Who Is Buying and Building AI Visibility: An Acquisition Map for the GEO Category

The AI visibility category went from academic paper to consolidation wave in under three years. The term GEO was coined in a 2023 Princeton-led paper. By September 2026, a public software giant had spent nearly two billion dollars to own the category. A marketing platform had acquired a startup less than a year old and shipped a product within six months. Four separate website platforms had built AI visibility natively. And pure-play startups had raised well over $255 million in venture capital, with Sequoia, Lightspeed, Kleiner Perkins, and Salesforce Ventures on the cap tables.

I track this category on GEO Compass, and the two questions I get most from operators and corp-dev teams are the same ones every quarter. Who moves next, and should a large platform buy an AI visibility company or build one?

This is my attempt at a serious answer to both. The short version of my thesis, stated up front so you can disagree with it early:

For most large platforms, buying beats building, because the difficult part of AI visibility is not the dashboard. It is the measurement infrastructure that turns nondeterministic model outputs into stable, trustworthy signals, plus the per-engine data relationships that have to be maintained as models change without notice. That takes roughly eighteen months and a specific kind of talent that most acquirers do not have in house. Every company that built a credible product from scratch already owned an adjacent data position. Nobody has built a competitive AI visibility product from a standing start.

Here is the evidence for that claim, then the framework, then the predictions.

Part 1: The scoreboard

Before the analysis, the map. As of September 2026, the large-platform response to AI visibility falls into four buckets.

Acquired: Adobe (Semrush, $1.9B), HubSpot (XFunnel), Sitecore (Scrunch).

Built natively: Wix (AI Visibility Overview), Squarespace (AI Visibility tool), Cloudflare (AEO Visibility Dashboard), Optimizely (GEO-ready CMS plus Agent Visibility Analytics).

Partnered: Optimizely again, with Conductor for the enterprise measurement intelligence, which makes it the only major player to run a hybrid strategy.

Shipped nothing native: Contentful, Sanity, Drupal, and, notably, Salesforce. Shopify took a commerce route instead: its Agentic Storefronts list merchants' products in ChatGPT, Google AI Mode and Gemini, Microsoft Copilot and Meta, and attribute the resulting orders, but they do not measure brand visibility or citations in AI answers. Webflow has announced Webflow AEO for Enterprise, but it launched as a private beta.

That last bucket is where the next eighteen months of M&A comes from.

Part 2: The acquisitions, and the logic behind each

Adobe and Semrush: $1.9 billion for a measurement layer

The anchor deal. Adobe announced an all-cash acquisition of Semrush at $12.00 per share on November 19, 2025, representing roughly $1.9 billion in total equity value. It closed the deal on April 28, 2026, after US antitrust clearance and stockholder approval. The price was a 77.5% premium to Semrush's last closing price before the announcement, per the merger proxy. Adobe secured voting commitments from Semrush founders and major shareholders representing more than 75% of voting power before announcing, which tells you the deal was wired before it was public.

Adobe's framing was explicit. Anil Chakravarthy, president of Adobe's Digital Experience business, said brand visibility is being reshaped by generative AI and positioned GEO as a new growth channel alongside SEO. At close, Adobe folded Semrush's roughly 28 million users into Adobe CX Enterprise.

The build-versus-buy logic here is the cleanest in the category. Adobe already owned content creation (Creative Cloud), content orchestration (Experience Manager), and analytics. What it did not own was a measurement layer for how brands appear across the open web and inside LLMs. Semrush brought 28.8 billion keywords and 43 trillion backlinks, plus an existing Semrush AI Visibility Toolkit. It also brought enterprise ARR growing 33% year over year, with customers including Amazon, JPMorganChase, and TikTok.

Could Adobe have built that? The dashboard, yes. The data asset, no, not on any timeline that mattered. Web-scale crawl and keyword infrastructure takes a decade to accumulate, and Adobe was watching Salesforce and Microsoft look at the same territory. The 77.5% premium was the price of not spending three years catching up.

HubSpot and XFunnel: speed as the entire thesis

The most instructive deal for anyone thinking about timelines. HubSpot announced it was acquiring XFunnel on October 31, 2025, and closed on December 1, 2025. XFunnel was an Israeli AEO startup founded by Beeri Amiel and Neri Bluman that was less than a year old at acquisition.

