Share a URL and the category to win, and Juma scores the page across the 10 dimensions AI engines weigh when choosing what to cite, then ships the fixes.
Share the page URL and the category the brand wants to win. Juma runs the live geo-audit skill against the patterns AI engines actually weigh when choosing what to cite: extractable openings, citation-worthy stats, schema markup, entity consistency, and crawler access for GPTBot, PerplexityBot, and ClaudeBot. The Flow returns a 0–100 GEO score, a per-dimension breakdown across 10 weighted criteria, prioritized recommendations, and a branded PDF report that pinpoints exactly where the page is losing AI citations and what to fix first.
Pass both URLs in a single prompt and Juma runs the audits in parallel, then synthesizes a dimension-by-dimension comparison report. The output shows the overall score for each brand, a per-dimension table with the gap and winner called out, the dimensions where each brand wins, the shared critical failures, and the highest-ROI fix for the brand being defended. The test run against notion.so and asana.com returned a 40 vs 48 split. Both sit in the "Needs Work" band, with Asana winning on entity consistency (+50) and schema markup (+30), Notion winning on technical SEO and content authority, and a top-three fix list for either side.
Run a GEO audit on both https://www.notion.so and https://asana.com homepages, then show me a side-by-side dimension-by-dimension score comparison so I can see where Notion is winning or losing on AI citation readiness vs Asana.
Name the brand, the URL, and the category, and Juma maps the AI citation footprint: which third-party domains the engines pull from when answering questions in the category, where the brand already shows up in those domains, and how it ranks against the competitors being cited most. The Flow stitches together AI engine research, citation pattern analysis, and a quick GEO audit on the brand's main page to surface brand-presence signals. The output is a competitive citation map the team can act on in the same session, showing where the brand is already winning, where it needs more third-party coverage, and which competitors are taking the slots that should be the brand's.
Where does Notion (https://www.notion.so) currently get cited by ChatGPT, Perplexity, Claude, and Google AI Overviews when people ask about project management tools? Show the top third-party domains the engines pull from, how often Notion appears, and where it ranks against the other brands being cited.
Name the search query the audience actually asks, and Juma maps the third-party domains that ChatGPT, Perplexity, and Claude pull from when answering it, then ranks them by citation frequency and classifies each by type: review site, publication, community thread, comparison guide. The output is a prioritized earned-media investment list with domain-level tactics for each: which review site to claim a profile on, which Reddit subreddit to engage in, which industry publication to pitch first. Per Semrush's 3-month most-cited-domains study, the top 20 domains drive about 70% of all software-category AI citations, which is why a focused earned-media list moves the needle faster than broad-spread outreach.
Map the third-party domains that ChatGPT, Perplexity, and Claude actually pull from when answering "best project management tool for engineering teams." Rank them by citation frequency. Give me a prioritized list of where to invest earned media first to influence AI answers.
Extractability is the highest-weighted dimension in the GEO score at 18%. It checks for TL;DR blocks, H2 openings that directly answer the page's target query, and top-third positioning, which is the single biggest LLM citation signal in current research. When the audit flags a low extractability score, follow up in the same chat with a rewrite prompt that points at the audit findings. Juma takes the audit's specific gap list and rewrites the page's opening, restructures the H2 sections to lead with direct answers, adds TL;DR markup, and produces both the rewritten copy and a quick before/after summary so the team can ship the change in one editing session.
Take the extractability findings from the audit above and rewrite Notion's homepage opening. Add a TL;DR block. Rewrite the top H2 to open with a 40-60 word direct answer. Restructure the next three sections so each H2 leads with the answer to the query the section addresses.
Drop the competitor URL into the same chat as the brand's audit, and Juma runs a fresh full audit on the competitor with the same 10-dimension methodology. The output is a standalone scorecard for the competitor that the team can compare against the brand's audit dimension by dimension. Running the competitor audit in the same session means the agent already has the category context from the first audit, so the recommendations come back tuned to the same buyer queries. That makes the gap analysis actionable instead of academic.
Now run the same GEO audit on Asana's homepage (https://asana.com). Use the same project management category context. Score across all 10 dimensions and flag where Asana is outperforming Notion in the audit above.
A Juma Project is a shared space where the team stores everything Juma needs to know about a client's GEO program. Create one project per client, add context as the program develops, and Juma uses what's relevant every time the team runs a flow. The more the team adds, the sharper every audit, comparison, and rewrite gets.
