DGTG logo DGTG

AI Marketing and Search: Growth Strategy for Agencies

AI and search are rewriting client acquisition. See the metrics that matter, what to build in-house, and where agencies still win. A practical playbook for gro…

How AI and Search Are Reshaping Growth: What Agencies Must Master
A mid-sized B2B software client came to us last year with a familiar problem. Their organic traffic had grown 40% year over year. Their demo requests had fallen 12%. They had more visitors than ever and fewer buyers. When we dug into the data, the answer was uncomfortable: their highest-intent keywords were now answered directly on the search results page by AI-generated summaries. People read the answer, nodded, and left. No click. No session. No pipeline. That gap, between visibility and revenue, is where agency work now lives. AI marketing and search marketing are no longer separate line items. They are the same problem wearing two hats. And the agencies that treat them as one system will take the accounts that the ones treating them as separate channels will lose. ## Search stopped being a funnel. It became a filter. For twenty years, search marketing worked like a funnel. You ranked, you got clicks, you converted a percentage. The math was stable enough to forecast. CAC was a function of keyword difficulty and conversion rate. That math is wobbling. Google's AI Overviews, Bing's Copilot integration, and a wave of answer engines ([Perplexity](https://www.perplexity.ai/), [ChatGPT](https://chat.openai.com/) search, You.com) now resolve a large share of informational and comparison queries before a user ever reaches a website. For top-of-funnel content, click-through rates on position one have dropped in many verticals. We have seen CTR declines of 20% to 35% on informational pages that previously carried a client's entire awareness strategy. Here is the part most agencies miss: this is not a traffic story. It is a qualification story. The people who still click through from an AI-summarized result are further down the funnel than they used to be. They have already read the summary. They want the specifics, the pricing page, the integration docs, the case study. That changes what content earns its keep. What this means in practice: - Informational content that only restates consensus gets absorbed by AI summaries. Stop funding it at scale. - Content with proprietary data, original research, or a named point of view still earns clicks because AI systems cite it, and citation drives the click. - Bottom-of-funnel pages (pricing, comparisons, implementation guides) matter more, not less. They are where the filtered traffic lands. ## The new acquisition math We rebuilt a client's content budget last quarter around this idea. They sell HR software to mid-market companies. Previously, 70% of content spend went to broad educational posts. We flipped it: 40% to original research and benchmark reports, 30% to comparison and alternative pages, 30% to retention content for existing customers. Six months in, organic sessions were down 18%. Demo requests were up 31%. Cost per demo dropped from $340 to $210. The traffic was worth more because it was later-stage and better qualified. That is the trade most agencies have not made yet. They are still reporting sessions and rankings because those numbers are easy and clients expect them. But a [digital growth strategy](/dgtg/blog/ai-search-social-2026-growth-playbook-for-brands) built on session volume in 2026 is a strategy built on a metric that no longer maps to revenue. ### What to measure instead | Old metric | Better signal | Why | |---|---|---| | Sessions | Qualified sessions (engaged 60s+, viewed pricing or demo page) | Filters out answer-engine bounces | | Keyword rankings | Share of AI citations for target queries | Rankings without clicks are vanity | | MQL volume | MQL-to-SQL rate by source | Volume hides quality collapse | | CTR | CTR on bottom-funnel pages | Top-funnel CTR is structurally declining | ## Where AI actually earns its keep in agency work We have tested a lot of AI tooling across client accounts. Most of it is overhyped. A few things are genuinely transformative, and they cluster in three areas. ### 1. Research and synthesis Tools like Perplexity and Claude can compress a competitive audit that used to take a junior strategist two days into two hours. Not because they replace judgment, but because they remove the grunt work of pulling and summarizing. We use them to map competitor content gaps, summarize sales call transcripts, and draft first-pass briefs. The strategist still decides what matters. ### 2. Creative volume with a floor Ad creative testing used to be limited by production cost. Now a team can generate 40 headline variants, run them through a [brand-safety and tone check](/tech/blog/ai-safety-vs-speed-how-to-choose-tools-won-t-get-you-burned), and ship the top 10 in a day. Meta's Advantage+ and Google's Performance Max both reward this kind of volume, but only if the underlying offer is strong. AI multiplies a good offer and accelerates a bad one. We have watched agencies burn client budgets generating more of what was not working. ### 3. Retention and lifecycle This is the most underrated application. AI-driven segmentation in tools like Klaviyo or HubSpot can now trigger lifecycle messages based on behavioral signals that used to require a data team. For one ecommerce client, we