Agency leads have not disappeared, they have moved. Long-tail blog traffic, directory listings, freelance marketplaces, and cold email are producing less, while AI answer engines, paid AI placements, and relationship-driven channels like podcasts, community, and partner programs are producing more. Computan CEO Sajeel Qureshi and CRO Joe Jerome recently sat down to compare notes on what they are seeing across their own pipeline and their clients' pipelines.[1]
The lead source shuffle is the shift in where qualified leads originate. The channels most agencies optimized for over the last decade (organic search, directories, marketplaces, email) are losing share, and new channels are absorbing demand. It matters because, as Sajeel put it:
"Most agencies are struggling with leads. And if they struggle with leads, they're struggling with sales." Sajeel Qureshi, CEO, Computan[1]
In a service business, a steady flow of leads is the oxygen. When the source changes and the pipeline does not adapt, the gap shows up in revenue within a quarter or two.
For years, one technical article on how to tell which CMS a website is using drove roughly half a million views to the Computan blog. That traffic has largely evaporated. As Joe explained:
"No one needs to come to our blog to figure that out. They did before AI, but they can just ask AI how to do that now."
Joe Jerome, CRO, Computan[1]
Simple, answerable questions now get resolved inside the AI interface itself. Content that only explains a fact is the first to lose clicks.
Directory leads are shrinking too. Computan still holds strong reviews on Clutch,[8] but the inbound volume from the site itself has dropped.
Buyers who once browsed a directory to find the best WordPress developer in Canada now ask an AI assistant. The assistant may still draw on those reviews, so directories can feed leads indirectly, just not through direct traffic.
Marketplaces are getting compressed. A job post that used to attract around 30 proposals can now attract 200 or more within a day, and many of those proposals are likely generated by AI tools. Sajeel called the result "almost counterproductive, counterintuitive."[1] More noise means fewer good matches on both sides.
A partner recently told the team that email marketing is not working as well as it used to. Sajeel's take was blunt:
"All that impersonal stuff is just not gonna work in a world of AI." Sajeel Qureshi, CEO, Computan[1]
Sajeel recently attended a HubSpot Partner User Group session on AEO, led by Tina Lavery, an AEO product manager at HubSpot. The core idea is that a growing share of everyday searching, such as finding the best accountant in a city, now starts in ChatGPT, Gemini, or Perplexity. Those models build answers from your website and from third-party citations that mention your company.[1]
The behavior that follows is the important part:
"No one's clicking on those links in the LLM. They're just saying, we recommend these guys. We go to the AI first to get some recommendations, and then we do our own research." Sajeel Qureshi, CEO, Computan[1]
Joe added that AI-driven clicks usually land on the home page rather than a deep article, so the home page and your brand signals carry more weight than they used to.[1] The session also suggested that buyers who use AI in their research are more likely to buy than those who do not, which is a good reason to make sure you show up there.[1]
HubSpot's AEO tool gives you a visibility score, prompt tracking, citation analysis, and prioritized recommendations inside your CRM.[2] It launched in April 2026 with competitor comparison and tracking across answers in tools like ChatGPT and Gemini.[3]
Paid placement inside ChatGPT is the newest option on the table, and regional targeting is now available. OpenAI added geo-targeting at the state, designated market area, and ZIP code level in Ads Manager, along with daily budgets.[5] OpenAI also reports that ChatGPT Ads reached a $1 billion annualized revenue run rate in under 200 days.[4]
One detail worth knowing: ads appear only to users on the Free and Go plans, not Plus, Pro, or business tiers.[5] That shapes who you can actually reach. Joe's view is optimistic but honest about how new this is:
"We're gonna dive in and become three-week experts like everyone else right now." Joe Jerome, CRO, Computan[1]
Anyone claiming a proven playbook after a few weeks is guessing. Treat it as a test budget, measure carefully, and learn in public.
The most consistent theme from the conversation was that human channels are working. Joe summed up the pattern:
"Social media, podcast, community, partnering. We're talking about very non-AI and very human oriented lead channels." Joe Jerome, CRO, Computan[1]
After a stretch of doubt about HubSpot earlier in the year, both leaders see momentum returning, especially around AI.
"Now we're starting to see HubSpot pull ahead in the AI space with tools like the AEO tool." Joe Jerome, CRO, Computan[1]
Joe also pointed to the Breeze Assistant becoming more useful and to the Agent Hub, which HubSpot describes as the home for every agent running your go-to-market and which was formerly called Breeze Agents.[6] For agencies, the practical point is that when lead sources multiply, your CRM needs to capture them cleanly. If AI referrals, podcast mentions, and community introductions all land in a generic "direct" bucket, you will not know what is actually working.
When lead flow is unpredictable, fixed headcount is risky. Joe made the case for a flexible model:
"Right now is a good time to have flexible staffing." Joe Jerome, CRO, Computan[1]
Instead of building internal teams that cannot sustain shifting demand cycles, agencies can bring in a block of hours, such as 40, 80, 120, or 160 a month, and draw on a set of specialists. Computan is seeing more requests for social media posting, video editing, landing page creation, and HubSpot internal work, and fewer for heavy custom development.[1] For more on how this works in practice, see our guide on how HubSpot agencies can scale without hiring more full-time specialists.[7]
Sajeel's analogy says it well:
"I'll just go to the ice cream store and get it whatever I need it. It's cheaper, it's easier, it's flexible." Sajeel Qureshi, CEO, Computan[1]
| Lead source | Direction | What is happening |
|---|---|---|
| Long-tail blog search | Fading | AI answers simple questions without a click |
| Directories (Clutch and similar) | Fading | Buyers ask AI first; reviews help indirectly |
| Freelance marketplaces | Fading | Proposal volume is exploding, quality is dropping |
| Impersonal email | Fading | Lower response from generic outreach |
| AEO (AI visibility) | Growing | Buyers shortlist vendors inside AI tools |
| Paid AI search | Emerging | New, regional targeting available, results unproven |
| LinkedIn and social content | Growing | People-first content starts conversations |
| Podcasts | Growing | Relationships that turn into referrals and projects |
| Community and partner programs | Growing | Trust built over years compounds |
The lead source shuffle is not a collapse, it is a reallocation. The channels built on being found by keyword are weakening, and the channels built on being known and recommended are strengthening. That is true whether the recommendation comes from a person in your community or from an AI assistant summarizing the web.
Agencies that measure their lead sources honestly, test the new ones carefully, and stay human in how they show up are the ones best positioned for the next stretch. Need help setting up HubSpot reporting, AEO tracking, or flexible development capacity? Talk to the Computan team.
It is the shift in where qualified leads come from. Traditional sources like blog search, directories, marketplaces, and email are weakening, while AI visibility, paid AI placements, and relationship-led channels are gaining ground.
People now ask AI assistants for answers and recommendations instead of browsing blogs and directories. Simple informational content gets answered in the interface, so fewer people click through.
AEO, or answer engine optimization, is the practice of improving how often your brand is retrieved and cited in AI answers. It matters because buyers increasingly shortlist vendors inside tools like ChatGPT, Gemini, and Perplexity before they ever visit a website.
Yes. HubSpot's AEO tool provides a visibility score, prompt tracking, citation analysis, and recommendations inside the platform.[2]
It is worth a small, well-measured test. Regional targeting is now available, but the channel is new, only certain user plans see ads, and proven benchmarks do not exist yet.[5]
Demand is shifting quickly, and fixed teams struggle to keep up. A flexible model lets you scale hours and specialties up or down without the cost of recruiting and carrying full-time roles.