From Clicks to Signals: How GenAI, Privacy, and Identity Are Rewriting B2B Marketing

How AI-powered search, privacy regulation, and identity shifts are forcing B2B marketers to architect outside the funnel and embrace a signal-driven future

The ground under B2B marketing is shifting

If the last decade was about mastering consumer-grade adtech for B2B, the next five years are about tearing up that playbook.

Generative AI is hollowing out the traditional search economy and everything which comes downwind of search (like content website visits); privacy regulation is dismantling the cookie spine of digital targeting; identity is fragmenting into competing ecosystems; agencies are using AI to flatten the technical delivery gap with vendors in a margin grab power move; the very fabric of programmatic is under threat from batching protocols. Change is afoot…

This isn’t a budget tweak. It’s a structural rewrite of how demand is created, captured, and measured. We need to design for ambient discovery, signal-layered targeting, and product-led activation - not the old linear, keyword-led funnel. [1]

1) GenAI is collapsing—and rewriting—the discovery path

Well-respected forecasts suggest that traditional search volume could decline by around 25% by mid-decade, but the speed of change in 2025 suggests even that might be conservative. I suspect it could be significantly out.

In January, LLM-powered search engines—think ChatGPT’s browsing mode, Perplexity, and Claude.ai—accounted for roughly 1.5% of all search-like queries. By June, that share had passed 10%, a near order-of-magnitude jump in less than half a year.

This isn’t a slow migration; it’s a behavioural landslide.

Younger cohorts already bypass Google entirely for research questions, starting instead inside conversational AI assistants. And Google is not resisting - it’s accelerating the shift. A giant business textbook game of innovators dilemma is playing out around Google as we watch. Until recently I thought they were hedging, but the latest releases mean their path is clear. Gen AI is the future of search across both Google and their new, and up until 2022 entirely unforeseen, competitors.

On Google mobile, the new AI tab in search is one big generative answer, often showing only a handful of organic links beneath, and already testing in-answer ads that will soon be sold the same way as sponsored search slots.

I do think we’re moving to a search experience which will not be democratising web traffic anything like the old version did.

For B2B, this is not just an SEO or SEM challenge, it’s a demand architecture shift:

• Discovery happens inside the AI layer before a buyer ever reaches a web page.

• The answer is the destination - organic click-through is a secondary outcome, not the default. Where previously discovery happened post click - now it happens pre-click — and often there won’t BE a reason to click. We see this in the collapse of referral traffic

• Attribution blurs as the research process fragments across multiple AI surfaces.

The most profound change? LLMs are acting as research assistants. They synthesise vendor content, analyst commentary, textbooks, premium B2B content, news, product specs, and user reviews (G2 etc) into a consolidated narrative, stripping away your controlled brand funnel and replacing it with a machine-curated one.

That has two major effects in B2B:

1. Vendor traffic disruption – The "long tail" of content that once pulled in niche research queries gets consumed, paraphrased, and presented without a click. This makes capturing net-new discovery harder and puts pressure on owned content to be both ingestible and influential in AI training corpora. Previously long form blog content had value in capturing early research, specially when paired with corporate traffic resolution layers - but now that content is more likely - if you’re lucky - to seed part of a Gen Ai answer.

2. Publisher disintermediation – Industry trade sites and niche B2B media lose direct traffic as AI assistants answer questions they used to rank for. Their monetisation models, already thin, will face existential strain.

Where does this net out?

• In the near term: Expect branded presence in AI answers to become a paid product - Google’s in-answer ads and OpenAI’s “sponsored citation” pilots are early signals. Winning here will require both content schema optimisation for AI ingestion and budget for premium placement inside AI summaries. This essentially becomes SEM 2.0 - however we used to think that one, two, three character search queries had a hell of a bell curve where there was actually very little volume for all but the biggest one or two word queries - I dread to think the impact on prompt based targeting and ad alignment. This all feels like it will be AI ad targeting AI generated content and brands left to sit back and hope the machines can nail their user journeys

• In the medium term: B2B publishers may fight back with data walls, holding premium content behind login or API paywalls that can be licensed rather than scraped. Some will partner directly with AI providers to secure revenue shares. Could an era of micro payments linked to single article reads come into its own? Apple News is growing fast with such models, and B2B content which oscillates in interest around buying cycles for corporate readers is a hard sell for a hard paywall in many cases. I do wonder if pay to play or in this case read/watch/listen and micropayment models could swing into focus

• In the long term: Expect a bifurcation - public-facing marketing content will be optimised for AI consumption (concise, structured, high-authority), while deep, high-value insight moves into gated ecosystems, private communities, and account-personalised experiences where LLMs can’t fully replace the interaction.

If your ABM model still assumes the majority of buying-committee research begins in Google and flows into your content hub, you’re already behind. The future play is to:

• Engineer for AI ingestion (structured content, schema markup, Q&A frameworks).