rebuilt their post-purchase flow around predicted churn risk. Repeat purchase rate went from 22% to 29% in a quarter. LTV moved more from that single change than from any acquisition campaign that year. ## What we recommend If you run an agency or manage growth in-house, here is where we would put money this quarter. **For search visibility:** Invest in being cited, not just ranked. That means original data, clear authorship, and structured content that AI systems can extract. Tools like Semrush and Ahrefs now track AI visibility alongside traditional rankings. Use them. If your content is not getting cited by answer engines, it is invisible to a growing share of buyers. **For paid acquisition:** Consolidate around fewer, better offers. Run Google Performance Max and Meta Advantage+ only after you have a proven offer and a clean conversion signal. Otherwise you are feeding a black box garbage and blaming the algorithm. **For retention:** Build one lifecycle automation this quarter. Not five. Pick the highest-churn moment in your customer journey and address it with a triggered sequence. Measure LTV impact, not open rates. **For the agency model itself:** Move from retainer-by-deliverable to retainer-by-outcome. Clients are getting smarter about what AI can do for free. They will not pay for content production that a model can handle. They will pay for judgment, strategy, and accountability to a number. ## The hiring question clients are asking Every founder and CMO we talk to is asking the same thing: what do I hire an agency for versus build in-house? The honest answer has shifted. In-house teams now handle execution well because AI tools have lowered the skill floor for production. What they struggle with is pattern recognition across accounts. An agency that has run 50 lifecycle programs has seen failure modes a single in-house team will never encounter. That cross-account intelligence is the product now. Not the deliverables. If an agency pitches you on output volume, ask what they have learned across clients that you could not learn alone. If they cannot answer that, they are selling you a commodity that AI has already deflated. The growth teams that win the next two years will be the ones who stop treating search and AI as separate budgets and start treating them as one system for finding, qualifying, and keeping customers. The technology is not the hard part. The discipline is. ## FAQ **How quickly should we expect AI search changes to affect our traffic?** For most B2B and informational content sites, measurable CTR declines have shown up within two to three quarters of AI Overviews expanding in a given vertical. Bottom-funnel pages are more resilient. Audit your top 20 informational pages now and check whether they still earn clicks. **Should we cut all top-of-funnel content?** No. Cut the generic kind. Original research, proprietary benchmarks, and strong point-of-view content still earn both clicks and AI citations. The rule is simple: if a model could write your post from public information, do not publish it. **What is the single highest-ROI AI investment for an agency right now?** Lifecycle automation tied to churn prediction. It is less glamorous than content generation, but it moves LTV directly and clients can see the revenue impact in one quarter. Acquisition gets the attention. Retention pays the bills. ## Related on this site - [How to Choose a Marketing Agency in the Age of AI: 7 Must-Ask Questions](/dgtg/blog/how-to-choose-a-marketing-agency-in-the-age-of-ai-7-must-ask-questions) - [2026 Digital Marketing Trends: A Strategic Guide for Agencies and Their Clients](/dgtg/blog/2026-digital-marketing-trends-a-strategic-guide-for-agencies-and-their-clients-2) - [2026 Digital Marketing Trends: What Agencies Must Know Now](/dgtg/blog/2026-digital-marketing-trends-what-agencies-must-know-now)

Frequently asked questions

How quickly should we expect AI search changes to affect our traffic?

For most B2B and informational content sites, measurable CTR declines have shown up within two to three quarters of AI Overviews expanding in a given vertical. Bottom-funnel pages are more resilient. Audit your top 20 informational pages now and check whether they still earn clicks.

Should we cut all top-of-funnel content?

No. Cut the generic kind. Original research, proprietary benchmarks, and strong point-of-view content still earn both clicks and AI citations. The rule is simple: if a model could write your post from public information, do not publish it.

What is the single highest-ROI AI investment for an agency right now?

Lifecycle automation tied to churn prediction. It is less glamorous than content generation, but it moves LTV directly and clients can see the revenue impact in one quarter. Acquisition gets the attention. Retention pays the bills.

What to measure instead | Old metric | Better signal | Why | |---|---|---| | Sessions | Qualified sessions (engaged 60s+, viewed pricing or demo page) | Filters out answer-engine bounces | | Keyword

The honest answer has shifted. In-house teams now handle execution well because AI tools have lowered the skill floor for production. What they struggle with is pattern recognition across accounts. An agency that has run 50 lifecycle programs has seen failure modes a single in-house team will never encounter. That cross-account intelligence is the product now. Not the deliverables.

Related articles