Before we run the audit, we need to make sure we're asking the right questions about the right competitors to the right buyers. This document presents what we've learned about Russell Innovation Center for Entrepreneurs' market — your job is to tell us what we got right, what we got wrong, and what we missed.
Before we measure how often AI assistants cite RICE in the Black-founder support and small-business incubator space, these three signals tell us whether AI crawlers can reach, trust, and read your site at all. They set the baseline every later section builds on.
AI search is quickly reshaping how people find a nonprofit entrepreneurship center and business incubator supporting Black and underestimated small business owners — and it is doing so on two fronts at once. Both the founders who might join RICE and the corporate, philanthropic, and civic funders who finance the mission are increasingly asking AI assistants "who does this well in Atlanta?" As an established, well-funded regional anchor, RICE enters this shift with real authority to defend — and the organizations that establish AI visibility now compound that lead as citation engines learn to trust their domains.
This Foundation Review is what we're validating together before the audit runs. It lays out the competitive landscape that shapes how we construct head-to-head queries, the buyer personas that determine which search intents we test, and the Layer 1 technical baseline that determines whether AI platforms can access and correctly read RICE's content in the first place. Each section exists to be confirmed or corrected by the people who know RICE's deals and donor conversations better than any analysis can.
The validation call is a working decision session, not a presentation. It resolves two kinds of questions: input validation (are the right buyers and competitors in the right tiers, and is RICE's dual entrepreneur-and-funder audience weighted correctly?) and engineering triage (what technical fixes can start immediately, before results come back?). Your answers to the first set directly determine the buyer query set we run across the selected AI platforms; the second set your engineering team can begin today. The TL;DR below names the specific items.
Three quick notes on what this document is for, what we need from you, and how to read the confidence signals throughout.
What This Is This is the foundation for a Generative Engine Optimization (GEO) audit — measuring how AI assistants like ChatGPT, Claude, Perplexity, and Google's AI answers represent RICE when people ask about a nonprofit entrepreneurship center and business incubator for Black and underestimated founders in Atlanta. It presents the knowledge graph we built (personas, competitors, features, pain points) and the technical baseline of russellcenter.org. It is not the audit itself and not a content plan — those come after we measure real AI responses.
What We Need From You Read each section as a draft we're asking you to red-line. Where you see a purple question, that's a point where your answer directly changes how we build the audit. Tell us what's right, what's wrong, and — most valuably — what's missing. You know your entrepreneurs and your funders better than any outside analysis.
Confidence Badges Every entity carries a confidence badge. High means it's well-sourced from RICE's own site or clear public signals. Medium means we inferred or partially sourced it and specifically want your confirmation. Low-confidence and inferred items are called out for validation — treat medium badges as our honest "please check this."
The baseline profile that anchors every query we generate. Confirm the identity signals — especially the name variants an AI assistant might use to refer to you.
→ RICE is a nonprofit serving two fundamentally different buyers: the Black entrepreneurs it supports (who choose to become stakeholders) and the corporate, foundation, and civic funders who finance the mission — and these two audiences search in entirely different language. Which audience should the audit weight as the primary one we optimize your AI visibility for? If it's funders, the query set tilts toward "measurable jobs/wealth outcomes" and "credible Atlanta economic-inclusion partner"; if it's entrepreneurs, it tilts toward "coworking, mentorship, and capital access for my business." Get this wrong and we misreport where you're visible.
5 personas — 3 decision-makers, 2 influencers — spanning both sides of RICE's mission. Personas drive the query set: each one searches differently, so each becomes a distinct cluster of buyer queries we test across the AI platforms.
Critical Review Area Personas are the single biggest lever on the audit. If a persona is wrong, missing, or mis-weighted, every query we generate for that buyer is off-target. RICE's dual entrepreneur-and-funder audience makes this especially high-stakes — please scrutinize both sides.
Data Sourcing Note Names, roles, seniority, veto power, and technical level are drawn from RICE's site, review/case-study mining, and category signals. The buying jobs and query focus areas are synthesized by us from those inputs — they're our best model of how each buyer actually searches, and exactly the kind of thing we want you to correct.
