I had a call with Brian on a Monday morning. He is the new CEO of an employee parking SaaS, six weeks into the role, and he wanted to talk about content. Specifically, whether live webinars were still worth doing, what a reasonable LinkedIn cadence looked like, and how to think about SEO when LLMs are eating search.
We did not get to any of that.
Twenty minutes in, Brian was describing a recurring customer problem. Employees at large tech companies are pinging each other in Slack channels asking colleagues to please run downstairs and move their cars so the next person can charge their EV. There is no good public answer to this anywhere. Brian sees it constantly. So I asked him: "What is your advice when a customer raises this with you?"
He gave me a clean, articulate, three-minute answer. He had thought about it. His team gives a version of it every week. As far as I could tell, that exact answer existed in no form on the public internet, no matter what tool you used to search.
The realization that surprised both of us was not about webinars. It was that Brian's most valuable asset, the asset his competitors with the same AI tools cannot replicate, was sitting unwritten in his head and in the heads of three or four people on his team who talk to customers every week.
This is the founder-led growth opportunity in the LLM era. Not the version most people are writing about. Let me lay it out plainly.
The argument most people are making, and where it is incomplete
The standard argument against SEO-first content in the LLM era goes like this: AI generates so much output that you have to be more authoritative, more deeply researched, more original. That framing is correct but incomplete. The deeper problem is not quality. It is that anyone can produce more authoritative, more deeply researched, more original content using the same tools you use. Whatever Claude or ChatGPT or Gemini surfaces for you, your competitor surfaces too. SEO optimization still gives you an uptick. It is just easy to copy, and it does not make you special.
What does not exist in any training set is the conversation Brian had with a head of workplace at Meta last Thursday. The exact sentence one of his customers used in their renewal call. The pattern his sales team has noticed across seventeen demos this quarter. None of that has been crawled. None of it can be replicated by a competitor with the same MarTech budget. It is, by definition, your private dataset.
In my experience working with B2B SaaS founders, this is the part most teams underweight. They invest in technical SEO because the playbook is clear. They invest in tooling because vendors line up to sell it. They under-invest in the one discipline that produces irreplaceable signal: actually talking to customers and writing down what they say.
Why "we already talk to customers" is not enough
The strongest counter to this argument is that most companies already capture customer voice. NPS surveys, win/loss interviews, sales call recordings, support tickets, customer advisory boards. The data exists. The problem is not access.
The problem is extraction.
When I asked Carson, Brian's solo marketer, how many one-on-one customer interviews he has run in his eighteen months at the company, the answer was zero. He has seen NPS reports. He has read transcripts. He has never sat across from a customer and asked an open-ended question. This is not unusual. Across the founders and marketers I work with, fewer than one in four have conducted a real customer interview in the last quarter. The rest work from second-hand sources. NPS verbatims the customer wrote in twenty seconds before closing the survey. Sales call summaries written for CRM hygiene rather than insight extraction.
Second-hand sources have a half-life. They give you the surface answer the customer was willing to commit to writing. They do not give you the pause before the answer. They do not give you the "well, actually" reversal. They do not give you the offhand comment that contains the real problem. None of that comes through Salesforce.
If your founder-led growth strategy depends on differentiated content, and your customer insight comes only from NPS, you are competing on the public dataset along with everyone else.
The framework: extracting signal that is worth scaling
Once you commit to running real interviews, the next problem is structure. An unstructured customer conversation produces anecdotes. A structured one produces content. I use two tools that pair well together.
The first is what I call the Why/How Ladder. It is a method I have been refining for years and wrote about in more detail in Syntropy. The short version: take any customer statement and climb up the why-ladder until you reach business consequences, then climb down the how-ladder until you reach concrete proof. Most interviewers stop at the surface. The signal is two or three layers deeper.
Let me show this with Brian's EV charging problem.
The surface answer is: "We need a way to manage EV charging for employees."
Climb up the why-ladder. Why does this matter? Because employees are interrupting work to move cars. Why does that matter? Because senior engineers are missing thirty to forty minutes of meeting time, multiple times a week, to manage their charging session. Why does that matter? Because the cost of disrupted senior engineering time, multiplied across a campus, runs into the tens of thousands of dollars per facility per year. And because employees who are frustrated with their commute and charging experience are measurably more likely to leave.
Now climb down the how-ladder. How are we different? They define rules and policies for who charges when, instead of leaving it to a Slack-channel free-for-all. How does that play out in practice? Reservation logic, nudges, queue management, rules that adapt by team and by office. How would a customer prove it worked? Before-and-after data on charging session completion, employee complaint volume, and Slack-channel chatter.
That is one customer problem, fully laddered. It is also a six-part LinkedIn series, a webinar, and three blog posts. Brian's team knows all of this. None of it is published.
The second tool is the P1/P2/P3 persona split. In any B2B SaaS deal, you have at least three personas. P1 is the user. P2 is the economic buyer. P3 is the influencer or blocker. The Why/How Ladder produces different content depending on which persona you climb for. For the head of real estate (P2), the why is operational efficiency and ESG reporting. For the engineer trying to charge their car (P1), the why is "I want to do my job without managing a parking lot." For the CFO (P3 in this case), the why is cost-per-employee and retention math.
