Being busy is not the same as being productive. That’s a cliche – but sometimes cliches are useful.
When I managed a team of people at a big news website, I knew what productivity meant.
It meant objectives and metrics which the team needed to meet (numbers of stories processed per shift, for instance) – and that they needed to be supported and encouraged and trained to meet those targets. (I also had my own metrics and goals.)
Over and above these measurables, there was an expectation of something less tangible: the quality of the work produced. Time came into it too: the meeting of deadlines, which meant, for instance, that a newsletter had to go out at 8am every weekday morning.
Now that I run my own small business (in which I am the only employee, though I do have fruitful collaborations with other people), the concept of productivity is less clear to me.
There are some deliverables
Some things are crystal clear: if a client wants a report done by the end of the month, then that deadline needs to be met. I must produce something of quality in order to get paid and maintain my reputation for reliability.
But on any given day, as I work to meet those deadlines and do all the other things that keep a business afloat, what does “being productive” actually look like?
I’m a big fan of Cal Newport, who has written a book called Slow Productivity. I read his newsletter whenever it drops in my email, and there’s always a lot there about productivity.
He says that because most knowledge workers have autonomy to control their workload, they are left to determine it themselves. They deal with this complexity by telling themselves the “workload fairy tale”: the idea that their current commitments and obligations represent the exact amount of work they need to be doing to succeed in their position. But, he says, experiments with a four-day work week show that the efforts that really matter require less than forty hours a week. “So why is everyone always so busy? Because in modern knowledge work we associate activity with usefulness… so we keep saying ‘yes,’ or inventing frenetic digital chores, until we’ve filled in every last minute of our workweek with action.”
This leads to what he calls “productivity rain dances” (borrowing a phrase from podcaster Chris Williamson): “When you… watch a solo-entrepreneur lose a morning to optimizing their ChatGPT-powered personalized assistant, you’re observing [a] rain dance.”
He says the solution is to turn your attention from inputs to outputs. “Identify the most valuable thing you do in your job, and then figure out what actually helps you do it better. This is what you should focus on.”
Which leads us to AI, and its impact (or not) on productivity
The productivity rain dance leads us to focus on the minutiae of what we do, and to wonder if Generative AI could make some of it go away, like a magic wand.
(Newport himself succumbed to the idea that AI might save him time and went on the hunt for a tool to help him sort out his overwhelmed email inbox. He found a tool that was able to sort his email – but he says could not find one that would actually answer his email for him.)
Having spent many months using Gen AI in a variety of ways, I think that yes, there are tools that will free up time, but alos that sometimes there are simpler ways to get the job done.
AI efficiency expert Tahnee Perry identifies the problem like this:
With all the hype around AI, it seems everyone is in a hurry to get more done, and churn out more content, regardless of the quality. I’ve caught myself slipping into that pattern at times. So I started looking for a way to work faster without letting my output slide into what Harvard Business Review calls workslop. That’s the low-quality content that looks productive but makes more work for everyone.
Her overarching guidelines, and some thoughts from me:
Human-in-the-loop is non-negotiable: There are a bunch of developments in the AI space at the moment that suggest that there are some things AI can do that only need a human to do the planning part. (Read Something Big Is Happening — matt shumer). For myself, and the work I do, I use AI as a tool in parts of the content production process, but do the thinking and writing bits myself.
Start small, measure everything: Don’t implement multiple AI tools at once, or hop from one to another without testing it to see if it really does the work you need done. Pick one thing you’d like to get off your plate and measure the time saved. Only then will you know if this works for you.
Protect the time you save: Don’t let the five hours saved by AI immediately fill with more admin tasks. Use it the time for learning or planning. Or block out some time to go to the beach (which is what I have done – see the picture at the top, taken on a Friday morning).
Invest in AI literacy: This field is changing all the time, and fast. Play with new tools, take all the free training you can find, ask other people how they use AI.
How to start thinking about your own productivity
At the risk of stealing all my ideas from Perry, I do like her steps for a productivity audit, which I reproduce here, along with an example of the steps I took in deciding how AI could help me with a small yet irritating problem.
Step 1. Identify the process: Pick a single workflow and define the start and end points.