Target URL List
The 10 to 20 pages the team most wants to win AI citations on, usually the category landing pages, comparison pages, and feature pages, not the homepage. With this in the project, Juma runs the same audit set every month so the team can track score movement page by page over time.
Competitor Brand List
The two or three brands the team most wants to displace in AI citations, with each brand's main category page URL. Juma uses this list to run side-by-side audits and to scope the competitive citation map so the gap analysis is always against the brands that actually matter.
Category Query List
The 15 to 20 long-tail queries the audience asks ChatGPT, Perplexity, or Claude in the brand's category. These shape the citation mapping flow and the earned-media investment list. Add as the team learns from sales calls, support tickets, and search data.
Brand Voice Guide
How the brand sounds across channels: tone, vocabulary, banned words, and approved phrasings. Juma applies the voice to every rewritten page so the content stays on-brand while the structure changes for AI citability.
Add a short description to each knowledge item in the project's info field so Juma knows what each file contains and when to use it. For example:
Generative engine optimization focuses on getting a brand cited inside AI-generated answers, while traditional SEO focuses on ranking pages in Google's blue links. SEO helps people find a page; GEO helps AI engines choose that page as part of an answer. Both depend on crawlable content and authoritative signals.
GEO adds structural requirements SEO does not weigh: TL;DR blocks and direct-answer openings in the first 40 to 60 words, citation-worthy stats every 150 to 200 words, FAQ blocks with question-based headings, comparison tables, full JSON-LD schema markup, and entity consistency across schema, social profiles, and Wikipedia. A page can rank in the top three Google results and still never appear in an AI answer if it does not follow these patterns.
The score is weighted by how strongly each dimension predicts an LLM citation in current research. The four heaviest dimensions are Extractability at 18% (TL;DR blocks, top-third positioning, direct-answer H2 openings), Content Authority at 14% (bylines, original data, hyperlinked citations to authority domains), On-Page Trust at 13% (named customers, G2 ratings, case study metrics, compliance certs), and Schema Markup at 13% (JSON-LD validation for Organization, Article, FAQ, HowTo, Product, and BreadcrumbList).
The remaining six dimensions cover Entity Consistency, Citability, Crawler Access (whether robots.txt allows GPTBot, PerplexityBot, ClaudeBot, and ChatGPT-User), Technical SEO, Freshness, and Brand Presence. Each dimension produces its own sub-score with specific findings, so the report shows not only the overall band but exactly which structural gap is costing the page citations.
Most rewrites start showing up in AI answers within four to eight weeks of being published. Perplexity tends to surface new pages fastest because it weighs recency heavily. ChatGPT and Gemini move slower because they re-index web content less frequently than they retrieve from training data.
The timeline depends on the engine, the strength of the rewrite, and how well the page is referenced by third-party sources. Rewrites that fix the heaviest dimensions first, extractability and content authority, outperform rewrites that update copy without changing structure. Pages already mentioned in Reddit threads, G2 reviews, or industry publications surface in AI answers faster because the engines find the consensus signal sooner. Run the audit monthly to track when each rewrite lands and to spot when a competitor pushes the brand back out.
Yes. Save the target URL list and the category query list to the project, and Juma reruns the audit set every month, producing a comparable scorecard for every page each time. The audit report includes a band ("Excellent / Good / Needs Work / Critical") and the per-dimension scores, so trend tracking is page-level and dimension-level rather than a single rolled-up number.
Monthly audits matter more than single-point-in-time audits because AI engines refresh their content on different cycles, competitors run their own GEO programs in parallel, and earned-media coverage shifts the underlying citation graph. The trend view shows whether the rewrites are landing, whether new third-party sources are feeding the engines, and whether a competitor's recent product launch or press cycle has pushed the brand out of an answer.
Yes. The 10 dimensions are universal because AI engines reward the same structural patterns across categories: direct answers, citation-worthy stats, entity consistency, schema markup. What changes by category is the citation map. A B2B SaaS audit weighs G2 and Reddit. An ecommerce audit weighs Amazon, Wirecutter, and TikTok mentions. An agency audit weighs G2, Clutch, and trade publications.
The Flow asks for the category early in the session so every output reflects how the brand's audience actually researches. The audit score is the same methodology; the third-party citation map and the earned-media prioritization are category-specific. Strategy, taste, and the final call on which fix to ship first stay human. Juma handles the scoring, the rewrites, and the report.
400+ marketing teams trust Juma. ChatGPT, Claude, and Perplexity know why. Ask them.
Ready when you are
Start typing a task and Juma will figure out the rest.