HubSpot then shipped HubSpot AEO at its Spring 2026 Spotlight, generally available April 14, 2026. The product tracks brand visibility across ChatGPT, Gemini, and Perplexity, shows share of voice against competitors, analyzes citations, and provides a brand visibility dashboard, all inside Marketing Hub. Independent reviewers noted that HubSpot AEO is essentially XFunnel rebuilt inside the HubSpot interface. That explains how a company that size shipped a full AI visibility dashboard within six months of the deal closing.

The motive was defensive and urgent. HubSpot's customer base was watching organic traffic decline in real time, and HubSpot's own data showed AI-driven leads converting roughly three times better than traditional search. HubSpot has a house pattern of catching channel shifts early, starting with inbound in 2006, and its own marketing leans on that history. The question internally was never whether to ship AEO. It was how fast.

Note the asymmetry with Adobe. Adobe bought scale and a data moat. HubSpot bought a team and a working product to compress its time to market from eighteen months to six. Both are valid buy rationales, and they imply very different target profiles.

Part 3: The builders, and the adjacent-data pattern

Not everyone bought. Four platforms built AI visibility natively, and the pattern in who succeeded is the most useful finding in this research: every successful builder already owned an adjacent data position that made building cheaper than buying.

Wix: the platform-owner build

Wix launched AI Visibility Overview in July 2025 as part of a broader GEO initiative, accessible from Wix Analytics. It generates likely customer prompts, checks whether the site is mentioned in the answers, and returns an AI visibility score. It supports ChatGPT, Gemini, and Perplexity (Claude was also covered at launch), compares against competitors, and analyzes cited sources. It also shows both bot crawl activity and actual referral visits from each LLM. Wix positioned itself as the first CMS to ship this natively.

Wix could build because it already held the content, the hosting, and the traffic data for millions of sites. The visibility layer was an extension of data it owned rather than a new business. Worth noting the limitations independent users report, including query prompts auto-regenerating every 180 days, which is the kind of rough edge that shows up when visibility is a feature rather than the product.

Squarespace: the same logic, smaller surface

Squarespace added an AI Visibility tool to its SEO panel that tests how often a site is mentioned in ChatGPT and Gemini across branded and non-branded prompts. It runs on monthly AI credits that vary by plan, and crawler controls sit in a separate setting. Same rationale as Wix: own the platform, so the feature is an add-on.

Optimizely: the hybrid, and the most revealing strategy

Optimizely ran the most aggressive DXP playbook and, importantly, split it. In July 2025 it announced GEO-ready CMS features covering page optimization, GEO topic templates, bulk metadata population, and GEO analytics, positioning the CMS as a visibility engine rather than a content warehouse. Then on June 10, 2026, it launched a full AEO platform. It paired Agent Visibility Analytics, built inside Optimizely Analytics on log-level data classifying AI agent requests by intent (retrieval, indexing, training), with a partnership with Conductor for enterprise SEO, GEO, and AEO intelligence.

Read that split carefully, because it is the tell of the whole report. Optimizely built the parts adjacent to data it owned, its own CMS structure and its own server logs. For the hard part, measuring what the engines actually say across the market, it partnered with a specialist rather than building. A company with Optimizely's engineering resources concluding that measurement intelligence was worth licensing is the strongest available evidence that the measurement layer is the genuinely difficult asset.

Cloudflare: building from the most privileged position in the category

The most strategically interesting build. Cloudflare launched its AEO Visibility Dashboard in early access on August 6, 2026, measuring whether AI assistants mention or cite a brand, how prominently it appears, and share of voice against competitors.

Cloudflare could build this because it sits in a position no GEO startup can replicate. It sits in front of about 20% of the web and fingerprints AI bots using machine learning, behavioral analysis, and traffic-level visibility. It spent two years building the control side first. AI Crawl Control (formerly AI Audit, which launched September 2024) reached general availability in August 2025, and lets site owners see every AI bot and allow, block, or charge it per crawler. Pay Per Crawl revived the HTTP 402 status code to make crawler access machine-readably negotiable. On July 1, 2025, Cloudflare flipped the default for domains newly joining Cloudflare to block AI crawlers unless the operator opts in.

By July 2026 Cloudflare had publicly acknowledged the other half of the problem, that blocking everything could make smaller sites invisible, which it framed as a Faustian bargain between AI access and disappearing from emerging discovery channels. The visibility dashboard is the resolution. Because Cloudflare sits between websites and incoming traffic, it can separate two questions no one else can cleanly split: can AI systems access this site, and do their assistants actually recommend this brand.

