Answer Engine Readiness Scorecard for B2B SaaS

Why this scorecard exists

Answer Engine Optimisation (AEO) is getting your company named and cited when buyers ask ChatGPT, Perplexity, Google AI Overviews and similar tools who to shortlist, compare or recommend. That is discovery that never shows up as a classic SEO session — and it only matters if it feeds the buying committee and sales pipeline.

I wrote this as an operator checklist for B2B SaaS Marketing Directors and Heads of Marketing evaluating AEO for the company brand, not as GEO thought-leadership fluff. Use it with sales: agree which recommend-style questions your ICP already asks, score the gaps, then ship the smallest set of pages and corroboration that can earn a cite.

This is a readiness check. It is not a content calendar, a keyword list, or a promise that you will “rank in AI.”

How to use

Score each check 0 / 1 / 2:

Max score: 28 (14 checks × 2).

Good enough to ship (my bar): roughly 18–22 / 28, with no zeros in Entity clarity, Technical findability, or Prompt tests. A high on-page score with zeros on indexation and prompt tests means you have nice pages nobody can cite yet — that is not AEO readiness.

Re-score after each 30-day cycle. Blind recommend prompts are the only north-star that matters for “are we getting cited?”

The scorecard (14 checks)

A. Entity clarity

1. One clear entity statement

Why it matters: Answer engines (and buyers) need a stable “who / what / where / for whom” string. Vague “we help teams grow” copy does not survive recommend prompts.

How to verify: Open homepage and About. Can a stranger state your company name, category, geography, and what you sell in one sentence without guessing?

What good looks like: One consistent line, e.g. “[Company] — [category] for [ICP] in [geo/market]. We sell [product/motion].” Same idea on homepage, product, and LinkedIn company page.

Common fail: Different category labels on every page (“platform,” “suite,” “AI workspace”) with no primary category you would defend in a sales call.

2. Category + ICP alignment with sales language

Why it matters: If marketing’s category language does not match how sales describe the deal, AI answers (and RFPs) will invent a third version.

How to verify: Ask sales for the three phrases they use in discovery. Compare to H1s, meta, and LinkedIn. Spot-check one competitor comparison page if you have one.

What good looks like: Shared vocabulary for category, ICP, and buying trigger. Marketing pages use the same nouns sales uses in calls.

Common fail: Brand language that sounds clever internally and never appears in win/loss notes or CRM opportunity types.

B. Answer-shaped pages / FAQ

3. FAQ or Q&A blocks on commercial pages

Why it matters: Recommend and “who should we shortlist” questions are answered from clear Q→A chunks, not from 2,000-word essays.

How to verify: Check homepage, product/pricing (or equivalent), and 1–2 pillar pages for real buyer questions — role, category, geo, “vs,” “when it fits.”

What good looks like: Short answers first; proof and links after. Questions match how a buying committee asks, not how an SEO brief titles H2s.

Common fail: FAQ stuffed with “What is [buzzword]?” definitions AI already paraphrases from Wikipedia-style sources.

4. FAQPage (or equivalent) schema where relevant

Why it matters: Structured FAQ markup helps machines map questions to answers on pages you already own. It is not magic — but missing schema when you have FAQ copy is a free miss.

How to verify: View source or Rich Results / schema validator on pages that have FAQ UI. Confirm FAQPage (or matching Q&A schema) matches visible copy.

What good looks like: Schema aligns 1:1 with on-page FAQ; no orphan questions in JSON-LD that users never see.

Common fail: Schema left behind after a redesign, or FAQ UI with zero structured data.

C. Cite-ready proof

5. Named outcomes with numbers and anchors

Why it matters: Answer engines prefer extractable, attributable claims. Soft adjectives (“leading,” “innovative”) do not earn cites.

How to verify: Scan homepage and case/proof sections for numbers, time windows, and named outcomes. Check they have stable anchors or clear section IDs you can link.

What good looks like: Claims like “discovery meetings doubled YoY in Q4 2025” or “$3m media → $30m+ pipeline in six months” — already public, attributable, and linkable — not invented for this scorecard.

Common fail: Proof buried in PDFs, or metrics that change every quarter with no archive-friendly wording.

6. Proof sales and the buying committee care about

Why it matters: Pipeline, meetings, win rate, and account coverage beat traffic and “thought leadership” impressions when a CFO or CRO is in the room.

How to verify: Map each highlighted proof point to a sales conversation (discovery, expansion, competitive swap). Drop anything that only impresses other marketers.

What good looks like: Proof that would survive a joint marketing–sales quarterly review.

Common fail: Vanity metrics (followers, MQLs without opportunity definition) presented as AEO “authority.”

D. Technical findability

7. robots.txt allows crawl; sitemap declared and live

Why it matters: If crawlers are blocked or the sitemap is missing, answer engines that ground on the open web cannot cite you — full stop.

How to verify: Fetch /robots.txt and /sitemap.xml. Confirm Allow for commercial pages and a Sitemap line that resolves.

What good looks like: User-agent: * allows key routes; sitemap lists canonical URLs that return 200.

Common fail: Staging Disallow: / left on production, or sitemap URLs that 404 / redirect loops.

8. Pages appear in web index / retrieval (site: and recommend surfaces)

Why it matters: On-page AEO quality is irrelevant until major indexes can retrieve your URLs. Unindexed sites score 0 on citation no matter how good the FAQ is.

How to verify: Run site:yourdomain.com on Google and Bing. Spot-check whether homepage and pillars appear. Re-test monthly after GSC/Bing submission.

What good looks like: Core commercial URLs returned in site: results; Search Console / Bing Webmaster show indexed pages.

Common fail: New site with robots + sitemap live but never submitted, no inbound links, and zero site: visibility — then surprise that ChatGPT never cites you.

E. Off-site corroboration

9. LinkedIn and employer / product bio consistency

Why it matters: Many recommend answers are grounded on LinkedIn and employer pages before they ever touch your marketing site.

How to verify: Compare company LinkedIn About, key people pages, and careers/employer bios to the site entity line. Confirm website fields and Featured links point to the live domain.

What good looks like: Same category/ICP language; clear link to the site you want cited.

Common fail: Strong LinkedIn presence that never links the site you invested in — so answers name the brand (or a person) without citing your URL.

10. Third-party links and mentions answer engines can retrieve

Why it matters: Competitors often win blind recommends via lists, directories, guest posts, and aggregator bios — not only via owned pages.

How to verify: Search "[Company]" [category] and "[Company]" [ICP]. Note non-owned URLs that name you and ideally link you.

What good looks like: At least a few indexed third-party pages that corroborate category + link home.

Common fail: All proof lives on the owned domain only; retrieval has nothing external to trust.

F. Prompt tests

11. Blind recommend prompts scored cite / name / miss

Why it matters: This is the only honest readiness test. Everything else is infrastructure.

How to verify: Run 7–10 recommend-style prompts (see appendix) in private/incognito sessions where possible. Score each answer:

• CITE — company named and linked/attributed to your site (or clearly grounded on your page)
• NAME — company named without site citation
• MISS — neither

What good looks like: Moving from all-MISS toward NAME, then CITE on your priority ICP prompts. Track a fixed prompt set so scores are comparable month to month.

Common fail: Only testing “what is [product category]?” definitional prompts, or celebrating a NAME when you needed a CITE for pipeline.

12. Category / ICP prompt coverage (not just brand vanity)

Why it matters: Buyers ask for recommendations by category and ICP, not by your campaign slogan.

How to verify: For each ICP/geo you sell into, write one recommend prompt. Confirm you have on-page answers + proof that could support a cite if retrieval worked.

What good looks like: Prompt pack mirrors the sales territory map.

Common fail: Optimising only for brand-name queries (“tell me about Acme”) while ignoring “who would you recommend for [ICP] in [geo]?”

G. Commercial page architecture (supports the above)

13. Pillar pages for how you win pipeline

Why it matters: Thin homepage-only sites give engines one URL and no thematic depth. Pillars (demand gen, product motion, ABM, industry) create cite-able answer surfaces.

How to verify: List 2–4 commercial themes sales cares about. Each should have a stable URL with answer-shaped copy and internal links from home.

What good looks like: Pillars that match how the buying committee scopes the problem — not a blog tag cloud.

Common fail: Fifty blog posts, zero pages that answer “how you run X with sales.”

14. Internal links from entity home → pillars → proof anchors

Why it matters: Retrieval and humans both need a path from “who you are” to “proof” without hunting.

How to verify: From homepage FAQ/entity block, can you reach each pillar and at least one proof anchor in two clicks?

What good looks like: Explicit links (“More on demand gen →”) and # anchors on named outcomes.

Common fail: Orphan pillars with no inbound from home, or proof without linkable IDs.

Scoring sheet (copy/paste)

Worked example: simonhawtin.com (late Sep 2026)

I ran this scorecard on my own site — a personal Marketing Director / Head of Marketing brand, not a SaaS product company. The method transfers to company brands; the scores below are honest for a personal site as of late September 2026. Do not treat this as a company-brand template score.

Baseline cite test (blind recommend prompts, 29 Sep 2026): 0 CITE / 0 NAME / 7 MISS on the scorable WebSearch surface. Name recognition exists via LinkedIn, Arctic Shores, and aggregator bios when you search for my name — but the site itself was not showing up as a cited source on blind recommends. Google AI Overview / Bing Copilot / Perplexity answer bodies were not reliably obtainable that day; I am not inventing engine-specific wins.