• Invest in direct relationships that bypass the open web entirely (community, events, owned data) which will work in cohort with dark funnels and gated [to Gen AI] environments like industry Slack channels, Whatsapp etc - all of which carry disproportionate influence to scale (low volumes but very influential, and equally very hard to penetrate)

• Prepare for sponsored AI responses and conversational ad formats to become key demand levers.

• Diversify discovery to platforms where AI is an input, not the sole gatekeeper—podcasts, video, in-product thought leadership.

The bottom line: In the signals era, visibility inside the AI layer is as strategically critical as ranking on page one was in 2015 and the brands that adapt fastest will own disproportionate share-of-voice in the new research funnel.

2) Privacy isn’t a constraint; it’s a competitive moat

Building a platform like FunnelFuel in the EU means fighting the privacy battle early. I would not have liked to build this business in the US, and then being tasked with retro fitting privacy by design into a system that would be attacking a smaller, more fragmented, more challenging set of EU markets. By fortune, doing it our way around let us lead on these principals, and it makes entering other markets which are moving in this direction, much lower stress.

That’s therefore an advantage. We’ve had to design for the world’s strictest regulatory environment from day one, rather than bolt on “compliance” later. And the truth is, there’s no magic cookie replacement coming to save lazy architectures.

Universal IDs promised to be that magic bullet, but the cracks are showing. ID5 has already pivoted its positioning, shifting from “universal cookie alternative” to a broader identity infrastructure narrative to account for scale constraints. UID2.0—the most talked-about in the U.S. faces friction in categories like healthcare and finance, where opt-in rates are vanishingly low and that has put hard analysis on their whole framework. Wherever these identifiers require true, informed consent, their usable audience size drops sharply, often to single-digit percentages of total reach.

This creates a structural reality: the identity layer is getting thinner.

But that thinness isn’t all downside, it’s where stronger signals emerge. When a user does click through from an LLM-generated answer into your owned property, that action carries far more intent weight than a generic organic search click ever did. It’s deliberate, not casual. The same is true for first-party data, especially first-party intent tracked through analytics platforms: fewer records, but each one a richer seed for activation. Publishers and vendors analytics have been full of noise, this shift is changing volume into richer layers of deeper insights

The by-product of privacy-by-design is this:

• Smaller, better seeds: lower volume, higher signal strength.

• World-class capture: you get fewer opportunities to see these signals, so the instrumentation has to be flawless.

• Smarter omnichannel use: one strong seed, activated across display, social, CTV, and email, can outperform a thousand low-confidence IDs.

• Probabilistic comfort: the old identity layer gave a false sense of certainty; the modern play is to work with confidence scoring, corroboration, and pattern recognition.

Between GDPR, ePrivacy, Apple ATT, and Chrome’s evolving move from blanket third-party cookie deprecation to user choice prompts, any model dependent on cross-site personal identifiers is a dead end—especially in EMEA. High performers are already building around composite, consented signals: context, firmographics, on-site engagement, geo/network quality, and the dynamic consent flows that turn trust into a strategic asset.

Treat privacy not as an obstacle, but as an architectural requirement—a forcing function that makes your targeting sharper, your measurement more honest, and your long-term moat deeper.

3) Identity is now multi-modal—and match rates are sliding

The cookie was crude but predictable. You could bank on a certain percentage of your target list being visible in the wild. That’s gone. In 2025, match rates are structurally lower, and pretending otherwise is a recipe for under-performance.

Why?

• Discovery happens in gated AI environments – When research is done inside ChatGPT, Perplexity, or private Slack channels, no pixel ever fires, no ID is dropped, and no “first touch” gets recorded.

• Hybrid work fragments the corporate IP footprint – Fewer people are browsing from identifiable office networks, weakening IP-based resolution and company targeting.

• Form fills are rarer – Buyers are more protective of their data, especially in early research stages, meaning fewer explicit identifiers to match against.

• Cookie and consent friction – Between Chrome’s user choice prompts, Apple ATT, and GDPR’s tightening enforcement, even on-site capture suffers attrition.

The net effect: the ABM “go signal” comes later in the buying journey, after multiple unseen research cycles. By the time a first-party signal lands in your analytics, the buyer is already far along in their consideration process. That delay compresses your available activation window and forces you to run harder with what you have.

Universal IDs don’t solve this. UID2, RampID, ID5—none of them can replace the breadth of cookie-based tracking, and all suffer steep drop-offs when true opt-in is enforced. Matching based solely on non-first-party signals is a losing battle. The winning play is to:

• Maximise first-party certainty – Form captures, event registrations, product logins, and authenticated media access.

• Work probabilistic signals harder – Network quality, content consumption patterns, geo-behaviour, and co-occurring account activity.

• Combine modes – Hashed emails, login IDs, clean-room overlaps, publisher/retailer graphs, and contextual/firmographic targeting in a single stitched-together identity layer.

In B2B, the move is from domain-only resolution to company-centric, multi-partner match networks—lifting match rates by blending deterministic anchors with probabilistic amplification. This isn’t “upload a list to a walled garden and hope.” It’s product design: building the identity graph you need for your specific TAM, regional privacy constraints, and activation channels.

