Which one is right for your team — and what you’ll actually hold in your hands after a month with each.
The deepest data in the category. Priced and sold like enterprise software — because it is.
Self-serve, transparent, and architecturally the most interesting. Now a Sitecore product.
Both tell you where you’re losing in AI answers. Neither writes what wins. More on that below.
Both founded within three years. Both rated 4.6/5 on G2. Then 2026 split them: Profound raised its way to a $1B valuation. Scrunch sold.
“The full stack marketing platform for AI search” — answer engine insights, demand data, and agent analytics for enterprise brands.
“The AI customer experience platform” — visibility monitoring, persona-based journeys, and an AI-readable mirror of your site served to crawlers.
Most Profound-vs-Scrunch articles were written in 2025 and are now materially out of date. Two events this year changed the calculus.
Led by Lightspeed. Roughly $155M raised in under two years. The round funds a push beyond measurement — an Agents execution layer, Profound Sheets, deeper enterprise. Profound is now the category’s default incumbent, with the pricing power that implies.
Announced June 3, 2026. Founder Chris Andrew stays on as division CEO; Scrunch folds into the SitecoreAI platform. For Sitecore customers, a clear win. For everyone else, open questions — roadmap priority, pricing continuity, and whether Scrunch stays a product or becomes a DXP feature.
| Core focus | Enterprise AI search intelligence — visibility, demand data, citations, agent analytics. | AI visibility plus a delivery layer that serves AI-readable versions of your site to crawlers. |
|---|---|---|
| AI platforms monitored | ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews, Copilot, Grok, Meta AI, DeepSeek. | ChatGPT, Claude, Gemini, Perplexity, Google AI Mode & AI Overviews, Copilot, Meta AI, Grok.Full platform list has historically been gated to Enterprise. |
| Demand data | Prompt Volumes — AI query demand estimated from a consumer panel Profound says covers 400M+ real conversations.Unique in the category; practitioners debate how verifiable the estimates are. | AI Search Trends — prompt volumes across 1,500+ curated topics.Reviewers describe topic coverage as thinner. |
| Methodology | Scheduled prompt runs across engines, layered with panel data and citation analysis at scale. | Prompt runs segmented by persona (3–5 by plan), topic, model, and region; share-of-answer and position scoring. |
| Hallucination detection | Sentiment and misrepresentation surfaced within answer insights. | Dedicated hallucination detection that flags false claims about your brand.Enterprise plans only. |
| Bot & agent analytics | Agent Analytics — AI crawler activity and AI-referred traffic.Reviewers note it reports page categories, not exact URLs. | Real-time AI crawl feed (GPTBot, ClaudeBot, PerplexityBot) plus LLM referral traffic with GA4 integration. |
| Action layer | Agents — template-based workers (FAQ generation, content refresh) that can publish to your CMS behind a human approval step.Launched recently; third-party reviews still describe it as nascent. | Agent Experience Platform (AXP) — serves a restructured, token-light mirror of existing pages to AI crawlers at the CDN layer.Transforms what you already published; does not create new pages. |
| Net-new content | Early, template-driven, approval-gated. | No — recommendations and page audits only. |
| Entry price | No public self-serve pricing. Reviewers report roughly $499/mo entry, with custom enterprise pricing for full coverage. | $250/mo billed annually ($300 monthly) for Starter; Growth $417/mo annually; Enterprise custom. |
| Free trial | No — demo-led sales process. | 7-day free trial, no credit card. |
| Security & compliance | SOC 2 Type II, HIPAA, SSO/SAML, dedicated account teams. | SOC 2 Type II, SAML/OIDC SSO, RBAC, multi-site and multi-region support. |
| Best fit | Enterprise and regulated brands with global footprints and budget for a sales-led platform. | Mid-market teams and agencies that want to start fast — now operating inside Sitecore. |
Compiled June 2026 from both companies’ public sites and pricing pages, G2 reviews, and third-party platform reviews. Profound’s pricing is not published; figures are as reported by reviewers and may vary by contract. No affiliate links. No sponsored placements.
Forget the feature names. This is what actually lands on your desk a month in.
What neither hands you: the published page. Both backlogs end at “someone should write this.”
Strip the marketing and both platforms share an engine: run prompts on a schedule, record the answers, score your presence. The difference is the layer on top.
The catch: practitioners question whether anyone can verify panel-based volume estimates. Directional, not gospel.
The catch: credits historically burned per engine — 125 prompts across 4 engines meant ~31 real queries. The 2026 tiers raised limits; check the math anyway.
