Tracking too many competitors. Start with the five who show up in deals.
  • Reporting noise. A typo fix on a pricing page is not news.
  • Using it to copy. Use it to decide, not to imitate.

09 questions

Common questions

Is this legal?
Reading public web pages is generally fine, but respect each site's terms. The agent does not log in or collect personal data.
Can it watch marketplaces like Amazon?
It can track public product listings and prices on marketplaces you name.

keep reading

All case studies ->

Tell us how your business runs. We'll tell you where AI fits.

A free 30-minute call. You talk, we ask questions, and you leave with a few concrete ideas — or an honest answer that AI isn't the right fix yet.

Working with teams around the world.