Sona

Most people who cannot write can talk perfectly well. Ask a founder to explain why their pricing changed and you get five minutes of clear, specific, opinionated argument. Ask the same person to write a LinkedIn post about it and you get an empty document and a sense of dread.

Meet Sona is built on that observation. Instead of a prompt box, it runs a ten-minute guided voice interview — asking follow-up questions the way a journalist would — and turns the transcript into content across five formats.

This review covers why the interview format extracts things a prompt cannot, what Verbal Identity actually captures, how it compares with the stylometry approach Spiral takes, and the honest limits of talking your way to a publishing habit.

Interview In, A Week Of Posts Out

• The input is your voice, not a prompt. A ten-minute guided interview with follow-up questions, rather than a text box asking what you want to write about.
• Verbal Identity captures tone, phrasing and beliefs so output across formats reads as one person.
• One conversation, five formats: LinkedIn posts, blog articles, newsletters, X posts and short-form video scripts.
• Publishes to LinkedIn directly through the official API, with scheduling and analytics on higher tiers.
• Built by Sona Narrative Foundry LLC, with ElevenLabs credited on the voice side.

Why An Interview Beats A Prompt Box

This is a genuine insight rather than a repackaging, and it is worth understanding before judging the output.

A prompt requires you to already know what you want to say. That is the actual bottleneck — not the writing, the deciding. Most people staring at a blank post do not lack words; they lack a settled view of which of their several half-thoughts is the interesting one.

An interview solves that from the other direction. A follow-up question — 'why did that surprise you?', 'what would you tell someone about to make the same mistake?' — extracts the specific, concrete detail that makes writing worth reading. It is the same reason journalists interview people rather than emailing them a form.

The second effect is on voice. Spoken language carries rhythm, hesitation, emphasis and idiom that written-from-cold prose flattens out. Working from a transcript means the raw material already sounds like you before any style processing happens.

The trade is that you have to actually do the interview. Ten minutes of talking is a smaller commitment than an hour of writing, and it is a larger one than typing a sentence into a box — which is why this suits people who publish deliberately rather than people looking for a shortcut.

The blank page is rarely a writing problem. It is a deciding problem, and questions are better at solving it than prompts.

Verbal Identity, And How It Differs From Stylometry

Verbal Identity is described as capturing tone, style, beliefs and communication patterns so that output stays consistent across formats. That is a broader claim than it first appears, and it is worth comparing with the alternative approach in this market.

Spiral does the same job by stylometry — statistical measurement of your existing written work: sentence length, punctuation habits, vocabulary. It reproduces how you write, and it needs a body of writing to learn from.

Sona works from how you speak, and it includes beliefs and positions rather than only mechanics. That is a meaningful difference: it means the system holds something about what you think, not just how you phrase things.

Which is better depends entirely on your starting point. If you have years of published writing you are happy with, stylometry has more to work with and will match your written register more precisely. If you have little written output but plenty to say, the interview route is the only one of the two that can start at all.

There is also a subtler distinction. Spoken and written voice are not the same — most people are looser, funnier and more digressive out loud. Content built from speech tends to read warmer and less formal, which suits LinkedIn and newsletters better than it suits a technical white paper.

What One Interview Produces

The output side is where the economics work, and the Remix function is the mechanism.

• LinkedIn posts — the primary format, with publishing, scheduling and analytics built around it
• Blog articles — the long-form version of the same material
• Email newsletters — a different register again, more direct and less performative
• X posts — compressed, with the tension at the front
• Short-form video scripts — spoken material returning to spoken form, which is the most natural conversion of the five
• Remix — the same conversation re-angled, so one interview yields several distinct posts rather than one post in five costumes

The LinkedIn Integration Is The Practical Half

Generating content and publishing it are different problems, and most writing tools solve only the first.

Sona connects through LinkedIn's official API for posting, scheduling and performance analytics on higher tiers. The word 'official' matters — tools that automate LinkedIn through unofficial means put the account at risk, and anyone who has watched a colleague lose a profile to an automation tool takes this seriously.

Scheduling is what converts intention into a habit. The reason most people publish inconsistently is not that they run out of ideas; it is that the idea and the free moment never coincide. A week of scheduled posts from Tuesday's interview removes that coordination entirely.

Analytics closes the loop, and is the part that should change what you talk about next. If posts about hiring outperform posts about product by four to one, the next interview should go there — and without the data you are guessing.

Making The Ten Minutes Count

The tool rewards treating the interview as the work, which is precisely the opposite of how most people use content tools.

1. Do the interview somewhere you can speak freely and at normal volume. Half-whispering in an open-plan office produces guarded, generic answers.
2. Answer the follow-up questions properly rather than steering back to your prepared point — the follow-ups are where the specific detail lives.
3. Talk about one thing per interview. A ten-minute conversation covering four topics produces four shallow posts.
4. Read the first drafts for factual accuracy before voice. You said it out loud, so an error is yours, and it will be published under your name.
5. Use Remix to get several angles from one conversation rather than running more interviews than you need.
6. Schedule rather than publishing immediately, so the habit survives a busy week.
7. Check analytics monthly and let it choose your next interview topic instead of your mood.

Who This Will Not Suit

If you dislike hearing yourself speak or freeze when recording, the input mechanism is a barrier rather than a shortcut, and no amount of output quality compensates.

If your content is technical documentation, research writing or anything where precision beats personality, spoken-derived prose is the wrong register. Warmth is the point here, and it is a liability in a specification.

If you publish beyond LinkedIn as your main channel, note where the integration depth sits — LinkedIn gets publishing, scheduling and analytics; the other formats are drafts you take elsewhere.

And if you have nothing particular to say, this will surface that faster than a prompt-based tool would. An interview about a topic you have no view on produces ten minutes of hedging, and the drafts will show it. That is arguably a feature.