From clicks to conversations: The CMO’s guide to winning in AI-driven search
Search has always been the workhorse of digital marketing. But the rules of the game have changed: Assistant-driven referrals, though small in number, already convert up to 20× better than Google clicks. The age of keyword search is ending, and AI-driven discovery is taking its place.
For leaders responsible for growth marketing, brand visibility and customer experience, this requires a re-wiring of how you and your team think about search and the way customers find and choose you. Marketers who adapt will win a disproportionate share of high-intent buyers. Those who don’t risk paying more for dwindling visibility while losing control of how their products and services are described.
What’s driving the new search imperative
Keyword search still delivers volume, but AI now delivers the value. Large-language-model–powered assistants and conversational interfaces have become the first stop for product research, service inquiries, and brand comparisons. While retail has been slow to move, financial services, technology, telecom, and healthcare leaders are already using AI to sharpen and scale their ad strategies. What began as a trickle is becoming a surge: ad spend on AI-based search is forecast to jump from $1 billion in 2025 to nearly $26 billion by 2029, totaling 13.6% of all search ad spending.
Caption: Evolution from keyword-based search to AI assistants changes everything – from the underlying technology to user behavior to the way brands optimize.
Figure 1 illustrates the stark contrast between the old “keywords era” of search and the emerging world of AI assistants. On the left is a model built on indexing, link ranking, and exact or partial keyword matching – an environment where users fired off short queries, opened multiple tabs, and performed their own comparisons while marketers optimized for SEO.
On the right is today’s reality: large language models, retrieval-augmented generation, and vector search power multi-turn research inside chat interfaces. Here, customers can compare, shortlist, and even purchase without leaving the assistant surface. For brands, success shifts from classic search engine optimization to Answer Engine Optimization (AEO) and AI Assistant Optimization (AAO) – ensuring information is structured, discoverable, and quotable inside the new conversational discovery channels.
This is a structural change in customer acquisition that collapses the old funnel, compresses journeys, and resets how success is measured. The old search playbook (i.e., optimize for keywords, bid for clicks, measure last-click attribution) no longer matches the buyer journey. Let’s take a moment to see what has to go away to make room for the new playbook.
What’s really changing
Search is still volume. AI is now value.
Zero-click answers and conversational UIs are shrinking the real estate your brand once owned on search engine results pages (SERPs). Around 80% of users rely on AI-generated summaries in at least 40% of their searches. This has already contributed to a global decline in organic traffic to your website of 15–25% versus 2024.
Discovery is fragmenting. Customer journeys now start in search, social, chatbots, marketplaces, and AI assistants — often without the explicit use of keywords. In June 2025, AI platforms (e.g., ChatGPT, Gemini, Perplexity, Grok, etc.) generated 1.13 billion referrals to the top 1,000 websites globally. That figure is up 357% year-over-year (with ChatGPT dominating 80% of this traffic).
Decisioning is opaque. Ranking logic churns faster than classic SEO signals — 70% of sources in Google’s AI Overviews changed within two to three months — and privacy changes make attribution even harder. Paid media auctions, meanwhile, are being driven by AI signals (audience, context, creative fitness), raising acquisition costs for brands that aren’t AI-ready.
The old way of doing search is no longer sustainable. It’s time for new tactics, new tools, and a and a new way of thinking about the buyers journey.
The new growth playbook: Step-by-step tactics for the AI era
The leaders who thrive in this new environment aren’t clinging to legacy tactics. They’re building entirely new playbooks designed for AI-first discovery. You need to rethink your playbook from the ground up. Think in two lanes:
- Answer engine optimization (AEO) for Google, Bing, AI Overviews, and social and marketplace search.
- LLM and RAG-readiness so that the third-party models and advertiser’s assistants can reliably surface and cite them. RAG, or retrieval-augmented generation, is an AI framework that enhances the output of large language models (LLMs) by connecting them to external data sources in real-time.
The framework below offers practical steps to address current gaps in the short term (0-6 months) with respect to the above lanes while also embracing the evolving landscape and preparing for the future. (Note: The examples are applicable to organizations across domains.)
Four moves stand out.
Move #1 — Own the answer
Customers increasingly get answers before they click. If your brand isn’t the answer, you’re invisible. That means your team must focus on creating a single source of truth — bite-sized, date-stamped, conflict-free facts that AI assistants can reliably quote.
How
- Equip marketers with the skills to craft LLM-friendly prompts for ad generation, content summaries, and audience insights.