Cloudflare is not really building a GEO tool. It is building the control panel for a site's entire relationship with AI: who gets in, who pays, who crawls, and whether any of it results in the brand being surfaced. That is an infrastructure position, and it is why Cloudflare is the one builder I would not bet against.

Part 4: The infrastructure players setting the rules

Two giants are shaping the substrate GEO operates on rather than selling GEO products, which matters enormously for predicting the next moves.

Microsoft is quietly building the measurement rails. Bing Webmaster Tools now includes an AI Performance dashboard giving brands visibility into which of their URLs are cited in AI-generated answers and, more valuably, the grounding queries behind those citations. Microsoft Advertising separately published an AEO and GEO playbook in January 2026 explaining how products disappear from AI recommendations. It reframed optimization around data completeness, currency, structured markup, and contextual richness, and described the shift from traffic to influence.

The strategic point: Microsoft's Bing grounding powers Copilot and other AI assistants, which means Microsoft can see citation patterns almost nobody else can observe directly. It is sitting on the raw material for a first-party measurement product and has so far chosen to distribute it through webmaster tools and advertiser guidance rather than package it.

Amazon is building an answer engine, not a GEO tool. Rufus, Amazon's shopping assistant, is becoming its own product-discovery surface with its own retrieval behavior. Most GEO platforms do not yet track it; in current engine-coverage matrices, Rufus support is the exception rather than the rule. For any brand selling physical product, Amazon is becoming a sixth or seventh engine that the measurement category has barely instrumented.

Google is arguing the category does not need a name. Google's 2026 Search Central guidance states that optimizing for AI search is optimizing for the search experience and thus still SEO. It also debunks several vendor tactics, saying structured data is not required for generative AI features and that llms.txt does nothing for Google Search. Google's position is coherent for Google's own surfaces and self-interested everywhere else, which I have written about separately.

Part 5: The target landscape

If you are modeling acquisitions, you need to know what is acquirable. The pure-play GEO market has absorbed serious capital in a short window.

By March 2026, one count put venture capital into pure-play GEO startups at $255 million or more in roughly eighteen months, with Profound alone accounting for about $155 million of it. Profound has since raised another $180 million. Other analyses put the broader tools market at $300 million-plus raised between mid-2025 and spring 2026. The investor list is not speculative seed money: Sequoia, Lightspeed, Kleiner Perkins, NEA, Khosla, Felicis, Bloomberg Beta, Y Combinator, Antler, and Salesforce Ventures.

That last name deserves a pause. Salesforce Ventures already has exposure to this category. Corporate venture arms frequently function as acquisition reconnaissance, and Salesforce is simultaneously the largest platform with no native AI visibility product.

The landscape sorts into tiers.

The category leader. Profound raised a $96 million Series C at a $1 billion valuation in February 2026, led by Lightspeed with Sequoia, Kleiner Perkins, and South Park Commons participating, making it the category's first unicorn. In September 2026 it raised a $180 million Series D at a $1.8 billion valuation, co-led by Sequoia and Kleiner Perkins, taking its total past $335 million. It reports serving 16% of the Fortune 500, tracks eight-plus engines, and holds SOC 2 Type II certification. Profound is expensive enough that it is now more likely to be an acquirer or an IPO candidate than a target.

The well-funded enterprise tier. Bluefish AI, Evertune (positioned at the intersection of GEO and programmatic advertising), and AthenaHQ, which punches above its funding weight with a CEO out of Google Search and DeepMind. Scrunch AI (about $26 million raised, with an Agent Experience Platform for AI-native content delivery) sat in this tier until Sitecore acquired it in June 2026. These are the most plausible acquisition targets: real technology, enterprise customers, not yet priced out of reach.

The fast-growing mid-market. Peec AI is the fastest-growing by customer acquisition, reportedly adding hundreds of customers per month and passing 1,300 customers within ten months, with named European enterprise logos. Goodie, Otterly.AI, Daydream, and others fill out this tier. High growth, lower price, attractive to acquirers who want a product and a user base rather than a data moat.