My example: Meetings: setting them up, attending them and then remembering what happened, and actually doing the things that were decided.
Step 2. Map the current state: List every step and the tool you use for it. Note the time each step takes and the output you get.
An example of my problem, as it used to look: It’s agreed in an email that a meeting will be held with someone in Chicago and someone in New Delhi, to interview both of them at the same time for a report. I’m the one asking for the interview, so I am the one that has to send the meeting invitation.
So I hunt around for some way to figure out the time zone issues with this meeting, which can take up to half an hour. Then I set up the meeting and email the invite to the people involved. That email invitation puts the meeting on my Google calendar, but I also write it in my diary.
We have the meeting, and I take notes on paper, or perhaps on a notepad app on my computer. Two days later, I look at the incomprehensible notes and hate my previous self. I look for the transcript or video (more time wasted if I can’t remember where the default Zoom folder is) and reconstruct what was said – this could take an hour or more.
Step 3. Assess how well it works: Track which steps run smoothly and which ones force manual cleanup. This gives you a sense of where you can optimize.
For me: The things that go smoothly? The meeting itself, and the system for recording it in both my paper diary and my Google calendar (which is backed up with a morning protocol in which I cross check analog and digital calendars.) The things that take time? The time zone thing, and the post-meeting notes and action items.
4. Decide where AI fits: For each step ask three questions:
Question one: Can AI replace this step? On one hand, it could not help with the time zone thing: I did a bunch of research on an AI tool to help with calculating time zones (I did a rain dance!!) and finally concluded that a browser bookmark to this particular website would do the job: Savvy Time Zone Converter (you have to dodge the occasional advert, but you can put in three or four different time zones at once.) On the other hand, , for the notes and action items, yes. See next step
Question two: Can it support it? Over time I have experimented with a bunch of meeting assistants, and finally settled on one called Granola – it takes beautiful notes, does a transcript and I can email the notes to the meeting participants directly from the app. (Note that it mercifully does not insert itself into the meeting as an “assistant” but that does mean you need to remember to tell people it is taking notes.)
Question three: Should I eliminate it altogether? Not applicable in this case – meetings sadly just will not die.
5. Make a call and act on it: Keep what works, replace what doesn’t and remove anything that adds friction (that you feasibly can). My new meeting note system has been in place for several months, and it is working well. I reckon I’ve saved an hour or two per meeting by using AI – and by deciding that, for parts of the problem, there is a simpler solution that doesn’t need a shiny new AI tool.
Other ways I use AI save time:
I filter my email – strictly speaking this is not a use of AI; it’s just using Gmail’s inbuilt rules so that an email from a client goes into a dedicated place, and I can tend to it first. (There’s also an ongoing commitment to unsubscribing, on the fly, from anything that doesn’t serve my needs any longer.)
Using NotebookLM – a Google tool that’s indispensable for research.
Specialised AI tools – I make Custom GPTs and Gems for ongoing projects, so that the AI tool has all the information it needs, and I don’t have to take time inputting everything, every time.
How’s that working for me?
I’ve been tracking my time for years, and I know for certain that just these uses of AI have saved me five to eight hours a week. Some of that saved time is being spent on more careful planning of my day (in which I try consciously to focus on outputs rather than busyness) and some of it is being spent on taking more breaks in both my days and week – because better rested Renee means more productive Renee.
Main picture: Elaborate sand castle at Surfer’s Corner, Muizenberg, Cape Town on a Friday morning when I would have been at my desk if it weren’t for Gen AI.
Previous Sensible Guide articles to check out
Filter, filter, filter – the key to email organisation – a blog post from 2018 which probably needs updating, but it does have the basic steps needed to use Gmail’s inbuilt filtering system.
The AI tool everybody should be using – NotebookLM – Research assistant and note-taker rolled into one – allow me to introduce you to NotebookLM.
How to build yourself a virtual coach with AI (for free) – So you know how to use ChatGPT (or Claude, or Gemini). But you’re sure there’s more you could be doing. Here’s how to build yourself a specialed virtual tool with AI (for free).
If any of this is relevant to your work…
You might find one of these useful: