AEO Gets Its First Unicorn: Profound Raises $180M While Google Gives Half the Product Away Free

On 15 September 2026, New York-based Profound announced a US$180 million Series D at a US$1.8 billion post-money valuation, co-led by Sequoia Capital and Kleiner Perkins. That lands less than seven months after its US$96 million Series C, taking total funding past US$335 million. The official announcement also disclosed more than 1,000 paying enterprise customers, roughly a third of the Fortune 100 — Walmart, Comcast and Estée Lauder among them — and 3x revenue growth in six months. AEO (answer engine optimization), a category that did not exist two years ago, now has its first unicorn.

The timing is what makes this interesting. Over the past 18 months AI search stopped being an experiment and became part of the traffic structure: eMarketer forecasts that 31.3% of the US population will use generative AI search in 2026. That opened a very specific gap on the brand side — nobody knew how often they appeared inside ChatGPT, Gemini or Perplexity answers. The crawler model behind Ahrefs and Semrush is useless here, because LLM answers are not indexed, they are sampled. So a wave of startups filled the gap with a brute-force method: run thousands of prompts, count mention rates. The category pulled in over US$350 million in two years. Profound's round is happening now not because of a technical breakthrough, but because it got onto enterprise procurement lists first — and marketing budgets get re-shuffled once a year.

Line the competitors up and the category has already split three ways. Bluefish AI has raised roughly US$68 million, including a US$43 million Series B in April. Scrunch took the acquisition route and is now inside Sitecore. Evertune sits in the eight figures. AthenaHQ raised a US$2.2 million seed. Europe's Peec AI attacks from below at around €89/month. The 20-odd platforms tracked by Ad Age total about US$300 million, with two-thirds concentrated in nine enterprise platforms. This is over-funded at the top, acquired in the middle, price-warring at the bottom.

For SMBs and freelance shops, a US Series D sounds remote. It isn't: it decides whether you buy an AI visibility tool this year, and whether you can package this as a paid service. The hinge is that Google gave away half of what Profound sells at the end of August.

The detail: two stories that have to be read together

The first is Profound's US$180 million. The second happened three weeks earlier and almost nobody connected it to the round: Google's Search Console generative AI performance report finished rolling out to every site worldwide on 31 August 2026. Anyone with a verified property — including your three-person company — can now see, for free, how many times their pages were shown in AI Overviews, AI Mode and generative features in Discover, broken down by page, country, date and device.

That matters because the core pitch of tools like Profound is exactly "did you show up in the AI answer?" Google now ships first-party data for its own ecosystem, and it is server-side logging, not sampled estimation.

But Google held something back. The official documentation states the report contains no click data, does not expose the prompts users typed, and does not give stable positions. On 16 September Google added that average position methodology may change over time as AI Mode evolves. So the free half solves impression volume inside Google's ecosystem. It does not solve what happens inside ChatGPT and Perplexity, and it does not solve whether impressions turned into money. The US$1.8 billion valuation is a bet that those two stay unsolved.

What each reader should do today

Brand owners and SMB operators

• Open the generative AI performance report today and pull three months of impressions before you decide on any subscription.
• Push AI visibility tool budget back a quarter. What you lack is not measurement, it is citable content.
• If 90% of revenue is local, the priority order is Google > ChatGPT > everything else. Ignore "10 engines covered" dashboards.

Marketing and SEO practitioners

• Cross-reference the generative AI report against the standard Web search report to find pages with high AI impressions and low clicks — that is your content being consumed without return traffic.
• Build a list of 30 to 50 questions customers actually ask. It is the required input for both DIY monitoring and any paid tool.
• Stop tracking position. The methodology will change, and AI Mode mixes in link cards and carousels, so the metric is losing meaning.

Developers and agencies

• Package an "AI visibility baseline report" as a one-off project rather than reselling subscriptions: multi-engine sampling, a report, a fix list.
• Fill in FAQPage, Organization and Product schema — currently the only technical lever that directly affects being extracted as an answer.
• Write a scheduled script that pushes Search Console API impression data into the client's Looker Studio every week.

SaaS comparison

Prices per published 2026 AEO platform pricing. Entry tiers almost always cover one engine and a small prompt quota; genuinely usable plans start above US$300/month.

