Hello Reader
I'm busy. You're busy. And every agency owner and fractional CMO I know is running the same two tracks right now.
Track one is "deliver more with less, with AI, faster." Track two is "keep growing pipeline so track one keeps running."
Track one wins my calendar most days.
I don't think much about track two until the pipeline dries up and then I panic and go looking for a quick fix.
I know I shouldn't, but I get sucked into the LinkedIn and YouTube hype.
Someone posts about the new agent that will get UNBELIEVABLE results. Someone else swears the tool they tried this week is THE ONE for driving leads.
I block an afternoon, wrangle the latest agent into place and come out the other end with zero pipeline and no new clients.
Then I remind myself. I'm a marketer. I know how this works. I start with my audience, their pain point and the job they want done.
I open my client acquisition project in Claude and start building an outreach plan around the strategies I know will work.
But what works today?
To find out, I talked to the agency owners and fractional CMOs in my network.
Why the old playbook isn't filling your pipeline
The channels most of us relied on three years ago are getting crowded.
Web traffic is not what it was.
In the first four months of 2026, 68% of US Google searches ended without a click, according to SparkToro's June study with Similarweb. That's up from 60% in 2024.
For every 1,000 searches, only 276 now reach the open web. Blog traffic that used to earn one signup every hundred visits now needs three times the volume to earn the same signup.
LinkedIn is cracking down on AI-generated content.
On July 30, the company rolled out a "seems like AI slop" reporting button and started privately flagging posts in creators' dashboards when the writing reads as inauthentic. Company-page organic reach already sits around 1.6% of followers.
If your calendar is running on AI-drafted posts with light human polish, the platform is going to decide your posts are not worth surfacing, and it's now inviting your audience to help make that call.
Prospects are tuning email out.
According to Instantly's 2026 Cold Email Benchmark Report, the average reply rate has fallen from 8.5% in 2019 to 5% in 2025 to 3.43% in 2026.
Volume keeps climbing and inboxes keep tightening. Every marketer I know is sending more outbound than a year ago and getting less back.
What used to work was a wide net across many channels. What works now is a shorter list of tactics where you have the resources go deeper.
"The strongest results I'm seeing come from highly personalized, signal-based outreach," says Chris Arden, CEO of Demand Lab Agency. "Connecting a prospect's pain points to a clear value proposition means I can develop a more thoughtful messaging cadence. Then the outreach feels timely and relevant, not like another automated campaign."
Where clients come from now
1. Referrals from existing clients
Referrals are still the highest-converting channel and they probably always will be.
What changed is that your referrer needs to know how to refer you.
Give your best clients a two-sentence version of what you did and how it helped them. And then ask for the referral.
How AI can help: Feed your client list into Copilot (with PII removed) and ask it to identify your top 20. Load the client project, the transcript from the last QBR or another source that shows why the client is thrilled with your work. Then ask Copilot to draft a custom outreach for each one. Bundle the set-up for your account team or ask Copilot to draft the emails and drop them into your inbox ready to send.
2. Pairing thought leadership with paid
Your newsletter, YouTube channel, webinars and LinkedIn articles are working hard to establish your credibility. You now need to pair them with paid to drive leads your team can convert.
Amy Baker, Fractional CMO at Flourish Marketing, says posting thought leadership on LinkedIn is performing well this year.
"We're having a lot of success pairing LinkedIn ads with long-form content, like an industry maturity model," she says. "The ad drives downloads of the report and that drives new leads."
Amy's version works because the ad gives people access to something they want. If the content earns its keep, conversions follow.
The same play works for speaking. Book podcasts and webinars for yourself or a senior person on the team, turn each one into a long-form asset and pair it with an ad campaign to drive views and downloads.
How AI can help: Drop the webinar or podcast transcript into Claude along with your audience's biggest pain points and ask it to build a 4–5 page white paper. Once the copy is right, drop the Word doc into Claude Design with your design system and ask it to design the finished piece.