→ Marcus is the only persona tied to the market-access/procurement pain — a very different search ("how do I land corporate contracts") than Tasha's early-stage queries. Do growth-stage founders search distinctly enough to justify their own query cluster, or do Marcus and Tasha collapse into one entrepreneur searcher (which would free that cluster for the funder side)?
→ Tasha is medium-confidence and pre-revenue — does an aspiring founder actually search for RICE (or an incubator) by name, or by generic top-of-funnel queries like "how to start a business in Atlanta"? If the latter, her cluster needs informational/awareness queries rather than branded comparison queries, which changes what "winning" looks like for her.
→ Danielle and Raymond are both funders (see the audience-priority question above), but a corporate CSR lead and a foundation program officer may vet very differently. Does Danielle evaluate RICE on the same impact-reporting criteria as a foundation, or does corporate procurement / supplier-diversity framing dominate her searches? If it's the latter, her query cluster leans toward "supplier diversity" rather than "grantee outcomes."
→ Raymond is rated technical level high — unusual for a grantmaker. Does a foundation program officer really vet RICE with data-heavy queries (outcomes methodology, jobs-created figures), or is "high" overstated? If he's genuinely data-driven, we test rigorous impact-evidence queries; if not, we lean on reputation and narrative queries instead.
→ Patrice is inferred, not sourced from named review data — we built her from RICE's documented civic funding relationships. Is a public economic-development director an actual buyer who evaluates RICE, or a downstream stakeholder who never independently searches? If she doesn't hold, we drop this persona and reallocate her query budget to the confirmed entrepreneur or funder clusters.
→ Who else shows up in your deals? A few roles sometimes appear in economic-inclusion and nonprofit-incubator conversations — do they show up in yours? Individual major donor / philanthropist (a person giving personally, who searches very differently from an institutional foundation officer); ecosystem partner (another nonprofit, university, or accelerator looking to co-program rather than fund); government grant administrator (federal MBDA/EDA/SBA officer where public grant compliance is its own conversation). If any of these drive real decisions for RICE, each warrants its own query cluster — which ones are missing?
6 primary + 4 secondary competitors identified. Tier assignments determine which organizations get head-to-head query coverage in the audit and which appear only in category-awareness queries.
Why Tiers Matter Each primary competitor earns roughly 5–8 head-to-head queries — comparisons like "RICE vs. digitalundivided" or "best Black entrepreneur programs in Atlanta." With 6 primaries, that's on the order of 35–45 queries riding on getting these tiers right. Three primaries carry medium confidence on tier — Start:ME Atlanta, The Village Market ATL, and 1863 Ventures — and 1863 Ventures in particular is a national, Washington-DC-rooted accelerator rather than an Atlanta place-based peer. If it rarely surfaces in RICE's stakeholder and funder conversations, moving it to secondary would shift roughly 6–8 queries out of the head-to-head set toward your true local rivals.
→ Three questions on this set: (1) Missing rivals — who else surfaces when a funder or founder asks an AI assistant for "the best organization supporting Black entrepreneurs in Atlanta" that isn't here? (2) Tier accuracy — do the three medium-confidence primaries (Start:ME Atlanta, The Village Market ATL, 1863 Ventures) actually compete for your entrepreneurs and funders, or are national/consumer-market players spending head-to-head query budget that belongs to your local peer set? (3) Irrelevant listings — is any competitor here (e.g., the for-profit Gathering Spot, or Operation HOPE, added by inference) one you'd never actually be compared against?
10 buyer-level capabilities mapped. These determine which capability queries the audit tests — "does RICE actually offer X?" — so the honest strength ratings matter more than marketing polish.
A professional place to work, meet clients, and host events without paying for my own office — desks, meeting rooms, and event space I can actually afford.
Experienced mentors and peer coaching groups that have actually built businesses and can help me solve real problems.
Hands-on programs and academies that teach me how to run and grow my business, not generic webinars.
A community of hundreds of other Black founders and partners who open doors, share referrals, and get what I'm going through.
Clear proof that my funding creates jobs, revenue, and wealth — numbers I can put in front of my board and community.
Help getting my business ready for funding and warm introductions to grants, lenders, and investors.
Can they actually write me a check or a loan, or do they just point me to someone else?
