Talk, Don't Type: Why Interviews Beat the Blank Page
Frankly Speaking Team
July 14, 2026 · 7 min read
Ask a founder to write a LinkedIn post about their pricing model and you'll get a delay of somewhere between three days and forever.
Ask the same founder, on a call, "why did you move away from per-seat pricing?" and you'll get four minutes of specific, opinionated, well-structured explanation. There will be an anecdote about the customer who broke the old model. There will be a number in it. There will be a view.
Same person. Same knowledge. Completely different output. The difference isn't motivation or ability. It's the format of the ask.
Writing is the wrong interface for most people
Writing a post from scratch is a compound task. You're simultaneously deciding what to say, how to structure it, how to open, how to close, and whether any of it sounds like you. We covered in The Blank Page Problem Is a Myth why that cognitive load, not a lack of ideas, is what actually stops founders from publishing.
Talking removes most of that load. In conversation, you don't design a hook. You don't worry about formatting. You respond to a question, and the structure emerges from the answer because that's how explanation works when a real person asks you something.
There's also a raw throughput argument. Conversational speech runs at roughly 150 words per minute, while most adults type at around 40. Stanford researchers found speech input about three times faster than typing on a keyboard, and that study measured transcription, not composition. When you add the thinking, drafting and second-guessing that come with composing a post, the real-world gap is far wider. Ten minutes of talking produces more usable raw material than an hour at the keyboard for most non-writers.
But speed is the lesser benefit. The bigger one is quality.
Spoken answers contain the good stuff
Read a founder's unedited LinkedIn draft and then listen to them explain the same topic on a call. The difference is consistent and always in the same direction.
The written version is more careful, more general, and more boring. Writing for a public audience triggers a self-editing reflex: hedge the claim, remove the customer detail, sand off the opinion. What survives is professionally worded and forgettable.
The spoken version has the things that make content perform. The specific example. The number they'd never have thought to include. The honest "we got this wrong for a year." The offhand phrase that becomes the hook. People are more candid, more concrete and more themselves out loud, because conversation doesn't feel like publishing.
That candour is exactly the raw material AI needs. As we argued in Why AI LinkedIn Content Sounds Generic, a model prompted with "write a post about hiring" produces the average of LinkedIn. A model working from six minutes of you actually talking about the hire that went wrong has something no other account on the platform has: your specifics, in your phrasing.
Why interviews beat dictation
"Just record a voice memo" is the obvious version of this advice, and it mostly doesn't work. Handed a microphone and no structure, people ramble, trail off, or freeze exactly the way they do at the blank page. Dictation replaces one blank page with another.
What works is being interviewed. Questions do three jobs that dictation can't:
They pick the topic for you. "Tell me about a decision you'd make differently" beats "record your thoughts" because the first is answerable and the second is homework.
They pull for specifics. A good follow-up, "what did that cost you?", "what did the customer actually say?", extracts the detail a solo speaker skips because it feels obvious to them. The detail is never obvious to the audience. It's usually the whole post.
They create natural structure. An answer to a well-formed question has a beginning, a reason and a consequence. That maps onto a post far more cleanly than ten minutes of unprompted monologue.
This is the same reason podcast guests sound sharp and articulate: the host is doing the structural work. An interview-based content workflow gives every founder a host.
What this looks like as a weekly system
The practical version of talk-don't-type is a short, recurring session, ten to fifteen minutes, once a week, on the phone or at a desk, answering a handful of questions about the week: what happened, what you learned, what you disagreed with.
From one session like that, a competent process extracts two or three distinct post-worthy threads. A conversation naturally covers more ground than a single written draft, so the ratio isn't one session to one post, it's one session to a small batch. That changes the economics of consistency. Publishing twice a week stops requiring twice-weekly effort.
The session also fits where writing doesn't. Founders reliably find ten minutes to talk, on a walk, between meetings, in the car, who will never reliably find the ninety-minute block that writing demands. The format matches the calendar of the person doing it, which is most of what determines whether a content habit survives past week three.
For teams, the same logic scales through one operator. A marketer or internal ghostwriter can run short interview sessions across a whole leadership team in an afternoon, which is how a single person keeps five executive profiles active without writing fiction on anyone's behalf.
The objections, briefly
"I ramble when I talk." Everyone does. Rambling is a feature of raw material, not a flaw in the final post. The transcript isn't the post; it's the ore. Structure is the easy part to add afterwards. Specificity and voice are the hard parts, and talking is how you get them.
"My spoken English is too casual for LinkedIn." Casual is closer to the target than you think. The posts that die on LinkedIn are the over-formal ones. Your speaking register, lightly tidied, is almost always the right register.
"I don't have time for another meeting." It's ten minutes, and it replaces the writing time you're currently not finding anyway. The comparison isn't interview versus nothing-in-the-calendar. It's interview versus the post that never got written.
Where the tooling fits
None of this requires software in principle. A colleague with good questions and a willingness to shape transcripts into drafts gets you most of the way there, and some teams run exactly that.
What software changes is the marginal cost. Frankly Speaking runs the interview itself, by text or by voice, asks the follow-ups that pull for specifics, and turns one session into three drafts written through your voice profile, so they read like your answers rather than like AI. Drafts then go through your normal approval flow before anything publishes. The workflow is the point; the tool just removes the parts that needed a second person.
However you implement it, the underlying shift is the same. Stop asking yourself to write. Start asking yourself questions, out loud, and let the answers become the content. You already talk about your work with more clarity, specificity and personality than you'll ever manage at a keyboard. The blank page can't compete with a good question.
Ten minutes of talking. Three posts in your voice.
Frankly Speaking interviews you by text or voice, drafts posts through your personal voice profile, and routes them through approval before anything goes live.