Content how-to

How to use Claude for thought leadership.

AI-written thought leadership has destroyed credibility for many executives because the AI tells are obvious. Claude can produce real, voice-matched thought leadership — but only with a specific workflow. Here is what works and what destroys credibility.

The premise

Why AI thought leadership usually fails

When an executive's LinkedIn post sounds like ChatGPT, their personal credibility takes a permanent hit. Audiences can spot it and they assume the executive does not actually think this — they just published it.

The version that works requires the executive to provide the spark (the original insight, the contrarian take, the specific story) and use Claude to help shape and refine, not to generate from scratch.

The "20-80 split" rule

How much is too much AI

Executive provides 20% (the spark). The specific insight, the contrarian take, the real story. This is the executive's actual thinking.

AI does 80% (the shaping). Structure, sentence flow, examples, headlines. This is where AI accelerates without destroying authenticity.

When the ratio flips — AI provides 80% of the substance — credibility dies. The 20% spark is non-negotiable.

The thought-leadership prompt

Use this

I am writing a thought-leadership piece on [TOPIC].

My original insight (in plain language): [THE SPECIFIC THING I THINK]
Why this is contrarian or new: [SPECIFIC]
The specific story or example that proves my point: [REAL STORY]
My audience: [SPECIFIC PERSONA]
My voice (load from voice guide in knowledge base)

Shape this into a [LinkedIn post / blog / email].

Constraints:
- The original insight must come through clearly in the opening
- Do not soften or hedge my claims
- Include the specific story or example (it is what proves I am not making it up)
- Match the voice in the voice guide
- No AI tells ("in today's rapidly evolving landscape", "it is important to note", "leveraging")

Generate 2 variants with different angles. Tell me which is stronger and why.
The AI tells to hunt

Hard-kill list

"In today's rapidly evolving landscape" — universal AI opener

"It is important to note" — pure filler

"It is worth mentioning" — same as above

"Navigate the complexities" — meaningless

"Leverage", "synergize", "transform" — buzzwords that signal AI

"What do you think?" closes — engagement bait everyone recognizes

Generic 3-bullet middles — visual tell of AI structure

Related

Related how-tos

Want this set up for your execs?
Implementation includes thought-leadership workflow design.
See Implementation → Book the AI Audit