Introductions to big corporate buyers and supplier-diversity programs so I can land real contracts.
Specialized help for my industry — getting retail-ready or adopting manufacturing and AI tools — not one-size-fits-all advice.
Professional space and support to produce content, tell my story, and build my brand so customers and investors notice me.
Prioritization Question Five capabilities are rated Strong — Coworking & Innovation Space, Mentorship & Coaching Circles, Business Education, Entrepreneur Community & Stakeholder Network, and Impact Measurement & Outcomes Reporting. The audit tests all 10 capabilities, but competitive differentiation queries will emphasize three. Which of these five best represents where RICE actually wins — where a founder or funder chooses you over a peer? Your answer sets which capabilities we press hardest in the head-to-head queries.
→ Two ratings we most want you to challenge: we scored Direct Financing/Loans as Weak and Capital Readiness as Moderate — because RICE convenes and connects to capital but isn't itself a lender/CDFI like ACE or 1863 Ventures. Is that fair, or do you deploy more direct capital than we credited? Buyers frequently search "who will actually fund my business," so mis-rating this either overclaims a capability you don't offer or hides a real strength — both distort your findings versus ACE. Separately: are any of the four "Moderate" capabilities really strengths, and should any two capabilities (e.g., Capital Readiness and Corporate Partnerships) be merged?
9 pain points: 5 high, 4 medium severity. The buyer language here is literally how we phrase queries — an AI assistant answers the question the way a frustrated founder or funder would actually type it.
→ Three checks: (1) Severity — we rated the market-access/procurement pain High but tied it to only one persona (the growth-stage founder), while rating the brand-visibility pain Medium; are those calibrated right for how often each actually blocks a deal? (2) Buyer language — does "nobody ever taught me how to actually run and grow a company" sound like your founders, or is the real frustration phrased differently? (3) Missing pains — do these show up for RICE's buyers that we didn't capture: founder isolation/burnout and mental load, navigating business licensing/regulatory red tape, or the digital/technology-adoption gap for legacy small businesses? Which are we missing?
A technical read of russellcenter.org for how well AI crawlers can access and extract your content. These are engineering and content fixes your team can act on now — separate from the content strategy the full audit will prioritize.
For Engineering — Act Now Good news first: robots.txt is present and explicitly allows every major AI crawler (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Googlebot), so crawler access is not the blocker here. The two high-severity items are: (1) injected Polish-language casino spam posts are published and indexed — likely a WordPress content-injection compromise that needs the site secured and the spam removed, and (2) nearly the entire site (all 9 blog pages and 11 of 17 program/commercial pages) is stale relative to the window AI assistants favor. No critical blockers, but the spam-injection finding is a security issue engineering should treat with urgency.
What we found: The published post sitemap (wp-sitemap-posts-post-1.xml) contains two Polish-language online-casino spam posts (lastmod 2026-06-26 and 2026-03-02). One appears directly in the /blog/ index alongside legitimate stakeholder stories. This content is unrelated to RICE and is a strong indicator of a WordPress content-injection / spam-post compromise.
Why it matters: AI crawlers use domain-level trust and topical-consistency signals. Off-topic gambling spam indexed under a nonprofit entrepreneurship domain pollutes the site's topical footprint, can trigger safety/spam classifiers that suppress the whole domain in AI answers, and wastes crawl budget on junk URLs instead of RICE's program pages. It also signals an unpatched security hole that could escalate.
Recommended fix: Delete the spam posts, audit the WordPress install for unauthorized users/plugins and injected content, rotate admin credentials, update WordPress core and plugins, and add a security plugin (e.g. Wordfence) plus a spam-post monitor. Regenerate the sitemap after cleanup and request re-crawl.
What we found: Category-weighted freshness is 0.19. Every content-marketing page scored (blog posts and hubs) is older than 180 days (content_marketing avg 0.13; 0 of 9 updated within 90 days). Key commercial pages carry last-modified dates well over six months old: Content Studio (2025-03-11), Host Your Event (2025-01-13), Big IDEAS (2024-03-30), the Beyond Innovation campaign (2025-02-06), and Support RICE (2024-01-30). 11 of 17 program/commercial pages are older than 180 days.
