DEV Community

Cover image for When ChatGPT Becomes Part of the Daily Workflow: What 86M Minutes of Usage Tells Us
TMetric Timer
TMetric Timer

Posted on

When ChatGPT Becomes Part of the Daily Workflow: What 86M Minutes of Usage Tells Us

TMetric recently analyzed 86 million work minutes from ~10,000 professionals across roles (Marketing, Support, Sales, Development) to compare how much time people are spending in ChatGPT vs Google during their workday.

What We Found

  • In the “ChatGPT + Google” mix, ChatGPT’s share increased from 8.9% to 14.5% between June and August.
  • That equals about 14 minutes per day on ChatGPT (compared to roughly 8.5 minutes before).
  • Weekly, this totals roughly 47 minutes of extra ChatGPT time per employee.
  • This usage is above and beyond other work tasks — it doesn’t seem to significantly cut into other core responsibilities.

Role & Behavior Insights

  • Marketing leads adoption: ~18 min/day on ChatGPT (vs Google) — roughly double many other teams.
  • Surprisingly, E-commerce users compete, nearly matching Marketing in ChatGPT time.
  • Support, Sales also leverage it (e.g. for response drafting, summarization) though their balance is more mixed.
  • Developers are less aggressive adopters; use tends to focus on boilerplate, scaffolding, or quick helpers rather than core architecture or deep research.
  • Peak ChatGPT use happens in the morning (06:00–09:00). The “polish window” in the afternoon also sees secondary bursts.

What This Means for Devs & Teams

  • Don’t expect ChatGPT to replace your research stack overnight — it’s more of a complement to Google, especially for draft generation, summarization, ideation, and internal tooling.
  • Encourage small experiments: Dev teams can use it for code sketches, spec drafts, documentation scaffolding, error explanations, and stub generation.
  • Share prompt patterns internally — the more that’s codified, the more consistent and higher-quality the output.
  • Track usage and outcomes: Which prompts led to real-time savings vs which ones needed heavy post-edit?
  • Reassess periodically: As models evolve, what’s feasible today may shift quickly.

Read the full research 👉 https://blog.tmetric.com/chatgpt-vs-google-usage-by-tmetric/

Top comments (0)