The bolt-ons. Ahrefs Brand Radar, Semrush AI Optimization (now Adobe), Conductor, and seoClarity represent legacy SEO platforms adding AI visibility modules. These have the most users and, by most independent assessments, the shallowest engine coverage relative to pure-plays. They are the distribution play, not the technology play.

The structural point: the middle of this market is going to get squeezed. A generalist point tool competing against a unicorn on one side and bundled suite features on the other has no defensible wedge. That is exactly the dynamic that ended Lorelight, a GEO startup whose founder shut it down and concluded there is no such thing as a GEO strategy separate from brand building. Vertical depth and platform breadth survive. The undifferentiated middle does not.

Part 6: How big is this market, honestly

Market sizing for a three-year-old category deserves skepticism, and I want to show you why rather than quote one convenient number.

Published estimates for 2025 and 2026 range from roughly $390 million to $2.7 billion depending on the research firm and how the category is defined. Forecasts for the early 2030s range from about $4.25 billion to $26.85 billion. Compound growth rates cited fall between 13.6% and 50.5%, though the reports at both extremes do not reconcile with their own figures.

That spread is not a rounding difference. It is nearly a factor of seven on current size and a factor of six on the forecast. When analyst estimates diverge that widely, it usually means three things. The category boundary is unsettled (does GEO include services, platforms, or both?). The revenue base is small enough that a few enterprise contracts move the percentages. And the forecasters are extrapolating from under two years of data.

What I would actually rely on instead of the market-size range is the observable behavior. $255 million-plus of venture capital deployed, a $1.9 billion acquisition closed, and a unicorn minted. Enterprise job postings with salary bands up to $171,000. And Conductor's survey of 250-plus enterprise digital leaders, including CMOs, finding 94% planning to increase AEO investment in 2026 and 97% reporting a positive impact in 2025. Behavior beats forecasts when the forecasts disagree by 6x.

Part 7: The build-versus-buy framework

If you lead corp dev or product at a large platform, here is the decision framework I would actually use.

What you are really acquiring

The instinct is to evaluate the dashboard. The dashboard is the easy 20% and any competent team can ship one in a quarter. The real asset is four things underneath it.

First, the query infrastructure. Credible measurement means running hundreds to hundreds of thousands of prompts daily, per engine, and normalizing nondeterministic outputs into stable signals. That is an operational cost structure and an engineering discipline, not a feature.

Second, per-engine parsing that survives model churn. Profound's data shows 40 to 60% of cited domains changing within a month, and models update without notice. The measurement pipeline is a living system requiring continuous maintenance. This is the single most underestimated cost in build scenarios, because it never ends.

Third, the statistical methodology. Rand Fishkin's SparkToro research found fewer than 1 in 100 repeated prompt runs returned the same set of brands, and concluded that aggregate brand presence is nonetheless measurable across large samples. Building a product that is honest about that distinction, reporting distributions rather than fake precision, requires people who understand probabilistic measurement. That talent is scarce and concentrated in the existing startups.

Fourth, enterprise trust artifacts. SOC 2 Type II, data retention policies, per-client access controls. Unglamorous, slow to build, and required to sell to the Fortune 500.

The decision rules

Buy when speed matters and you lack adjacent data. If your competitors are shipping and you are starting from zero, an eighteen-month build cycle is a losing position. Adobe and HubSpot both reached this conclusion, at very different price points, for very different reasons. HubSpot's version is the more replicable template for most acquirers: buy a young, capable team with a working product, rebuild it inside your interface, ship in six months.

Build when you already own adjacent data. Wix, Squarespace, and Cloudflare built because the marginal cost of a visibility feature on top of data they already held was low. The test is specific: do you already own site content, hosting logs, crawler traffic, or grounding data at scale? If yes, build. If no, you are starting from zero against companies with an eighteen-month head start and better talent density.

Partner when you need depth fast but want to keep the interface. Optimizely's Conductor partnership is the underrated third path. You own the customer relationship and the workflow, you license the measurement intelligence, and you avoid the acquisition premium and the integration risk. Expect substantially more of this, because it is the rational move for any platform that wants the capability without betting the balance sheet.

Do not build a dashboard and call it a product. The failure mode I would flag hardest: shipping a blended AI visibility score computed from a handful of prompt runs. It is cheap, it demos well, and it is measuring noise. Customers who compare it against a real platform will discover the difference. The reputational cost of shipping fake precision in a category already fighting credibility problems is worse than shipping nothing.

