AI & Technology
Stop Letting AI Write Your First Draft
What if the fastest AI summary you share today actually erodes the insight you need tomorrow?
2026-10-211 min read
The paradox of instant AI drafts is that they hide the very thinking that makes a recommendation valuable. An AI model can ingest a briefing, spit out a polished outline, and deliver it to a client within minutes, but the shortcut also skips the mental rehearsal that surfaces hidden assumptions. When a strategist spends the first half‑hour wrestling with the raw data, they discover contradictory trends that the model never questions because it merely mirrors the dominant narrative.
That friction is the crucible where novel angles are forged; without it, the final product feels safe but shallow. In a recent internal review, a consultancy’s senior partners noticed that proposals generated by a generative assistant consistently received lower follow‑up engagement, a symptom they traced back to the missing “question‑storm” that usually happens after the first draft. The remedy is not to abandon the assistant, but to rewire the workflow so that AI augments, not replaces, the initial sense‑making phase.
By forcing a brief pause between AI output and human refinement, teams retain the speed advantage while preserving the critical insight engine.
Key insights
AI drafts accelerate output but also compress the mental friction that uncovers hidden assumptions.
Insert a short, structured “question‑storm” pause after the AI output to recover that friction.
Why it matters
Ignoring the hidden thinking gap means future decisions will be built on unchecked premises, leading to costly revisions later.
Over‑reliance on AI drafts can dull the team’s ability to spot emerging risks that only surface through deliberate mental probing.
Use this tomorrow
1Open the latest AI‑generated briefing you plan to send and count the number of bullet points that you can immediately defend without further research; note any that feel “obvious” but lack evidence.
2After the count, spend ten minutes rewriting one “obvious” bullet in your own words, explicitly stating the underlying data source or assumption; if the rewrite reveals a missing piece, you’ve confirmed the gap.
Go deeper
The practice builds on the concept of “cognitive friction,” where deliberate difficulty forces deeper processing. By treating the AI’s first pass as a raw material rather than a finished product, you turn speed into a scaffolding for insight rather than a shortcut that skips it. This approach also aligns incentives: the faster the draft, the more valuable the subsequent human refinement becomes, keeping expertise in the loop.
The trade‑off is that the pause adds a tiny overhead, which can feel wasteful in high‑velocity environments. However, teams that embed the pause as a ritual find that the later stages of client interaction require fewer revisions, offsetting the initial time cost. Moreover, the habit cultivates a culture where AI is seen as a partner, not a replacement, preserving long‑term analytical depth.