Why it matters: AI assistants concentrate citations on recently updated content — AI-cited pages run 25.7% fresher on average than typical results (Ahrefs, August 2025), and 76.4% of ChatGPT's most-cited pages were updated within the last 30 days (ConvertMate, ChatGPT-scoped). When RICE's program and story pages look stale, fresher content from Atlanta peers is preferred for the informational queries RICE should win.
Recommended fix: Establish a refresh cadence for the highest-value program pages (Retail Readiness Academy, Supply Chain Accelerator, Coaching Circles, Big IDEAS, Content Studio) and the RICE Report — updating cohort dates, outcomes, and last-reviewed dates at least quarterly. Add visible "last updated" dates to program and blog pages so crawlers can read recency.
What we found: The page sitemap exposes numerous live, crawlable near-duplicate and working-draft pages — including about-us-new, about-us-new-2, home-new, russell-innovation-center-for-entrepreneurs-new, big-ideas-2, contact-us-new, stakeholder-directory-new, rice-report-draft-page, and one literally named "blog-for-review-page-delete-after." Separately, two blog hubs run in parallel: /blog/ ("Stories from Our Community") and /the-rice-effect/ ("The RICE Effect Blog").
Why it matters: Multiple live copies split internal-link authority and create duplicate-content ambiguity, so crawlers and AI models can't tell which About/Home/Big IDEAS page is canonical. It wastes crawl budget on throwaway pages and risks an AI model quoting an outdated draft instead of the live page.
Recommended fix: Unpublish or 301-redirect every draft/duplicate/template page to its canonical equivalent, delete the explicitly-marked "delete-after" page, and consolidate the two blog hubs into one. Add canonical tags and regenerate the sitemap so only production pages remain.
What we found: The Stakeholder Directory (/stakeholder-directory/) is a live, linked page whose body reads "Coming Soon..." with three sample bios. The /support-rice/ page renders with the H1 "Welcome to RICE Tech Support" and title "RICE Tech Support" even though the URL and navigation present it as a donation/support page — a title-to-purpose mismatch. The /schedule-a-tour/ page is a bare booking form with almost no descriptive content (content depth 0.2).
Why it matters: A page's title and first heading are among the strongest signals an AI model uses to understand what a URL is about. A donation page titled "Tech Support" will be miscategorized, and a "Coming Soon" directory that is nonetheless linked and indexed gives crawlers an empty, low-value page where users expect the flagship stakeholder network.
Recommended fix: Correct the /support-rice/ title and H1 to match its donation purpose; either finish and populate the Stakeholder Directory or unpublish it until ready; and add a short descriptive intro paragraph above the tour booking form.
What we found: On multiple pages the only H2s in the rendered output are the repeated footer navigation blocks ("Who We Are," "Get Involved," "Tools," "Connect") while the real content headings are pushed to H3 or absent. This appears on /support-rice/, /schedule-a-tour/, /stakeholder-directory/, /news-press/, /the-rice-effect/, and the Melissa Bradley and Kristen Dunning blog posts. Many program pages also rely on all-caps stylistic headings ("GET RETAIL READY!", "WHO IS IT FOR?").
Why it matters: AI models use heading hierarchy to segment a page into citable passages. When the meaningful headings are demoted below repeated footer nav, or replaced by all-caps stylistic labels, the model struggles to identify which section answers a given question — lowering passage-level extractability and citation odds.
Recommended fix: Ensure each page has one descriptive H1 and a logical H2→H3 body hierarchy distinct from footer navigation. Convert footer nav to non-heading markup, and rewrite stylistic all-caps headings as descriptive noun phrases (e.g. "Who the Retail Readiness Academy Is For").
What we found: Program, event, and commercial pages (Coaching Circles, Content Studio, Host Your Event, Big IDEAS, Join Our Community, Georgia AIM, Partner With Us) show no on-page published or last-updated date; recency could only be inferred from sitemap lastmod values. Blog posts do carry visible dates, but the program pages that most need to signal "current cohort / current offering" do not.