- Use chain-of-thought prompting to guide AI through reasoning steps for better targeting.
- Build a living facts hub with clear, date-stamped answer snippets (30-60 words) and up-to-date specs or pricing tables.
- Layer in schema.org markup and JSON-LD feeds so assistants can easily find and use your information. Include a short, reuse notice (“You may quote with attribution.”).
- Use modern tools such as LangChain to break large docs into intuitive sections (e.g., “Router Setup,” “Signal Issues,” etc.) so an AI assistant can answer questions such as “Why is my internet slow at night?” with pinpoint accuracy. Store them in a vector db for fast retrieval.
- Use RAG-powered systems to generate ad creatives based on real-time user queries and retrieved context.
Examples
- “Minimum for 4K streaming is 25 Mbps per stream; our Fiber 300 supports two 4K streams.” Updated: 2025-08-15. Evidence: /network-tests-H1-2025. Reuse: Quote with attribution to
.
Owners
- Content Lead, SEO/CMS Tech, Legal
Result
- You control your category and brand narrative before someone else writes it.
Move #2 — Shift your media mix
Assistant surfaces are no longer experiments — they’re growth engines. Testing conversational placements across Google AI Overviews, Bing Copilot, Perplexity, and Meta Click-to-WhatsApp gives your brand a seat in the in-answer economy. Success depends on offers that are short, clear, and quotable in a single turn.
How
Platform
- Google AI overviews/AI mode: Make sure your Search (and PMax, where it fits) is well-funded for high-intent terms. You can’t target AI Overviews directly, but eligibility flows from your existing campaigns and signals, so prioritize clean sitelinks, up-to-date location assets and call extensions to boost your chances of being included.
- Microsoft copilot or bing chat: Keep your Microsoft Ads Search and Shopping campaigns active. Microsoft extends popular formats into chat and even offers a Chat Ads API that plugs into partner “Ask AI” experiences, opening new conversational touchpoints.
- Perplexity sponsored follow-up questions: Tap into upper- and mid-funnel exploration with sponsored follow-up questions that nudge users toward a related, brand-framed question.
- Meta click-to-whatsApp: Run Facebook and Instagram Ad placements that launch a prefilled WhatsApp thread, making it effortless for prospects to start a conversation.
Campaign & creative
- Set up test matrices (e.g., intent, geo, offer, etc.) and use unique codes or UTMs for each assistant touchpoint and review weekly lift analysis to see what’s working – and what’s not.
- Write ad copy the way people naturally ask assistants for help. Phrases like “best for beginners” or “eco-friendly option” feel conversational and increase the odds your ad will show up.
Examples
- Promote offers in a conversational way: “Fiber 500 installs in Pune in 48 hours – use code FIBER48.” This type of one-line, location-specific incentive works naturally inside an assistant’s answer.
Owners
- Marketing Analytics, Performance Marketing, Finance
Result
- Capture attention in the exact moment of intent.
Move #3 — Measure beyond clicks
Last-click attribution undercounts the real influence of your search activities. Assistants compress journeys because customers don’t have to hop site to site to get a reliable answer. When journeys compress, clicks are no longer the currency of measurement. Brands need zero-click attribution frameworks that capture when the decision was made – and what influenced it. This includes AI-specific promo codes, conversational IDs, assistant UTMs, and post-purchase surveys.
How
Start tracking three kinds of signals:
- Deterministic signals: Use AI promo or offer codes, conversational IDs, assistant UTM parameters or referral codes to tie assistant activity directly to revenue.
- Declarative signals: Ask customers where they decided – via quick post-purchase surveys (“Where did you decide to choose us?”) or by capturing the decision source in call center or retail systems.
- Contextual signals: Monitor a “share-of-answer” panel to see which brands assistants cite for your top intents, and map spend and impressions in assistant placements against those insights.
Examples
- Add “ASSTCODE123” or a similar code to offers quoted in assistants and track redemption to revenue to measure impact directly.
Owners
- Data Engineering, Commercial, Martech, CX Ops
Result
- See — and value — the influence that assistants exert on buying decisions.
Move #4 — Govern with confidence
AI doesn’t just scale opportunities. It scales risk. Leaders must establish guardrails for pricing claims, eligibility, privacy, and consent within conversational flows. That means pre-approved copy blocks, disclaimers, logging of assistant interactions, and consent services that keep offers compliant by default.
How
Start tracking three kinds of signals:
- Include pre-approved copy blocks, offers (with IDs, geo, start and end dates, and stackability), and disclaimers in place, along with a clear escalation path for edge cases so assistants always give accurate answers.
