AI Copilots: Catalyst for true employee involvement in AI development
- Posted on January 23, 2024
- Estimated reading time 5 minutes
Leaving the beach and returning to work for 2024 I am reflecting on 2023. I remember in the first half of the year wondering whether the rate of change in the world of work would plateau after the incredible spike post Covid. Since then, with LLM based AI, and more specifically many user friendly ‘Copilots’ hitting the market, the pace of change continues.
Having worked with AI in the Employee Experience space for over a decade, I have seen the incredible potential and impact of this technology across various domains and industries. From reimagining learning, onboarding or knowledge management, to enhancing customer experience delivery, supporting wellbeing and resilience, AI used well can support us to create new insights, solve complex problems and optimise entire processes.
But despite the availability and maturity of various AI solutions, uptake has sometimes been ‘in the background’, niche and specific. Now I have seen an explosion of interest in one specific tool in the last six months: Microsoft Copilot, particularly M365 Copilot.
But as research has been trying to tell us for multiple years now, you can’t just drop new technology into an organisation and expect to get full value from it. You need to set the organisation up for success – considering everything from operating model to attitudes, behaviours and skills, data shape and aligned processes.
We have been working with multiple clients across multiple countries over the last six months, helping them to maximise their use of Copilot. From this experience we have identified several themes and areas organisations need to get right in order to maximise the value they get from this pop star of a tool.
Why the momentum with M365 Copilot?
Copilot is probably exploding on to the scene because it is a relatively safe, pre-packaged but flexible version of a LLM AI tool that can provide value ‘in the flow of work’ using apps many employees already use - in different ways that work for different types of employees. Suddenly it’s not just the Digital innovation team or the curious experimenter using AI.
What themes are we seeing from working with clients to maximise value from M365 Copilot?
From our support of clients in multiple countries, we are seeing some common themes:
- Most are seeing a huge potential for value, both in terms of productivity and the ‘enjoyment of work’ by removing some of the more manual, process heavy elements of work
- Many want to work through issues of ‘trust’ – everything from trusting that the wrong data is not accessed (eg confidential files, people’s pay or age), or that the tools are not being used for the wrong reasons, or that the LLM model is not ‘hallucinating’ an answer
- Most recognise quickly that it takes time to get ‘good’ at using copilot, partly because there are millions of potential ways to use it. This includes core skills such as getting good at prompts, to seeing (through experimentation and sharing) ways to connect multiple apps together in sequences (eg how do I combine Teams, OneNote, Word and Outlook together to maximise meeting outcomes?), to actually getting into the habit of using copilot at the right times, every day
So, what do organisations need to do to maximise value from Copilots?
If an organisation is wanting to move uplift from ‘small’ or ‘good’ to ‘great’ there are a number of common things they need to get right:
- Shaping data and access so that only the right data is used, but as much of it as possible in order to maximise utility
- Identifying and prioritising use cases – to help different types of workers find value from copilot, but also to help the organisation work out who should have Copilot licenses
- Calculating value and business cases – to help them work out how far they should go with their initial Copilot experiments and broader rollouts
- Helping employees to get the most from Copilot – not just ‘how to prompt’, but actual cultural initiatives to build trust and encourage and celebrate curiosity, innovation, experimentation, resilience and sharing when it comes to use of copilot across multiple apps
- Planning for scale up from a very small number (eg IT testers) to a ‘pilot’ (eg 300 business users) to organisation wide usage (for example, how do we do adoption for 20,000 employees?)
- Setting up ongoing measurement, processes and rhythms for ‘continuous improvement’, envisioning and building ‘next level’ opportunities.
Where to from here? How could this create an AI renaissance?
From the momentum I am seeing, there is little doubt that copilots (across many platforms) will continue to build momentum. They will become an expectation for knowledge workers and many ‘frontline’ workers in modern corporates. They will improve productivity and experience for many roles, but they are unlikely to fundamentally transform businesses on their own.
The real question is whether the popularity of these copilots can be used by organisations to fuel a surge in trust, interest, ideation and uptake of future AI opportunities. The sort of opportunities that can transform entire processes, functions, businesses.
Now that a broader range of employees have a tangible example of the sorts of things that can be done, I am seeing them suddenly ‘get’ the potential, the risks, the need for them to be actively involved in identifying opportunities whilst making sure AI is not ‘mis-used’.
This is the time for organisations to tap into that ‘get’. There is an opportunity to combine existing metadata within the business and qualitative sessions with business representatives to engage with employees, find practical genius and build trust that things are being done the right way – by involving them in the process.
Done well, a range of employees can be supported to identify, ideate, prioritise and de-risk serious AI opportunities. Furthermore, if the approach is right, these can be seamlessly fed in to a ‘factory’ to triage ideas, formalise business cases, undertake responsibility, risk and feasibility assessment and develop solutions. With this process in place, solutions will be more practical, testing and uptake will be accelerated, and trust will be present throughout.
In a world where we all know AI will create disruption - both good and bad - today’s Copilots do more than enhance day to day experience. They bring employees into the world of AI potential – to be part of the solution rather than nervous observers.