Why it matters: When no date is visible in the rendered page, AI crawlers can't confirm recency and withhold freshness credit even when the content is current. For time-sensitive offerings like application windows and cohort programs, an undated page reads as potentially outdated and is deprioritized against dated competitor content.
Recommended fix: Add a visible "Last updated" or "Program year 2026" date to program and event pages, and ensure cohort/application dates reflect the active cycle. Keep sitemap lastmod accurate so header-level and on-page signals agree.
The following items could not be assessed through our analysis method (rendered markdown). We recommend your engineering team verify these manually before the validation call.
What to check: Our analysis reads rendered markdown, which doesn't expose JSON-LD blocks, so we couldn't confirm whether Organization, FAQPage, Event, or Article schema is present. RICE has strong candidates for structured data: an FAQ page, a recurring events calendar, program pages, and blog articles.
Recommended action: Verify current schema with Google's Rich Results Test. Add FAQPage schema to /faq/, Event schema to RICE events, and Article schema to blog posts if not already present.
What to check: Our method returns rendered page text, not the HTML head, so meta descriptions, canonical tags, and Open Graph/Twitter tags aren't visible. Given the duplicate-page issue above, missing or inconsistent canonical tags are a plausible risk worth confirming.
Recommended action: Spot-check meta descriptions, canonical tags, and OG/Twitter tags using view-source or a crawler like Screaming Frog. Ensure each canonical page self-references and that duplicate/draft pages canonicalize to production or are removed.
What to check: Every page fetched returned substantial server-rendered body text (this is a WordPress site), so we saw no evidence of client-side-rendering gaps that would hide content from crawlers — but we can't inspect raw HTML or run the page with JavaScript disabled through this method, so CSR can't be positively ruled out.
Recommended action: Load key templates with JavaScript disabled, or use Google's URL Inspection "view crawled HTML," to confirm body content is present in the initial server response — paying attention to embedded/widget-driven sections (the directory once built, events widget, video repository).
Read With Care Two metrics could not be scored by our method and are marked "Unable to assess": schema coverage (all 35 pages — JSON-LD isn't visible in rendered output) and structural/reference-page freshness (9 pages with no detectable date). These are on the Manual Verification Checklist above, not failures — your engineering team should confirm them directly rather than treating them as zero.
Why Now GEO is a timing opportunity, and the window is open:
• AI-driven discovery is shifting quarter over quarter — both founders and funders increasingly start with an AI assistant rather than Google.
• Early citations compound: domains AI platforms learn to trust now get cited more often as those systems accumulate signal.
• Peers who establish AI visibility first create a structural disadvantage for late movers in the same Atlanta market.
• Black-founder support and economic-inclusion nonprofits are still early-innings in GEO — right now RICE competes mostly against inaction, not against entrenched strategies.
Once you validate this foundation, the full audit measures how often AI assistants actually cite RICE across the queries your buyers really run — from "best coworking and mentorship for Black founders in Atlanta" to "which Atlanta nonprofit creates measurable jobs and wealth." You'll see exactly which of those queries return your competitors but not RICE, and what it would take to appear in them. Fixing the spam-injection and staleness issues now improves your baseline before we even start measuring — so the audit captures RICE at its cleaned-up best.
A 45–60 minute working session to walk through this document, confirm the inputs, and resolve the open questions — especially the entrepreneur-vs-funder audience priority.
We generate the buyer query set from your validated personas, competitors, features, and pain points, then run it across the selected AI platforms.
Visibility analysis, competitive positioning, and a prioritized three-layer action plan showing exactly where and how to win more AI citations.
Engineering Can Start Now Three Layer 1 technical fixes don't depend on the rest of the audit and will improve your baseline before we even measure it: (1) delete the two Polish casino spam posts and audit the WordPress install for the injection vector (rotate credentials, update core/plugins, add a security plugin); (2) unpublish or 301-redirect the ~12 draft/duplicate/template pages and consolidate the two parallel blog hubs; (3) verify JSON-LD schema with Google's Rich Results Test and add FAQPage/Event/Article schema where missing. Note: robots.txt is already confirmed open to all major AI crawlers, so no crawler-access work is needed there.
Two jobs before we meet. The questions on the left require your judgment — no one knows your business better than you. The engineering tasks on the right don't require the call at all.