Skip to main content
When a competitor already ranks at scale, their sitemap is a list of the keyword patterns they’ve bet on. Ask Pi to work out who’s winning at programmatic SEO in your space, crawl their sitemaps, and pull every pattern they target into grouped files, then reshape those groups into your own pillar coverage. Pi does the research and crawling with the keyword-exploration skill and lands the output against your landing-page pillar. In the walkthrough below, a seed set of roughly 100 keywords becomes about 200 page groups covering close to 1,500 keywords across two passes.
Use this once you have your base. It assumes you’ve already done company research, a Product Bible, and a first batch of BOFU keywords, and that you have competitors executing programmatic SEO at scale. If you’re still finding your first keywords, start at How to find keyword opportunities.

​
The idea

Programmatic SEO means one page template stamped across hundreds of near-identical queries: an AI video tool builds a text-to-video page, a blog-to-video page, a youtube-shorts-maker page, and so on. A competitor who’s been at it for a while has already tested which of those templates earn traffic. You don’t have to guess the map. You can read theirs. The method has five moves, and Pi runs each one for you:
  1. Find who’s best at programmatic SEO in your category, and roughly how much traffic each pattern pulls, so you copy the winners rather than the noise.
  2. Read their sitemaps to see the templated page families they run (often a nested sitemap groups them for you, like a /make/ or /platform/ section).
  3. Brainstorm the commercial angles first, the use cases where people actually spend money, so the expansion leans toward buyers rather than high-volume filler.
  4. Have Pi crawl each competitor’s sitemap and dump the keywords behind their dominant strategy, one file per competitor, worked in parallel with subagents.
  5. Regroup everything into your own pillar categories, dump all observed keywords one file per group, and flag what’s unique to you. The big set then flows into evaluation.

​
Steps

Each step below has the real prompt to send Pi, followed by what comes back. Copy the prompt, swap the [bracketed] parts for your own, and follow along.

​
1. Ground the account

Confirm the Product Bible, base keywords, and (ideally) a horizontal landing-page pillar are already in place, so Pi expands against what you actually sell. See the Product Bible and Design a landing-page pillar. Starting from scratch, this one prompt builds the base and passes it forward: 
i'm working on [my product], help me:
1. do a website research
2. build a product bible
3. update my org info
4. do deeper research on what BOFU keywords i need, see the 2d keyword labeling framework
5. expand on the keywords and think about programmatic SEO templates that would be useful
6. (in parallel to 5) using the product bible, help me plan a horizontal landing page pillar (create it)
do in sequence, passing data from prev to the next, use subagents.
Pi runs each step in order and hands the base keywords and pillar into the next stage.

​
2. Get a competitor report and the commercial angles

In one pass, ask Pi to brainstorm the commercial use cases (the ones where someone makes money) and research who ranks best, how mature their SEO is, and how they group their keywords. The report tells you whose sitemap is worth mining; the use cases keep the expansion pointed at buyers. 
i want you to brainstorm on more use cases that are more commercial,
ie people who make money from what we sell (ugc ads is one example).
find more use cases in that "[product] for x" shape. research some
other competitors and see how they group and structure their pages,
then give me a full competitor report on who does SEO best and how
they group their keywords.
Pi ranks the competitors, maps their strategies, and groups the commercial angles.
A competitor SEO report table listing each competitor with their organic strategy, top traffic page and monthly visits, and SEO maturity.
A table of commercial video use cases in two tiers, each with a head keyword, monthly volume, CPC signal, and a note on why it is commercial.
Finding the commercial angle (mostly a B2C thing). For B2C products, the useful question is who treats it as work: the people who make money from what you sell are the likeliest to pay for it, so lead with their use cases. Focus there first before the casual, hobbyist queries. B2B is more straightforward, since almost every buyer already has a budget and a job to get done. Either way, expand the keywords tied to helping someone make money or save money first.

​
3. Mine each competitor’s sitemap, one file each

Point Pi at each competitor’s sitemap and have it pull the keywords behind their dominant strategy, one file per competitor, worked in parallel with subagents. 
now go to the sitemap for each of these and pull out their keywords for their dominant
strategy, and tell me what keywords i should target next. use subagents, one file each.
Pi writes a file per competitor with their target keywords, ordered by their dominant strategy, and flags the biggest structural gaps. If a long run stalls partway, have it resume and persist to files so nothing is lost: 
previously you were researching various competitors. continue the task and write it out to
files as you go, so we don't lose the data.
An analysis showing the biggest structural gap, a prioritised list of keyword clusters to build, and angles no competitor has covered.

​
4. Regroup into your pillar and dump every keyword

Ask Pi to reshape the observed categories into your pillar and dump every keyword it saw, one file per group, into a folder. Be explicit that you want it comprehensive. 
these are simply different categories on the pillar. help me:
1. group all these categories so i can put them in the /uses pillar
2. dump ALL keywords observed from competitors, one file per group, be comprehensive,
   filed into a new folder.
Pi files a folder of grouped keyword dumps plus an overview of which categories are biggest and which are yours alone.
A summary table of seven keyword categories, each with a slug pattern, page count, keyword count, and top volume signal, totalling around 1,487 keywords.

​
5. Read the map and the dumps

Open the category map to see how each group maps to a slug pattern and a keyword pattern on your pillar, then open the individual files for the exact target keywords per page.
A category map table with columns for category, slug pattern, keyword pattern, and the markdown file for each group.
A single keyword-dump file for the input-type category, showing a summary, page candidates with primary and supporting keywords, which competitors run each page, estimated volume, and priority.

​
6. Prioritise your first-mover pages

Take the flagged unique angles seriously. Pi calls out the pages no competitor has built, which are your first-mover pillar pages.
Previous: How to find keyword opportunities · Next: How to check if you can rank for a keyword · Theory: Horizontal and vertical landing-page pillars, BOFU-first · Related: Design a landing-page pillar