The brands that win here will be the ones willing to get heavier-weight, deeper-thinking partners—not point-solution vendors. In a world of thinner signals and later triggers, you need strategy, not just tech.

4) Welcome to the Signals Era

Search used to be your early-warning radar for intent. Cookies gave you a crude—but dependable—ID system. Both are fading fast. The replacement isn’t a single silver bullet—it’s a mosaic of signals that you score, weight, and orchestrate into action.

The winners in 2025 and beyond won’t be the ones with the longest target account list—they’ll be the ones who can see warmth building before it’s obvious. That means:

• Diversity of inputs – First-party intent data, account analytics, content engagement, contextual patterns, attention metrics, and even inferred signals like sudden cross-domain topic interest.

• Confidence scoring – Recency, corroboration, and relevance all matter; not every signal is equal.

• Activation mapping – Knowing which channel, creative, and format logic fits each signal—and each stage of the buying journey.

This is a shift from “find people who match my audience” to “know when an account is warming up—and meet them there.”

The GenAI discovery collapse, the privacy-driven thinning of identity layers, and the later-arriving ABM triggers aren’t obstacles here—they’re filters. The noise drops, the signal gets stronger, and if you’ve got the instrumentation to capture it, you can run an ABM program that’s sharper, leaner, and more resilient than anything in the cookie era.

5) Product thinking for the next wave

This new environment is where FunnelFuel—and my approach—thrives. We’ve built for constraint from day one, and that constraint is now the competitive advantage. Here’s what a next-gen B2B stack looks like when you embrace the Signals Era:

• AI-native discovery – Your brand and expertise are optimised for LLM ingestion, appear in answer engines, and are tested in emerging conversational ad formats.

• Privacy-designed data architecture – Consent, regional logic, and performance without personal identifiers baked into the core. GDPR and Chrome choice prompts aren’t retrofits—they’re the blueprint.

• Resilient identity graphs – Company-centric, multi-partner match networks combining UID frameworks, clean rooms, and publisher coalitions to recover match rates without over-reliance on any single layer.

• Signal orchestration – Real-time ingestion and weighting of diverse intent and engagement signals, scored with explicit confidence levels and human override for the moments that matter.

• Outcome-focused measurement – Moving beyond CTR to attention-adjusted, account-level outcomes, with modernized MMM approaches for a post-cookie reality.

The takeaway? B2B works brilliantly in this new era—if you stop chasing the old playbook.
When you can:

• Surface in AI-native environments

• Capture and score thinner, stronger privacy-compliant signals

• Stitch deterministic and probabilistic identity into a usable graph

• Activate at the exact moment an account warms

…you’re not just surviving the shift—you’re weaponising it.

The takeaway

Clicks are relics. Signals are the new currency—rich, varied, privacy-safe, confidence-scored, and mapped to the right activation points. Generative AI is reshaping discovery, privacy is forcing architectural clarity, and identity is now a multi-source design problem. The marketers who think like product builders—who design stacks that thrive amid constraints—won’t just keep up. They’ll set the pace.

---

Sources and references

[1] Macro shift overviews on AI + privacy + identity from: Gartner trends coverage; Basis/Insider Intelligence 2025 forecasts; IAB Europe privacy guidance.
[2] Gartner prediction: traditional search volume down ~25% by 2026 (primary), with reputable secondary coverage.
[3] Google AI Overviews impact on clicks and zero-click behavior (Wired / Search Engine Land analyses).
[4] Agency/brand behavior on Answer Engine Optimization and “AI search units” (Digiday); Gen Z/younger cohorts using LLMs vs search (survey coverage).
[5] Chrome third-party cookies: CMA/Google updates on user choice prompts; EMEA privacy impacts (IAB Europe, ICO).
[6] Apple ATT: measurement and ad effectiveness studies (Apple docs; independent analyses by SKAdNetwork partners).
[7] Advertiser concerns and adoption trends for the cookieless future (Insider Intelligence/eMarketer).
[8] UID2 technical explainer (The Trade Desk).
[9] RampID and LiveRamp Safe Haven clean room documentation.
[10] Publisher/retail media coalitions and clean-room partnerships (e.g., Albertsons Media Collective + LiveRamp).
[11] Industry POVs on company-based vs domain-based B2B resolution (IAB Tech Lab, B2B identity vendors).
[12] Attention metrics frameworks from IAB/Attention Council/MOAT/Advangelists; platform/publisher studies correlating attention with outcomes.
[13] Unified measurement and MMM resurgence for post-cookie attribution (Insider Intelligence/eMarketer; Google/MMM open-source libs).
[14] Perplexity/OpenAI/Google publisher guidance on optimizing brand presence in AI answers; SEO industry pieces on AEO.
[15] Practical experimentation: geo-lift/holdouts; modern MMM stacks and cadence (NCS, Meta/Google MMM resources, Neustar papers).