Listen to audio version
I ran a webinar last week to unpack my current thinking on AI and how we can leverage it to enhance our personal productivity. You can access a link to the recording here if you missed it. Obviously, this is a rapidly changing space, and my thinking is probably already out of date, but it provides a good starting point to formulate a plan for how we can get the most out of a technology that is changing the way we work.
One of the things I talked about was the three different types of AI capability that I see already at our disposal or coming down the track. If we can understand what each type is capable of, it becomes much easier to develop and plan for how to use it effectively.
It is important to note before I get into the detail that I am specifically talking about tools like Copilot and their application to our personal productivity. AI is being used by many organisations to enhance workforce productivity and can be harnessed to do everything from analysing data, to writing code, to synthesising information from multiple systems. This is not my space. I am interested in how tools like Copilot can help us to process our communications, organise our work, prioritise our tasks and plan our time.
Tell AI
In this context, the first type of AI capability is Tell. Most of us have access to this type of AI and are using it every day, whether we realise it or not. This basic type of AI can tell us information or make information more understandable. Anytime you do a Google search, AI is helping you find what you want. In Outlook, the appropriate Copilot licence allows you to summarise a complex email thread, or even prioritise your Inbox based on criteria you provide. Tell AI is really useful when it comes to sorting through lots of noise and complexity in our roles. This type of AI delivers immediate value while presenting relatively low risk.
Task AI
The next level of capability is Task AI. While many of us do not yet have access to these tools, emerging agentic AI capabilities are already being tested in organisations around the world and are likely to become more widely available in the near future. Because it is agentic AI, Copilot Cowork can go beyond providing information to actually taking action on your behalf. For instance, you could build an agent to monitor your Inbox for you, reducing noise while organising higher-priority emails into meaningful groups. Or you could ask it to clear space in your schedule, and it will provide you with options around what meetings could be cancelled, delegated or moved. It can then do the cancelling, delegating and moving for you once you approve.
This type of capability is very useful but carries more risk. We will need to think very carefully about letting AI prioritise for us, or interact with others on our behalf, even if we get the final approval.
Think AI
My prediction is that coming down the track very quickly will be the next generation of AI which can even go beyond doing what we tell it to do. Think AI may eventually be able to monitor our work, understand our patterns and preferences, and make increasingly autonomous decisions. For many, that sounds like the ultimate productivity solution, having a fully capable assistant doing the work and the thinking for us. Exciting, but can you see the increased risk here? This is where we will need to be sure of the AI’s competence, and that it truly knows and understands us, our role, our priorities, our preferences, and our stakeholders.
Right now, we don’t have to worry about that. But we do need to put some foundation thinking in place to ensure when we get to that future, we are ready to remain in control. We need to be clear about how much we are willing to give away to tools like Copilot.
For me, Processing, Planning and Prioritising are the higher order personal productivity skills that I refuse to give over totally to AI, no matter how many efficiencies can be gained. I am constantly looking for ways to engage AI to do the lower order work, but my effectiveness depends on me holding on to certain disciplines even when the technology is so inviting.
Rather than asking, “What can AI do?”, start by asking, “What work should AI do?”.
The answer may be different for each of us, but being intentional about where we use AI and where we retain control will become one of the most important productivity decisions we make. AI can help, but it should not be a replacement for good work practices or habits.n and see if there are opportunities to collapse the time involved by leveraging concurrency?