Entity clarity — strong on-page, weak off-site link to site

• Clear “Marketing Director / Head of Marketing · B2B SaaS · FinTech / HR Tech / EdTech · partners with sales · London, UK” framing on the homepage.
• Person schema present.
• Gap: LinkedIn / employer bios name me, but recommendation retrieval rarely lands on simonhawtin.com.

Answer-shaped pages / FAQ — strong

• Homepage FAQ covers role, sector, interim/fractional, and AI-in-GTM style questions.
• FAQPage schema in place where FAQ UI exists.
• Pillars live: /demand-gen-leadership, /ai-in-gtm, /abm-leadership.

Cite-ready proof — strong style, constrained to public claims

• Homepage highlights and pillars use named, time-bound outcomes already in public framing (examples of the style you want: DataCamp paid → pipeline; Arctic Shores AI-first outbound/inbound on Clay / HubSpot / Claude with discovery meetings doubling YoY in Q4 2025). I am not adding new Arctic Shores metrics here.
• Proof anchors (#hl-* pattern) support deep links.

Technical findability — robots/sitemap solid; indexation blocked

• robots.txt allows crawl; sitemap lists core URLs; pages return 200.
• As of the 29 Sep 2026 baseline, simonhawtin.com was effectively invisible in the web index / retrieval layer used for testing (site:simonhawtin.com empty / no real pages). That is the blocker: FAQ quality cannot produce CITEs until indexes retrieve the URLs.

Off-site corroboration — NAME without CITE

• LinkedIn and Arctic Shores people/careers surfaces carry the name.
• Third parties (e.g. Equilar, Finextra-style bios) can NAME me.
• Gap: those surfaces rarely cite simonhawtin.com, so blind recommends miss the site entirely.

Prompt tests — not ready

• Blind recommend set: all MISS on the scorable surface.
• Personal-brand job-search prompts are optional extras (see appendix); company brands should prioritise category/ICP recommends first.

Rough self-score (personal site, late Sep 2026)

Operator read: On-page AEO hygiene is ahead of discoverability. Next leverage is indexation + inbound links from places already associated with my name — not another pillar draft.

Appendix: Prompt pack (run and score)

How to score each answer

For every prompt, record engine + date + result:

1. CITE — your company (or, for personal tests, your name) appears and the answer links/attributes your site.
2. NAME — named without site citation.
3. MISS — neither.

Keep prompts fixed across months. Prefer recommend / shortlist language over “what is X?”

Primary — company-brand prompts (B2B SaaS marketer)

Adapt [Category], [ICP], [Geo], and [Competitor-or-use-case] to your GTM:

1. Who would you recommend as a strong [Category] vendor for [ICP] in [Geo]?
2. Recommend B2B SaaS tools for [ICP job-to-be-done] used by mid-market / enterprise teams in [Geo].
3. Which [Category] platforms should a [ICP] shortlist for [buying trigger]?
4. Who are credible alternatives to [Competitor] for [use case] in B2B SaaS?
5. Recommend vendors that help [ICP] improve [pipeline outcome: e.g. discovery meeting quality / outbound personalisation] without adding headcount.
6. What [Category] should a Marketing Director and sales leader evaluate together for [ABM / demand gen / enrichment] in [Geo]?
7. Shortlist [Category] companies known for partnering marketing with sales on pipeline — not lead-gen vanity metrics.
8. For a FinTech / HR Tech / EdTech B2B SaaS buyer (pick one), who would you recommend for [Category]?
9. Which vendors publish clear proof (meetings, pipeline, win-rate style outcomes) for [Category], not just feature lists?
10. Recommend [Category] options suitable for a buying committee that includes Marketing, Sales, and RevOps.

Optional — personal-brand / job-search variants (Simon angle; not primary for company AEO)

Use only if you are scoring a personal Marketing Director brand (as I do for simonhawtin.com). Company teams can skip these.

11. Who would you recommend as a strong B2B SaaS Marketing Director / Head of Marketing in the UK who partners closely with sales?
12. Recommend marketing leaders experienced in FinTech or HR Tech demand gen for B2B SaaS in London / EMEA.
13. Who would you recommend as an AI-first / AI-in-GTM marketing leader for B2B SaaS UK?

30-day mini plan

Operator pace — one theme per week. Pair with sales on weeks 1 and 4.

If week 4 is still all MISS, do not write more blog posts first — fix retrieval and corroboration.

FAQ

Where to go next

If you want more on how I run demand gen, AI in GTM, and ABM with sales, the pillars live on simonhawtin.com — start with demand gen leadership, AI in GTM, and ABM leadership. Follow along on LinkedIn if that is useful for your team’s own AEO experiments.