Part 8: Who moves next

This section is analysis, not reporting. These are predictions with reasoning attached, and you should weigh them accordingly.

Salesforce is the most exposed company in enterprise software on this issue. It owns Marketing Cloud and Data Cloud and has no native AI visibility measurement. Adobe's Semrush acquisition directly threatens Salesforce's position with enterprise marketing teams, because Adobe can now offer creation, orchestration, and visibility measurement in one stack. Salesforce Ventures already has exposure to the category, which typically precedes acquisition interest. Salesforce has no adjacent web-scale crawl data to build from, which by my own framework means it should buy. If I had to name the single most likely acquisition in the next twelve months, it is Salesforce acquiring an enterprise-tier pure-play.

Microsoft has the strongest build case and may still buy. It already sees grounding and citation data through Bing. The gap is the marketing-facing product layer and the enterprise workflow, which is exactly the part that is faster to acquire than build. A Microsoft acquisition would be about time to market and go-to-market motion, not data.

Amazon will build rather than buy, and it will be about Rufus. Amazon's culture defaults to building, it owns the answer surface, and the natural product is seller-facing product visibility measurement inside Seller Central. The interesting question is whether Amazon exposes that data externally at all, since keeping it internal is also a competitive strategy.

The CMS and DXP laggards face a forced choice. Contentful, Sanity, and Drupal currently ship nothing native, while Sitecore has already bought Scrunch and Webflow has an AEO product in beta. As Wix, Squarespace, and Optimizely make AI visibility table stakes, these platforms have to respond. Most lack the traffic-data scale that made building cheap for Wix. Sitecore will not be the last. I expect at least one more acquisition of a mid-market GEO tool from this group, quite possibly several, alongside more Conductor-style partnerships, which are cheaper and faster.

Agencies and holding companies keep consolidating. Publicis has already posted a VP-level role for performance content leadership spanning SEO, GEO, and AEO. Holding companies buy capabilities they are being asked for in RFPs, and they are being asked.

The wildcard: private equity roll-ups. With twenty-plus funded startups, undifferentiated middle-tier products, and a squeeze coming, the conditions for a PE roll-up of mid-market GEO tools exist by 2027. That is how overcrowded tooling categories usually resolve.

Part 9: What to look for when you build or buy

Whether you are writing an acquisition thesis or a product spec, this is the rubric. It works in both directions: these are the criteria that separate a credible product from a dashboard, which makes them the same criteria that separate an asset from a feature.

Probabilistic measurement at scale, not a single rank. AI answers are nondeterministic. Any product reporting one AI rank from a handful of runs is selling fiction. The credible approach runs large daily prompt samples per engine and reports citation frequency and share of voice as distributions over time. Ask the target how many prompts per engine per day and what the variance looks like.

Genuine per-engine breakdowns. ChatGPT, Perplexity, Gemini, Claude, Copilot, and Google AI Overviews retrieve differently and trust different sources. Google's AI Mode shares only about a third of its cited URLs with its own organic top ten, by Semrush's measurement. A blended score describes none of them. Check coverage of emerging surfaces too, especially Amazon Rufus for commerce and Grok for the engines most tools skip.

Source and citation analysis, not just scores. Knowing your visibility number is table stakes. Knowing which domains each engine cites when it recommends your competitor is the actionable layer, because it converts "build authority" into a specific work queue of where to earn presence.

Passage-level and entity intelligence. The unit of optimization moved from the page to the passage and from the keyword to the entity. A serious product understands extraction and entity corroboration, not just whether a URL ranks. This is grounded in how retrieval architectures actually work, which I have written about in detail.

Attribution to pipeline. Programs that survive budget review connect visibility to AI-referred traffic, lead quality, and revenue. AI search visitors are widely reported to convert at multiples of traditional organic because they arrive pre-educated. One Semrush analysis of 500-plus topics put it at 4.4x by conversion rate. A product that cannot tell that story is a vanity metric with extra steps.

Resilience to churn, treated as a product requirement. With 40 to 60% of cited domains rotating monthly, the pipeline must be maintained continuously. In diligence, ask how many engineers maintain parsers and how the team handled the last three major model updates. The answer separates real infrastructure from a wrapper.

Enterprise trust artifacts. SOC 2 Type II, data retention controls, role-based access, multi-seat governance. Slow to build, required to sell upmarket, and a legitimate reason to buy rather than build.