Each vendor attacks the other’s methodology. Profound calls rival prompt sets “synthetic.” Scrunch notes Profound can’t deliver content to crawlers. Both critiques are self-interested. Both contain some truth.
This is the real fork in the comparison — and where most write-ups go shallow.
Evaluating Profound? Ask to see Agents live on your own site — not the slide.
The debate: competitors say serving bots different content flirts with cloaking. Scrunch says Google and Bing see the normal site. Make your own call.
Profound’s Agents fill templates. Scrunch’s AXP reformats what exists. The article that doesn’t exist yet — stays not existing.
Same category. Opposite go-to-market. This is the clearest difference between the two — and the easiest way to disqualify one of them today.
Skip it if you’re a lean team without enterprise budget. The entry tier is deliberately limited, and the platform is built — and priced — over your head.
Skip it if you need heavy reporting and exports, full engine coverage on a starter budget, or roadmap certainty — post-acquisition, those calls belong to Sitecore.
Run either platform for a month. You’ll end up with the same artifact: a list of prompts where competitors get recommended and you don’t.
The dashboard’s job ends there.
Someone still has to research the gap. Write the page that deserves the citation. Publish it. Check whether the answer moved.
That’s the part we build. Roman tracks the same engines — then its agents research, write, publish, and re-measure. We start where measurement stops.

Neither is universally better — they sit at different ends of the same market. Profound has the deepest data in the category (including Prompt Volumes, its consumer-panel demand data) and the strongest enterprise credentials, but it is sales-led and priced accordingly. Scrunch is faster to start, self-serve, and ships a genuinely unique delivery layer (AXP) that serves AI-readable versions of your pages to crawlers.
A reasonable shorthand: Fortune 500 brand with budget and procurement — Profound. Mid-market team that wants visibility data this week — Scrunch.
Profound does not publish self-serve pricing. Third-party reviews report an entry point around $499/month, with the features most teams actually want — multi-platform coverage, Prompt Volumes, API access — gated to custom-priced enterprise plans. Every purchase starts with a sales call, and there is no free trial.
Scrunch publishes its pricing: Starter at $250/month billed annually ($300 monthly) with 350 custom prompts and 3 personas, Growth at $417/month annually with 700 prompts and 5 personas, and custom Enterprise plans that add SAML/OIDC, the data API, and expanded limits. There is a 7-day free trial with no credit card required.
Yes. Sitecore announced its acquisition of Scrunch on June 3, 2026 — Bloomberg reported the price at roughly $225 million. Founder and CEO Chris Andrew is staying on to run the division, and Scrunch is being combined with the SitecoreAI digital experience platform. What that means for Scrunch as a standalone product is still an open question.
Mostly no. Both are measurement-first platforms. Profound launched an execution layer called Agents — template-based workers that can draft FAQ updates and content refreshes and publish to a CMS behind a human approval step — but third-party reviews still describe it as early. Scrunch does not create content at all: its AXP restructures pages you already published so AI crawlers can read them, and everything else is recommendations.
Closing the gap from "we are invisible for this prompt" to "a page now exists that wins it" is a different job — and it is the job Roman was built for.
AXP detects AI crawler traffic at the CDN layer (Akamai, Cloudflare, Vercel) and serves it a parallel, machine-readable version of each page — server-rendered, structured, and dramatically lighter in tokens. Human visitors and traditional search crawlers see your normal site.
It is Scrunch’s clearest differentiator, and also its most debated feature: competitors argue that serving different content to bots and humans carries cloaking risk. Scrunch maintains that traditional crawlers are unaffected.
Prompt Volumes estimates how often topics actually come up in AI conversations, built on what Profound describes as a panel of 400M+ real, double-opt-in user conversations. No competitor has an equivalent, which makes it Profound’s most cited advantage — though some practitioners question whether prompt-volume estimates can be independently verified.
The AI visibility category is crowded — Ahrefs Brand Radar, Semrush AI toolkit, Peec, Otterly, and others all track AI answers. The bigger question is what happens after measurement. If your bottleneck is not knowing where you stand, any strong tracker works. If your bottleneck is shipping the content that changes the answer, look at Roman — it tracks the same engines and then researches, writes, and publishes the pages that close the gaps it finds.
Researched June 2026 from tryprofound.com and scrunch.com (product and pricing pages), G2 reviews of both platforms, Fortune’s coverage of Profound’s Series C, and Sitecore’s acquisition announcement with Bloomberg’s reported deal terms. Both platforms ship fast — if something here is out of date, we’ll fix it.
See where AI recommends your competitors — then watch the content that changes the answer get researched, written, and shipped.