What they won't tell you

• The numbers are not reproducible. LLMs are non-deterministic; the same input returns different results across runs. Brainlabs states it plainly: none of the visibility data from Profound, Peec and the rest is fully accurate. When a competitor's score jumps 15 points you cannot tell whether citations really increased, the tool changed its prompt set, or sampling caught a favourable moment.
• "Prompt volume" is invented. The prompt volume on every dashboard is Google keyword volume multiplied by an assumed conversion factor. No LLM provider publishes real prompt-level query counts.
• Paying still does not buy click data. Neither ChatGPT nor Perplexity exposes citation clicks to third parties. Paid tools tell you whether you were mentioned, not whether it produced business.

The no-subscription SMB alternative

A weekend to build, half an hour a month to maintain:

• ☐ Enable the generative AI performance report and export 90 days of impressions.
• ☐ List 30 real customer questions, including local phrasing.
• ☐ Write a Python script that hits the OpenAI and Perplexity APIs weekly, matches responses for your brand name and domain, and writes to a Google Sheet. API cost is roughly US$3 to US$8 a month.
• ☐ Have the same script log other brands appearing in the answers — a free competitor citation list.
• ☐ Add FAQPage and Organization structured data, and open every article with a 40 to 80 word direct answer.
• ☐ Check crawl frequency for GPTBot, PerplexityBot and ClaudeBot in server logs to confirm your CDN is not blocking them.
• ☐ Push impressions into Looker Studio monthly via the Search Console API.

Coverage is about 70% of a paid tool at 3% of the cost. You lose historical baselines, multi-seat permissions and SOC 2 — none of which matter unless your client is publicly listed.

FAQ

Is the Search Console generative AI report available outside the US?

Yes. It completed worldwide rollout on 31 August 2026. Any verified, indexed site sees it under Performance, broken down by page, country, date and device — but without click data.

Do I still need Profound or Peec AI?

If almost all traffic comes from Google's ecosystem, the free report plus a DIY script covers around 70% of the need. Paid tools pay off only when you need cross-engine comparison across ChatGPT, Claude and Perplexity with a historical baseline for formal reporting.

Can I report these numbers to clients as a KPI?

Report trends, not absolutes. Sampling is non-deterministic and 10 to 20% month-to-month movement is normal. Use a three-month moving average and write the data source and sampling method into the contract.

What can an agency charge for this?

Position it as a one-off baseline report plus quarterly tracking, not subscription resale. A one-off report covering 30 to 50 prompts, multi-engine sampling, a schema audit and a fix list is a defensible fixed-fee project.

My take

The consensus read is "the AEO category is validated, get on board." My judgement is the opposite: the measurement half of this category gets commoditised to free within 12 to 18 months, and the valuation multiple will not survive it. Google has already moved, and it ships log-grade truth rather than samples. When OpenAI has to build a business-facing console to make advertising work, the second dataset opens up too. A business selling sampled estimates of someone else's platform has a moat made of data opacity, and data opacity is never a stable state. Profound clearly knows: this round's language is "the AI platform for marketing teams," not "AI visibility monitoring" — it is escaping the dashboard toward the action layer.

The implication for an agency like ScriptWalker is direct: do not resell subscriptions, sell a one-off AI visibility audit plus schema implementation. Monitoring gets eaten by free tools, but fixing structured data, rewriting answer-first paragraphs and unblocking wrongly-filtered AI crawlers is engineering work a dashboard cannot replace.

Sources

• Primary | Profound press release: Raises $180M Series D at $1.8B Valuation (2026-09-15)
• Primary | Google Search Central Blog: Introducing Search Generative AI performance reports
• Primary | Search Console Help: Generative AI performance report (Search)
• Primary | Bloomberg: Profound Hits $1.8 Billion Value to Boost Brands in AI Search
• Secondary | TechCrunch: AEO startup Profound hits unicorn valuation
• Secondary | Search Engine Land: GSC AI performance reports rolling out globally
• Secondary | Search Engine Roundtable: GSC AI Reporting Will Change Over Time
• Secondary | Brainlabs: Your AI Visibility Data Is Wrong
• Secondary | AEO Platform Pricing 2026: Profound, Peec, Otterly side-by-side
• Secondary | eMarketer: FAQ on GEO and AEO
• Secondary | Ad Age: GEO and AEO startups that marketers need to know

Contact

• Email: [email protected]
• Phone: 0916-224-047
• LINE: @ufv9089p