Everybody Bought the Same Superpower in the Same Quarter

To kick off Boston’s AI week, I attended a MTLC session on Building the AI-Powered GTM Engine. The room was full, and the content was actionable while still feeling like déjà vu. I've been here before; we’ve been here before.

Mike Grandinetti, an executive fellow at Harvard Business School, opened the workshop with the number every brand owner should know: reply rates to cold email are down to just above 3%. He explained that everyone picked up the same capability (ahem, AI) at the same moment, so whatever edge it bought lasted until competitors did it too. David Chang, co-founder of Via AI, continued, noting that there are roughly 10 million GTM professionals in the U.S., and there’s 4x the outbound volume competing for the same inbox, plus about three hours of a seller's day is spent deciding who to contact. Nobody who has looked at their own reply rates this year is going to argue with these numbers, including me.

Chang put warm outbound response somewhere around 20-30% against 2-3% for cold. As someone who runs a brand marketing and communications agency, I tried to temper my enthusiasm for any finding that says reputation outperforms volume and focused on the point, which is that having a warm pipeline isn't a channel anyone can switch “on.”  

Nobody is sitting on an unused warm outbound leads list, but what is underutilized is the creativity and asset development needed to manufacture that warmth before it happens.

Things got interesting when Grandinetti made the point that the first reviewer of this communication and content is no longer a person. It's a model. We have been saying this was the inevitable path since the day GPT entered the chat, but Jen Griffin Smith, CMO at Acquia, had the numbers to back it up.  She cited that 98% of her site traffic is bots. Traffic arriving from LLMs converts at roughly 2x the rate of her SEO traffic. Video converts at roughly 3x. Reddit, LinkedIn and YouTube are the top three sources cited in LLM answers, which means the content deciding whether you get recommended is sitting somewhere other than your website.

Adobe's retail data says the same thing. In March 2025, traffic from AI assistants converted 38% worse than everything else. In March 2026, it converted 42% better. That is a full reversal in a year.

Put this data side by side, and you’ll see that the assets that manufacture warm leads are the same assets a model has a reason to cite.

• The winner will be the data nobody else can produce. For example, Carta publishes State of Private Markets because it sits on cap table records. Ramp publishes an AI Index because it sees corporate spend. If your content restates what your five competitors are also saying, there is no LLM citation to win and nothing for a buyer to remember. A number only you can generate is the one thing in your program that can't be paraphrased from somebody else.

• Use a person rather than a logo. Tom Orbach's list of 60 growth ideas in Lenny's Newsletter last week has a good example: Torq, a security company, employs a character called Trevor From Torq, played by an actor, who posts memes on LinkedIn. A named human accumulates an audience, and a company page doesn't, and that audience is what makes the outreach warm two quarters later. Most executives I talk to agree with this, then put the budget behind the logo.

• Do the work before you ask for the meeting. This is something we do at Greenough when talking to prospective brand partners. We don’t just provide generalities of what can be done with their brand, we do the discovery, look at the AI visibility, share of conversation and share of voice data. We take the time to understand what the brand does and where it wants to go BEFORE we meet with them and before we ask for a single dime. This is as helpful to us as it is useful to the prospect. Staying fresh on how brands are performing across the industries we work in and keeping our outbound conversation backed by data and the time they spend with us valuable even if we don’t move forward is something we feel passionately about. It was great to hear this approach is a gold standard in standing out in the sea of sameness. Also, if we’re willing to make this level of investment before earning a dime, imagine what it’s like to be our partner?!

Megan Anderson, Chief Customer Officer at Constant Contact, had a story that flipped the script. She described an AI agent called Luna that places and takes customer calls and says up front that it's AI. She read a transcript of a two-hour call where the agent offered to hand the customer to a human four times, and the customer said no every time. What that customer needed was patience, and no company is staffing a person for two hours on a sub-hundred-dollar account. So, humans may not always win, especially when the job is patience and repetition. But as we note in our own AI Charter, AI loses when the job is judgment and accountability, which is critical for brand positioning, messaging, and deciding what a company should be known for.

We all left the workshop with the same question we all came in with: how do you cut cost with AI without cutting the only thing that makes you worth citing?

Griffin Smith said every board she talks to is asking how AI takes cost out, and mentioned a portfolio company where the CMO's goal is to have zero people in marketing. Meanwhile, the channels that decide whether you're seen are discounting machine-made work, and the content that converts is the unique content only one company could have produced. So, the next time your BOD favors the AI cost savings, be quick to point to the “machine” working against us when it comes to developing differentiated value.

What I’m taking with me as I prepare our own GTM plan for 2027: If your 2027 pipeline looks weaker than you’d like, the problem isn't your outreach stack. It's that you haven't decided what your company knows that nobody else does.

‍