- Log and audit assistant interactions regularly to catch issues early and show proof of compliance.
- Maintain SLA tables by pin code and publish outage or capacity feeds frequently so your information stays current.
- Use PII detectors and redactors in prompts and logs to protect customer data by default.
- Run a consent service to ensure every offer and ad quoted inside an assistant meets regulatory requirements.
- Eliminate discriminatory offers, dark patterns, and unsupported superlatives (“fastest in city”) unless you have evidence — trust is earned through accuracy.
Examples
- Auto-append disclaimers such as “Promo valid in UK only; standard install fees apply” to assistant quotes to avoid missteps.
Owners
- Privacy/DPO, Security, CX, Product
Result
- Protect brand integrity while embracing conversational commerce.
Emerging AI-driven advertising metrics
As AI reshapes how people search and discover, the old KPIs no longer tell the whole story. Click-through rates and last-click attribution can’t measure influence when decisions happen inside conversational answers. New metrics are emerging to help leaders see how visible, quotable, and competitive their brands really are in AI-driven environments.
Success in the AI era will be measured less by clicks and more by how often your brand is cited, surfaced, and trusted inside AI-generated answers. As you can see in Figure 2, these emerging measures are signals of how well your organization is adapting to AI-first discovery. Incorporating them into your current KPIs will lay the groundwork for the next phase: building an operating model where your content, systems, and teams are architected for AI discovery from the ground up.
The long game: Architecting for AI discovery
The short-term playbook helps brands adapt. The long-term strategy is where leaders win. That means reimagining your advertising and marketing operating model.
Your organization will need these tactics to ensure a long-term winning strategy:
Develop assistant-ready multimodal brand assets (text, images, video) that assistants can surface as media-rich answers.
Stand up your own branded internal assistant to handle acquisition, support, and scheduling. Layer RAG over your fact store and tools so it can manage address checks, quotes, payments, reschedules, and more.
Integrate identity and CRM so every interaction is connected. By linking conversation IDs and household graphs to your CRM, your assistants can go beyond simply answering questions to actively guiding customers through their journeys.
Move beyond last-click metrics to value the influence of assistants and use marketing mix models (MMMs) that factor in conversational exposure, promo-code anchors, and contribution modelling.
Establish AI Centers of Excellence to anchor your conversational commerce strategy. Bring together regulatory readiness, data governance, and CX design so teams can innovate safely and scale best practices across the business.
Build regulatory and commercial readiness into your operations. Stay audit-ready across ads, pricing, and data use with legal engineering reviews, model cards, audit trails for claims and quotes, regional policy packs, and a consent-friendly UX.
Adopt a content-as-data operating model to create one source of truth. Power your web, partners, and assistants from the same headless CMS or knowledge graph with versioned feeds, evidence pages, and reusable, licensable snippets.
The mandate is clear: evolve your operating model or risk watching competitors rewrite the story your customers hear first. And the clock is already ticking. In a landscape moving this fast, you can’t afford to wait for everything to fall into place—you need to scale as you learn.
EXL helps you turn strategy into immediate action, so your brand doesn’t just keep up with change, it stays ahead of it.
How EXL makes it real
Success in this new era demands more than tools. It demands a partner who can weave data, design, intelligence, and strategy into a seamless narrative of growth. That’s where EXL stands — not just as a service provider, but as a catalyst.
With deep expertise in digital transformation, EXL helps leaders move from uncertainty to clarity, from pilot projects to scaled advantage.
Future-proof digital strategy. Rapid AI-readiness assessments and executive roadmaps.
Technology enablement. Bespoke integration of AI into your marketing stack and operations.
Content & experience transformation. AI-optimized content portfolios and real-time optimization frameworks.
Analytics & measurement. Hybrid analytics stacks that capture new AI-driven metrics and build holistic attribution.
Operational change management. Training, governance, ethics, and best practices for generative AI across teams.
Your next move
AI-driven search is already reshaping your customers’ journeys. The question isn’t if you’ll adapt. It’s how quickly.
EXL can help you take the first step with a clear-eyed diagnostic of your AI-search readiness — and chart the path toward building an organization that doesn’t just survive this shift but thrives because of it.
Let’s start the conversation.
Author: Tanuj Arora Head of Digital & AI Solutions
Co Author: Deepa Yadav AEO Solution Lead
Reviewer: Aakash Gupta Vice President, Head of AI and Analytics Services