We are hiring! If you'd like to join our awesome team, check out our open roles
Insights

Teamwork in the era of Human-AI Teams

With the recent hype around artificial intelligence and LLMs in particular, there's a disproportionate focus on the tools -- there is little focus on how this will impact teams.

This post explores how:  

  • how teams can start to think about AI in a team context
  • common challenges you might encounter
  • actionable ways for you to establish a shared understanding

AI and teamwork

By now, we all understand the promise of AI – it has the potential to save us time, help us be more productive, and enable us to focus on more meaningful and higher value work.

But where are we in terms of how we’re using AI at work?

In a recent Microsoft WorkLab’s report (May 8th, 2024):

  • 75% of knowledge workers use AI at work today
  • 46% of users started using it less than six months ago

The results are clear, in users saying AI helps them save time (90%), focus on their most important work (85%), be more creative (84%), and enjoy their work more (83%).

In contrast, in a report from 2023 from McKinsey, 79% say their organisation has not yet set Generative AI policies.

This disparity represents a substantial gap – on one hand there’s a lot of individual usage of Generative AI tools and technology, on the other hand, organisations are not yet providing their employees guidance on safe use of generative AI at work.

This gap puts the responsibility on teams to navigate a highly ambiguous and rapidly changing landscape, which has the potential to introduce potential challenges to team dynamics and ways of working.

The era of the Human-AI team (HATs)

Researchers have described the next phase of teamwork and AI as the era of the Human-AI team (oftentimes abbreviated to “HAT”), where humans and AI agents work more interdependently toward common goals.

Henrik Kniberg theorised that this might look like smaller teams, potentially as few as 2 human team members. Smaller teams equipped with prompt engineering skills, collaborating with an AI agent, to deliver much faster than all-human cross-functional teams (which tracks with what XP originally intended with pair and ensemble programming, with faster feedback loops).

Realistically, no one actually knows how we’re going to get there!

Navigating the uncertainty of AI and teams

This topic is rife with unknowns, which are leading to a real sense of uncertainty from every level – from organisations, to teams, to individuals.

Questions that are still unknown include everything from how quickly AI will evolve in the future, to how AI will change the nature of the work that we do, to how AI will impact our roles and what teams will look like in the future.

Until those questions are better understood, it still leaves ample challenges to teams today, including:

  • The use of Shadow AI: without clear organisational policies, people are using AI anyway, which leads to…
  • Uneven distribution of benefits: where AI is only benefiting individual productivity, which isn’t being passed on to the team or organisation…
  • Reduced capability to innovate: in this common scenario, learnings aren’t shared, which means skills don’t grow (and the potential to innovate isn’t being met)
  • Unclear ways of working, which leads to team tensions: where the lack of clarity of how teams can collectively use AI introduces tensions (which are experienced through their ways of working)
  • Risks aren’t known and visible: and more importantly, when use of AI isn’t transparent, that means risks aren’t visible and adequately managed, whether security, privacy, mitigating bias

Ethan Mollick recently published an insightful book titled Co-Intelligence: Living and Working with AI. In that book, he shares a fundamental truth about our collective future with AI: 

There is no single manual or instruction book that you can refer to in order to understand AI’s value and its limits. 

Ground rules for Teamwork and AI

Until there’s better understanding of the unknowns of AI, we need to embrace that teamwork with AI is emergent.

This means that there is no “best practice” with AI and teams. We won’t know until we start applying AI to our context: organisations, teams and our day-to-day work.

The best way to navigate this uncertainty is with our teams using the following ground rules:

  • Use AI (in the open): the simplest thing you can do is to start being transparent about when you’re using AI, whether it’s different use cases, learnings, failures – move your use of AI from the shadows into the open so that your team can learn from it
  • Be intentional in building AI capability: reframe AI with your team from “how do we use AI more” to “how might we build our team’s AI capability for the future” – not only are there benefits to your organisation, but growing AI skills will also benefit each individual team member
  • Explore the unknown through experimentation: embrace unknowns as a team and discover AI’s value and limitations through testing, learning, and adapting as a team; in emergent domains, experimentation is a necessity to tackle unknowns
  • Double down on your team to co-create the future: when faced with uncertainty, it’s a natural instinct to wait to see – but the rapidly accelerating capability of AI means that it’s more important than ever to lean into teamwork to build what the future looks like together

What you can do about it

While there is no instruction manual, you can start bringing AI more into the open, create a clear set of working agreements for how your team uses AI.

AI Working Agreements Workshop Template on Miro

We have created the AI Working Agreements Workshop to give you a starting point. It’s an essential starting point to create clarity, establish norms and set expectations for how your team incorporates AI into your day-to-day work, including topics like: 

  • What AI tools are we using (and what's supported by a company policy)?  
  • What tasks are we using AI for (which tasks have a best practice and which need further experimentation)?
  • What are we optimising for when we use AI?
  • How and when will we experiment and share learnings around our use of AI?
  • How do we use AI safely and what do we need to make sure we don't compromise?

To learn more on how to run this with your team, check out our free workshop template on the Miroverse.

Summary

While AI introduces complexity and challenges, it also presents a unique chance to redefine teamwork. Let’s embrace this opportunity to build a future where AI and teams work seamlessly to create greater value together.

This blog post is adapted from a presentation that was delivered at Atlassian Team 24.