Most B2B content collapses these into one undifferentiated message. Three sharper, persona-specific pieces will outperform one generic piece every time.
This was the moment on the call that surprised both Brian and me. He had been treating the user as toothless, assuming the head of real estate discovered the problem independently. When I pushed, Brian agreed: the head of real estate at Meta does not wake up worrying about EV chargers. They learn about the problem because employees complain. Which means the most efficient marketing motion may be to talk directly to the high-income, sophisticated EV-driving employee and give them language and evidence to escalate the problem internally. That is a fundamentally different content strategy than the one Brian came on the call to discuss.
Thought leadership, demystified
Here is what most founders get wrong about thought leadership. They think it requires a futurist take, a contrarian prediction, or a TED-style framework. It does not. It requires one thing: an opinion grounded in something you know about your customers that nobody else knows in the same form.
If Brian publishes a one-page article titled "Why Slack channels for parking are a $200K-per-year problem at large tech employers," that is thought leadership. It is also probably the only article on the internet that frames it that way, because Brian is one of a few hundred people on earth who have heard the complaint from twenty different enterprise customers and synthesized it.
You do not need a manifesto. You need an opinion that is downstream of customer evidence. Take a position. Help the customer solve the thing.
The daily content engine that actually works
Here is the workflow I use myself and recommend to founders I work with. It assumes you have done the customer interview work above. Without that input, no engine will save you.
- Have everyone in a customer-facing role document one customer question per day that the LLMs answer poorly. By "poorly" I mean the answer Claude or ChatGPT gives is generic, abstract, or subtly wrong in ways only an industry insider would catch. Three out of four of these questions will not be useful. The fourth one is your content.
- Voice memo the answer while you walk, drive, or wait. iPhone now transcribes natively. Android has good alternatives. The point is to capture the answer in your spoken voice, because that is the voice your customers will recognize.
- Drop the transcript into Claude with a prompt like "tighten this without changing the voice, flag any claims that need a source, and suggest a sharper opening sentence." Treat the model as your editor. Not your author.
- Edit lightly. Ship.
- Repurpose. The same answer becomes a LinkedIn post on Monday, a section in the email newsletter on Friday, a slide in next month's webinar, and the basis for a longer blog article a quarter from now.
Three things make this work that most founders skip. The source material is real, not invented from a content calendar. Your voice carries through because the input is your spoken answer, not a written prompt. The LLM edits, it does not author. That last distinction is the one that determines whether your content sounds like you or like everyone else.
Where to put it: webinars and the newsletter
A quick word on distribution.
Live webinars are having a quiet comeback, specifically because they are visibly human. In a feed dominated by generated text, a live person answering live questions about a real problem they have seen at twenty companies cuts through in a way pre-recorded polished content does not. Webinars do not need production value. They need to be true and specific. A founder talking off the cuff for thirty minutes about EV charging policies they have watched work and fail will outperform a glossy webinar on "the future of workplace parking" every quarter.
Newsletters are the channel most B2B SaaS teams underweight. Curation is content. If every Friday you send your customers a tight email with two of your own articles, three pieces from elsewhere that you found valuable that week, and one anonymized observation from a customer call, you have built an asset most of your competitors will not be able to replicate in the next three years. Curated attention is a different product than generated content, and the tools that flood the latter make the former more valuable, not less.
The thing to remember
The trusted guide status is not won by publishing more, optimizing harder, or buying more attention. It is won by paying attention to the conversations you are already having and committing to write them down.
Your customers are telling you exactly what content they need. Most teams hear it and do not capture it. The teams that capture it, ladder it, segment it by persona, and ship it daily are the ones that earn the trusted guide position over the next three to five years.
Brian came on the call wanting to know whether webinars still worked. He is leaving with a different question, which is the question every founder should be asking right now: what do I know about my customers that nobody else knows, and what stops me from writing it down today?
Discussion items
- How many one-on-one customer interviews has your team run in the last ninety days? Not surveys, not sales calls, not advisory boards. Real, open-ended interviews where you went in to learn rather than to sell or measure.
- Of the customer questions your team answered last week, how many made it into a published piece of content? The honest answer for most companies is zero, and that gap is where founder-led growth lives.
- Which of your three personas (P1 user, P2 buyer, P3 influencer or blocker) is currently underserved by your content? Most B2B SaaS companies write almost exclusively for P2 and ignore P1 entirely.
- What is your current ratio of "generated content" to "customer-derived content"? If you cannot answer this, you do not have a content strategy. You have a content output.
Questions to ask yourselves and your team this week
- What is one customer question we got this week that the public internet does not answer well?
- Who on our team is closest to that question, and have we captured their answer in writing or audio?
- If we had to write a single short LinkedIn article today based on something we learned from a customer in the last seven days, what would it be?
- Are we treating our LLMs as authors or as editors, and what would change in our content if we flipped that?
- Which of our customers would be willing to do a thirty-minute interview this month if we asked them straight?
- What is one channel (webinar, newsletter, podcast appearance) we have been deprioritizing because it does not scale, and what would happen if we scaled it deliberately for the next two quarters anyway?
Sources and further reading