Side note: if you want to see how this is done, sign up for the Creative Session I'm hosting in September.
3. Account-based, not lead-based
The shift from lead gen to ABM isn’t new, but the tooling around it has changed enough that marketers can now run this play without a full enterprise stack behind them.
Tiffany Teixidor, VP Marketing at Educate 360, is running exactly this play at scale.
"Paid search is still our strongest bottom-of-funnel channel and it's helped offset declining organic traffic as we shift to an AEO-focused strategy," she says. "We're moving beyond a traditional lead gen model, where an individual SQL is passed to sales, toward an ABM model that looks at engagement across the entire account."
Tiffany's team runs cross-channel campaigns that hit different pain points in different formats. That activity increases signal volume, surfaces which accounts are warming up and shows which topics and solutions are generating interest. Only when multiple contacts from the same account start engaging does her team map the broader buying committee and pass it on.
Stop chasing individual contacts at target companies. Track engagement at the account level and only reach out when two or three people are showing signals.
How AI can help: Build a project in Claude or ChatGPT with your ICP, the accounts you're targeting and the signals that matter (job changes, hiring, tech installs, content downloads). Ask it to review the list weekly and flag accounts where three or more people have shown activity. That's your outreach list.
4. Targeted events
With everything moving to AI, people are craving face-to-face time. Events are a higher-lift channel but still well worth the effort. The relationships you build carry through to pipeline at a much higher rate than digital advertising.
Toby Frowen, from R.E.A.L. Advisory agrees but says the key to success is to have a set of goals across all the events in your portfolio, not just event-by-event.
“Do you have clean attendee data, a defined follow-up process, and a way to connect the event to your pipeline?” she says. “That's what most orgs get wrong. They focus event-by-event, skip those questions, and then wonder why nothing's sustainable or ties back to ROI.”
AI can plan an event, but it can't build that foundation for you. Once you have your goals and plan in place, every event reinforces the same outcomes instead of starting from scratch each time.
How AI can help: Drop your ICP into Perplexity and ask it to research relevant events. Ask for a score on each so you can prioritize. Then create a project in ChatGPT with your event planning info (budget, timelines, goals, who can attend), load the list from Perplexity and ask for an execution plan with steps and a timeline.
5. Signal-informed outreach
Volume outbound is dead. Precision outbound still works when you do it well, and it needs the right signals and the right plumbing.
You need a system that identifies your target contacts and the signals that show they're ready for what you sell. Most people I know pull signals like ICP fit, job title changes, hiring patterns and competitor connections on LinkedIn.
Elise Oras, Cofounder of Wheels Up Collective, takes signal-informed outreach in a different direction. Instead of cold email, she uses the signals to earn a warm intro.
"LinkedIn is our best channel this year, and it has nothing to do with posting or our feed," she says. "When we want to reach a company, we find who we're already a second-degree connection to and ask for a warm introduction. Before we even ask, we know something about their need. From research, from what they've read on our blog, from landing pages they've visited and from tools like Clay and Apollo. So when the intro happens, the email is targeted and specific, filling a real gap instead of adding more noise."
Elise's approach is worth studying if you're an agency owner without an SDR function. You don’t need a big outbound engine. You need a small number of well-timed, well-researched conversations that start with someone your prospect already trusts.
Avoiding AI slop in your outbound
Copywriters often complain to me that AI-generated copy is rubbish and needs to be completely rewritten.
"If that's the case, they're doing it wrong," Chris says.
"I hear people say AI can't write outreach without producing slop. I vehemently disagree. Sure, it can sound like slop if you haven't built a clear voice and tone guide or aren't using the right models. But if you know what you're doing, AI can write like a veteran copywriter."
How does Chris do it?
He feeds his outbound platform a voice and tone doc along with case studies and testimonials. As the tool writes each message, it checks the target company's industry and pulls in a matching case study. The prospect gets an email that references their space using proof points that map directly to them.