The number one concern people have about AI-written newsletters is voice. "Will it sound like me?" The honest answer: it depends entirely on how well you define what "sounding like you" actually means.
Most people have never articulated their writing voice. They know it when they see it. They cannot describe it. That makes it impossible for any writer -- human or AI -- to replicate. Here is how to fix that.
What Voice Actually Means
Voice is not just tone. It is a combination of:
- Vocabulary: The words you naturally use and the ones you avoid. Do you say "leverage" or "use"? Do you say "folks" or "people" or "individuals"?
- Sentence structure: Short, punchy sentences? Long, flowing ones? A mix?
- Perspective: First person? Second person? Do you include personal anecdotes? Do you reference current events?
- Formality level: Academic and precise? Casual and conversational? Somewhere in between?
- Opinion strength: Do you state opinions as facts? Do you hedge with qualifiers? Do you acknowledge counterarguments?
- Humor: Dry wit? Self-deprecating? None at all?
- References: Industry jargon? Pop culture? Historical parallels? Sports metaphors?
The Voice Discovery Exercise
Take 30 minutes and answer these questions honestly:
Who are you writing for? Not demographics. Psychographics. What keeps them up at night? What do they aspire to? What do they already know? What do they need to hear that nobody else is telling them?
What three writers do you admire? Not because they are famous, but because something about their style resonates with how you want to come across. What specifically do you like about their writing?
Read five of your best emails, Slack messages, or social posts. The ones where you were being authentically you. What patterns do you notice? Sentence length? Humor? Directness? Use of questions?
What topics make you angry? Your strongest writing comes from conviction. The things that frustrate you about your industry are often the most compelling content for your audience.
What do people compliment about your communication? "You explain complex things simply." "You are really direct." "You always have a unique take." These compliments point to your natural strengths.
How Ted Learns Your Voice
When you onboard with WrittenByTed, we go through a version of this exercise on your kickoff call. But we also do something most ghostwriters cannot: we analyze patterns at scale.
Ted ingests your existing content -- blog posts, social media, email communications, presentations, even Slack messages if you share them. From this corpus, Ted builds a statistical model of your writing patterns: sentence length distribution, vocabulary frequency, structural preferences, opinion framing, and dozens of other micro-patterns that collectively constitute your "voice."
The result is not a template. It is a generative model that writes new content in your style. Not mimicry -- synthesis. Ted does not copy your sentences. Ted internalizes your patterns and produces original content that follows the same patterns.
The Feedback Loop
Voice matching improves over time. After each issue, your feedback -- "this paragraph sounds too formal," "I would never use that word," "add more of my personal experience" -- refines Ted's model. By issue five or six, most clients stop providing significant edits. By issue ten, they often cannot identify which sentences Ted wrote versus which they might have written themselves.
The Honest Limitations
Ted is very good at matching voice for informational and analytical content. Ted is less good at replicating deeply personal stories, highly emotional content, or humor that relies on specific personal experiences that Ted has not been exposed to. For these moments, the best approach is a hybrid: Ted writes the framework, you add the personal touches.
Most newsletter content is informational and analytical. For the 80-90% of content that falls in this category, Ted's voice matching is indistinguishable from the real thing.