Breaking the 3% Barrier: Time advantages compound…
Over the last few weeks, I've been writing about an interesting disconnect in how we use Large Language Models (LLMs): They offer tremendous potential to transform our work - yet the measured personal productivity gains so far have been modest, often just 3-5%, or 10-20 minutes saved per day. That's not nothing, but it's far from life-changing.
This gap likely stems from several challenges - some structural (like organizational resistance in larger companies), some technical (LLMs lacking clear manuals or workflows), and some personal (not knowing exactly what we need help with, or not focusing on tasks with the most value).
I already shared one practical idea last week about getting more personal value from LLMs: Move away from a command mindset toward mini-projects
Building on this, here's another practical idea: Let the LLM take over tasks where it does a better or faster job than you - because time advantages compound
LLMs excel particularly where speed matters and 60% or "okay" output quality is sufficient, at least to get started. Remember that in project work, speed advantages compound over time. If you can speed up a project by 20 minutes per day over three months, you gain 20 hours or 2-3 work days - time you can use to either finish earlier or put more effort into improving the outcomes.
Here are several use cases where LLMs have an edge, at least for getting started. perhaps this is useful:
Generating a First Draft
Feed the LLM your unstructured thinking from your notes app or emails. Ask your colleagues to share their notes too, or create a joint document in the cloud. The LLM can structure these ideas, highlight key decisions, and outline next steps. The more context you provide, the better - so consider setting up dedicated chats (ChatGPT, Gemini) or projects (Claude) for your most important meetings, returning to them after each session when you need support.
Basic Concepts and Structure
For new topics at work, ask your LLM to provide quick context that you can verify and expand upon. For example, summarize basic concepts or "rules of the game" for an industry, outline business challenges or key success factors, map out customer requirements or a typical customer journey, or compile industry terminology. While you should review and add nuances to the output, it's much easier to rework a first draft from your LLM than start with a blank page.
Structuring Your Own Thinking
Sometimes we're not entirely clear on what we're thinking - that's where feeding your notes, longer emails, or rough workshop ideas into your LLM can help. Let it know you're still uncertain, and ask it to structure your thoughts, identify key themes, and suggest next steps. LLMs don't judge, so share your concerns or doubts freely. They typically respond with solid ideas for sharpening your thinking, testing it, and moving forward.
Here's an example prompt sequence:
"Hey LLM, I'm sharing extensive meeting notes from the last few weeks. We want to manage these meetings more programmatically - setting agendas earlier, following up on to-dos more consistently. Please review the notes and identify key themes and challenges."
"My hypothesis is that our work could benefit from setting a few goals to work toward long-term. Based on the notes, can you suggest 5 potential long-term goals? Please explain why each would make sense."
"Next, let's define key priorities to reach these goals. Based on my experience, [these] are some crucial next steps, as they allow us to do [this]. Please review the minutes and identify other priorities we should consider."
"Can you draft a 3-month timeline to achieve these priorities (which should be sufficient based on my experience)? Please break down tasks and activities for discussion in the next three weekly meetings."
Drawing Parallels and Getting Inspired by Studies
LLMs excel at quickly summarizing studies and research. Even if you don't typically read studies for work, they can provide great reference points and inspiration. Start with a good briefing on why you're seeking inspiration, then find quality studies via Google from reputable sources like research institutes, top-tier consultants, or investment banks. Try searching "[subject matter] studies from [McKinsey, BCG, Bain, Goldman Sachs, UBS]". Alternatively, use Perplexity.ai for research, ensuring high-quality sources (as it sometimes returns less rigorous sources like blogs or Reddit). Then ask the LLM to review the studies, summarize key takeaways, and identify parallels or inspiration for your current challenges.
Preparing for Pushback or Challenges
Your LLM can help challenge your thinking and anticipate potential opposition to your ideas. Ask it to compile a Q&A based on your emails or memos, predicting questions or challenges that might come up, and prepare good responses. If you are not happy with the answers you can come up with to these questions, you should revisit your thinking. Remember though that LLMs aren't great at traditional reasoning, so they can't easily challenge your logic directly.
In summary, there's significant value in being more strategic about which roles you and your LLM play best when working together. LLMs generate "good enough" thinking with usually much lower time investment. Just remember that the initial prompt needs to be precise and thorough - see some tips in this post.
Also, as always, I'm interested in your thoughts: Feel free to message me. Or, if you prefer, you can share your feedback anonymously via this form.
All opinions are my own. Please be mindful of your company's rules for AI tools and use good judgment when dealing with sensitive data.