To lead in the AI era, take a human-centred approach
As hype transforms into reality, organisations are embedding AI-powered tools into the fabric of their operations. How can AI tools such as Microsoft 365 Copilot be implemented in an ethical way while minimising risks and maximising benefits? The answer: by taking a human-centred approach.
Organisations across industries and sectors are increasingly turning to artificial intelligence (AI) to enhance productivity, boost innovation, and unlock hidden value. AI can empower workforces through significant benefits, from administrative support to enhancing internal sales tools. In fact, every $1 companies invest in AI generates an average return of $3.5. And AI deployments are happening fast – 92 percent of deployments take 12 months or less, and 40 percent within six months.
As AI tools revolutionise the workplace, many technology firms are developing their own AI-based solutions. One such solution is Microsoft 365 Copilot, which has far-reaching use cases from sales to customer service across over one billion Microsoft Office users. However, new applications bring new risks. AI tools are only as good as their users, and ‘good’ will differ across organisations. The task for leaders is not simply to make AI tools available, but to ensure their effective usage. This means adopting a human-centred approach.
Keep it simple
Clarity about the impact of AI on people and their roles – based on a strong workforce strategy, usage policy, and a clear implementation plan – is critical to build trust in AI tools (and their purpose) as they are rolled out. Employees may fear job displacement, or question transparency around AI-backed decision-making. Sharing a compelling vision about how AI will augment (not replace) humans and encouraging organisation-wide engagement, will help to mitigate the risk of AI tools being underused or misused.
Identifying a broad, diverse range of pilot users early can reassure employees that leaders are taking a bottom-up, practical approach to exploration. Two-way engagement and transparency about organisational plans and current capabilities also builds trust and buy-in. This is important for any digital transformation programme, particularly where the deployment of unfamiliar AI tools carry their own unique considerations.
At PA, we’ve adopted this approach while participating in Microsoft’s M365 Copilot exclusive Early Access Programme. Initially, our pilot group includes 300 internal users who come together each week to discuss usage, advantages, and challenges, and to share experiences to inform a wider rollout when the trial ends. Exceptionally high engagement demonstrates participants’ enthusiasm to understand how AI can positively impact their work. Alongside the Copilot trial, we’ve made Bing Chat Enterprise available to 4,000 employees, providing further insights on how real-world users interact with the technology. By identifying use cases in a safe environment, and stress-testing the journeys our clients will take, we can bolster our policies and guidance for users overall – using individual empowerment to achieve collective enhancement.
Identify risks and set robust governance up front
AI is designed to provide a response that best responds to requests, even if inaccurate. Risks associated with disinformation and inaccurate data only grow without sufficient mitigation measures. This may be as simple as building critical thinking skills, and frequent peer reviews of work. Flagging specific AI risks in advance, and setting up appropriate mitigations and governance, are critical to responsible adoption. It’s important that a cross-functional team designs and oversees development and change management, including individuals from the domains of technology, compliance and legal, HR, and finance.
When using AI tools in interactions with clients and external parties, an AI risk policy can mitigate potentially difficult situations such as potential IP infringement or unintended capture and use of personally identifiable information. Giving internal and external audiences sight of these policies will ensure transparency, build trust, and align with best practice, enabling iteration and improvement.
Upskill, with a focus on shifting mindsets
Understanding how AI tools like Microsoft 365 Copilot augment delivery in line with the wider organisation’s technology landscape is key to effective user adoption. Obtaining information on existing behaviours and capabilities helps to identify the most appropriate upskilling future talent management initiatives. Successful upskilling initiatives need to focus on shifting mindsets as well as skills to drive a culture of collaboration around AI tools. Understanding people’s starting position reveals the scale of change management required, as well as the depth and breadth of training.
To keep up with the accelerating pace of adoption, start small and scale fast. Setting up pilot groups and identifying specific use cases that complement existing tools and skills – as well as the organisation’s DNA – makes tasks simpler, faster, and less time-intensive. For example, we’re investing in custom large language models (LLMs) to better understand when to use AI embedded in everyday productivity tools, and when to extend its use.
AI tools are like a Swiss Army Knife – multi-purpose, with open-ended capability. But selecting the right tool relies on a clear understanding of the task at hand. Early piloting and guided hackathons can not only help identify which tasks new AI tools can be used most effectively for, but also help highlight skills gaps. For instance, the skill of prompt engineering – interacting with generative AI to retrieve the most effective answers – is critical to success. We’ve developed a prompt engineering course that is being rolled out to all employees to build the AI capabilities of our global network.
Leaders can proactively encourage knowledge-sharing through drop-in sessions and by identifying ‘super users’ who act as AI champions, empowering peers and taking those all-important first steps. For Copilot specifically, digital collaboration platforms like Microsoft Teams offer a forum to test and share the results of AI pilots with users who aren’t part of the initial trial, drumming up excitement for wider rollout.
Build the business case, measure value
Predicting the business impact of new and fast-moving AI transformation is challenging. But, by identifying key value drivers, and linking them to the business strategy, organisations can measure associated benefits from the deployment of AI. For example, we’re one of 1,000 organisations taking part in the global Sales Copilot trial. We’ve used Sales Copilot to transform how we use and input data around our opportunity and sales pipeline. We met with Microsoft data scientists to build capabilities to provide feedback, shape the product, and understand how we and our clients can benefit from AI-embedded tools. This enables us to measure and track value through custom dashboards, supporting the business case for adoption.
The question of how to deploy AI successfully and reap the rewards is front of mind for all organisations. Proactive, good management to ensure that knowledge is shared, innovation encouraged, and risks mitigated is as crucial as ever. With tools such as Microsoft Copilot, and generative AI applications like Chat GPT, this question is no longer theoretical. Making AI work in a practical way calls for a human-centred approach – grounded in a clear value-focused vision, robust governance, and accelerated piloting and upskilling.