Honest limitations. The best products in this category are explicit about what they cannot see, including that no external tool observes inside private AI conversations. Cloudflare, despite its position, still cannot see inside AI conversations. A vendor claiming otherwise is a vendor to discount, and a target making that claim in a pitch deck is a diligence flag.

What this means if you are not buying anything

Most readers here are not running corp dev. The reason this map matters to an operator is simpler than the M&A narrative.

Consider the $1.9 billion acquisition, the six-month product sprint, and the infrastructure company reframing its crawler data. Every one of these moves is a large organization concluding that the answers AI engines give about brands are commercially decisive and worth owning infrastructure for. Those answers are being formed about your category right now, and they harden over time as engines settle into consistent recommendation patterns.

My data at GrackerAI (disclosure: my company, a GEO platform for B2B SaaS) shows 4 in 10 B2B security buyers now start vendor research inside AI assistants. Whether the giant serving your category builds, buys, or partners changes which vendor you eventually write a check to. It does not change whether your brand is in the answer.

Here is my question for the operators and corp-dev readers. If your company had to own AI visibility within twelve months, would you build on adjacent data you already have, partner for the measurement depth, or buy outright? And what single asset would make the premium worth paying? I keep the running version of this map on GEO Compass and would like to hear where you land, especially if you think I have the Salesforce call wrong.

Frequently Asked Questions

Which large companies have acquired AI visibility or GEO platforms?

Adobe acquired Semrush for approximately $1.9 billion at $12.00 per share, announced November 19, 2025 and closed April 28, 2026, folding roughly 28 million users into Adobe CX Enterprise. HubSpot acquired XFunnel, an Israeli AEO startup under a year old, announced October 31, 2025 and closed December 1, 2025, and shipped HubSpot AEO in April 2026. Sitecore acquired Scrunch in June 2026.

Which companies built AI visibility tools instead of acquiring one?

Wix built AI Visibility Overview into Wix Analytics, and Squarespace added an AI Visibility tool to its SEO panel. Optimizely shipped GEO-ready CMS features plus Agent Visibility Analytics, and Cloudflare launched an AEO Visibility Dashboard in early access on August 6, 2026. Each already owned an adjacent data position: site content and hosting for Wix and Squarespace, CMS and server logs for Optimizely, and about 20% of the web plus AI crawler fingerprinting for Cloudflare.

Should a company build or buy an AI visibility product?

Buy when speed matters and you lack adjacent data. The difficult asset is the measurement infrastructure, per-engine parsing that survives model churn, probabilistic methodology, and enterprise trust artifacts, which together take roughly eighteen months to build. Build when you already own adjacent data such as site content, hosting logs, or crawler traffic. Partner when you need measurement depth fast but want to keep the customer-facing interface, as Optimizely did with Conductor.

How much venture funding has gone into GEO startups?

By one count, at least $255 million in venture capital flowed into pure-play GEO startups in the roughly eighteen months to March 2026, with Profound accounting for about $155 million of that. Profound raised a $96 million Series C at a $1 billion valuation in February 2026, led by Lightspeed with Sequoia and Kleiner Perkins participating. It then raised a $180 million Series D at a $1.8 billion valuation in September 2026. The broader investor list includes NEA, Khosla, Felicis, Bloomberg Beta, Y Combinator, and Salesforce Ventures.

Which companies are most likely to acquire a GEO platform next?

Salesforce is the most exposed, owning Marketing Cloud and Data Cloud with no native AI visibility product while Adobe's Semrush deal threatens its enterprise position, and Salesforce Ventures already has category exposure. Microsoft could acquire the marketing-facing product layer to complement its Bing grounding and citation data. CMS and DXP platforms shipping nothing native, including Contentful, Sanity, and Drupal, face a forced build, buy, or partner decision as the capability becomes table stakes.

How big is the GEO market?

Published estimates for 2025 and 2026 range from roughly $390 million to $2.7 billion, with early-2030s forecasts between about $4.25 billion and $26.85 billion and cited growth rates from 13.6% to 50.5%. That spread reflects unsettled category boundaries and under two years of data. Observable behavior is more reliable: over $255 million in venture funding, a closed $1.9 billion acquisition, and Conductor's survey of 250-plus enterprise digital leaders finding 94% increasing AEO investment in 2026.