Happy Triple Threat Thursday.
Here’s one Signal to notice, one thing to Spark growth and one Shift to consider.
This week's theme: AI writes to the lowest common denominator, so everyone sounds the same.
The model is trained to produce the safest, most universal version of any prompt. Same prompt, same output, no matter who is running it. Run "write a LinkedIn post about retention" through ChatGPT or Claude this morning, your competitor will get something almost identical tomorrow. The only thing that breaks the convergence is specifics from your business the average cannot generate.
📡 Signal — What’s Changing
What are readers actually detecting when they spot AI content?
Audiences can tell when content is just the average. The em-dash is not what gives it away.
What they are detecting is convergence: the same opening hook across three brands in their feed, the same three-option list with no recommendation, the same generic example ("a B2B SaaS company struggling with retention") that could fit any of those brands, the same closing line that summarizes without committing. The patterns get blamed for the convergence, though they are actually downstream of it.
This happens because AI is trained to produce the safest, most universally acceptable version of any prompt. The model averages every business writer in its training data and outputs something near the center. When several brands run similar prompts, the drafts converge. At scale, an entire category develops a shared voice, and no individual brand is recognizable inside it.
The market data confirms the trust cost. Klaviyo's 2026 consumer report across 8,000 consumers in seven countries found 31 percent trust a brand less for visible AI content. The most fluent AI users penalize the hardest. Klaviyo found 39 percent of self-identified AI enthusiasts trust brands less when customer-facing content reads as AI. People who write with AI every day are the quickest to spot a piece that did not commit.
Why it matters now: Trust degrades when readers feel like no one was paying attention. That feeling is what they call AI, whether the patterns are present or not. The convergence breaks when you add specifics from your business that the average cannot generate.
What to do this week: Read one piece you published last month and count the specific decisions in it. A named customer, a real number, a stance someone could push back on, a detail that could only have come from your business. If the count is zero, the patterns were never the problem.
⚡ Spark — What to Try This Week
How can you make AI content read like you wrote it on purpose?
AI removed the friction that used to force decisions. First drafts used to demand choices because the writer had to fill in the blanks themselves. AI fills them in by default, and the defaults are generic. The audit puts the choices back in.
Underline the specifics. Read the draft and mark every named person, real number, specific example, and clear stance. Any paragraph with nothing underlined needs rework before any copy-edit pass.
Replace generic with specific. "Many B2B companies struggle with retention" becomes "Three CEOs in my network told me their renewal rate dropped six points last quarter, and the fourth was up four points doing something different." The replacement adds conviction more than words.
Take a stance someone could disagree with, on a question that has two real sides. If everyone in the audience would nod along, the piece said nothing. Contestability is the threshold for content worth reading.
A note on the patterns: The em-dash, the rhythmic triple, the negative parallelism, the staccato three-sentence chunk, the significance hedge, the banned vocabulary list, and the wrap-up closer are still worth cleaning. Clean them in service of the noticing. Specifics with a few em-dashes will outperform clean copy that says nothing.
Time required: ten minutes per piece under 800 words.
🔄 Shift — How to Rethink It
Is your voice in your words, or in what you pay attention to?
Default belief: AI sounds like AI because of how it phrases things.
Flip: AI sounds like AI because it does not notice anything specific. Voice is mostly noticing.
I recently worked with a VP of Marketing of a $55MM eCommerce brand, who told me she spent five hours a week rewriting AI drafts to strip patterns. The output got cleaner while engagement stayed flat, so she decided to try a different exercise. Before any AI draft went out, she added one specific thing she had learned that week: a customer's exact words, a number from last quarter, a moment from a Tuesday meeting. Engagement moved even though the typical AI patterns were still in the drafts.
Why it matters: Teaching AI your vocabulary grooms the draft. Teaching it what to notice changes what the draft is about, which compounds across every piece you ever publish.
Before any AI session, write down one specific thing you noticed this week. A number, a quote, a pattern you saw across two customers. Feed it to the model as prompt context.
Make the AI draft commit. If it hedges, prompt back with "pick a side here." If it offers three balanced options, ask which one it recommends and why.
Build the voice doc later. The noticing comes first because nobody else can fake it.
The decision to publish stays human even when the draft does not.
📚 Worth A Look
What should you be reading about AI content and audience trust this week?
A survey of 8,000 consumers across seven countries on how AI in marketing affects brand trust. 31% trust a brand less when content reads as AI, and the heaviest AI users in the audience penalize the hardest.
A breakdown of why every AI model falls into the same corporate-helpful voice without specific direction, and a 25-minute checklist for getting a draft out of it before it ships.
A practical walkthrough on building a brand voice doc that AI will actually follow, with real prompt examples for uploading past work and asking the model to learn from it.
📈 TL;DR
Fix what's generic before you fix the em-dashes.
📈 One Question
If a competitor published your last LinkedIn post under their name, would anyone notice the difference?
Thanks for reading Triple Threat. See you next Thursday with another Signal, Spark, and Shift.
— Alexandria Ohlinger
p.s. If this helped you think sharper or move faster, share it with someone who builds the way you do. And if you want more practical insight between issues, connect with me on LinkedIn or schedule a strategy session.