Could AI help solve inclusion challenges?

Could AI help solve inclusion challenges?

By Jo Faragher on 29 January 2026

One of the biggest hurdles in building everyday inclusion is supporting managers and staff to overcome their fear of saying the wrong thing. The intention is positive and inclusive, but often conversations go unspoken or feedback is not given because people from different groups or backgrounds don’t feel comfortable. So imagine a tool that could help someone practice that conversation without fear of failure, or that could offer feedback on a difficult email before it was sent. Increasingly, companies are turning to AI to help them with these inclusion challenges.

There are a number of scenarios where an AI tool could support an employee or manager in this way. For example, there could be a performance conversation where an employee is highly skilled but often quiet in meetings, and the manager wants to encourage participation without making assumptions about their culture or thought processes. Rather than simply saying ‘You need to speak up more’, an AI tool might suggest a more nuanced approach, asking why that person shares insights one-on-one after the meeting and what might make it easier for them to participate. And while there are a number of specialist AI tools coming onto the market that can help organisations with such scenarios, some of this functionality is already embedded in workplace systems such as Microsoft (Co-Pilot) and Google (Gemini).

Not without risk

Mary-Clare Race, CEO of Talking Talent, believes there is a role for AI in helping people find the language they need in difficult conversations, but adds some caveats. “They need to have a level of training to ask the right questions and use the right prompts. Leaders are often working at such a fast pace that AI can help them prepare a conversation,” she explains. “But AI might also give you the answer you want to hear. It really depends on how you ask and the product you use.” Talking Talent coaching clients can access AI to support their progress, but this is built on the company’s own content and context, with suggestions made based on real-life interactions with the user’s coach. In a coaching context, AI can also support simple nudges such as prompting someone to book a meeting, or complete their journal, making the process more efficient.

“We’ve really found that clients have related to it,” adds Race. “But we’ve been really intentional about the role of the human, as no algorithm can replace human empathy, curiosity, or body language for example. But it can enhance the coaching process. For example if you have an ‘aha’ moment one week, it can give you a prompt about it later so you don’t forget. It’s a companion, a support on the journey,” she adds. However, as organisations build out the use of these tools, they also need to ensure they are intentional about how they are used, she advises, for example promoting in-person feedback and collaboration between new recruits.

Closed insights

John Kolm, CEO of Team Results USA, has developed an AI assistant called Dorothy that mines insights from a proprietary database, sourced from a panel of 70 experts who were interviewed about life in the American workplace. Rather than using data from the wider internet, Dorothy uses ChatGPT to mine a closed set of insights and is also governed by a set of around 25 rules that “keep her professional, on-subject and free of hallucinations” (false or misleading outputs from an AI prompt). “If we accept that a more online and remote work environment is inevitable for at least some workers, AI holds the promise of mitigating the human downside by offering human strategies,” he says.

“Dorothy was designed for quick and practical answers like ‘How do I get noticed at meetings?’, but also to have soulful conversations with people in which they often pour their hearts out,” explains Kolm. “Her great strength in this mode is that underlying the emotional connection is half a million words-plus of accurate, reliable, properly provenanced information from eminently qualified people whose names and references she provides. This is a huge difference from AI systems like vanilla ChatGPT that simply scrape the internet and report whatever generic information they find.”

Assist rather than substitute

“The most effective uses of generative AI in this space are deliberately assistive, not substitutive,” agrees Anne DeAngelis, global head of employee engagement at PR company Zeno Group. “Organisations using AI well are clear about one thing: it is a tool to support managers, not replace them, particularly in moments where efforts to foster inclusion tend to break down. One of the most effective uses is coaching and preparation. Leaders can use generative AI as a private, low-stakes practice space to rehearse difficult conversations, including performance feedback, wellbeing check-ins, workplace adjustments, or career development discussions. For managers who worry about saying the wrong thing, that practice builds confidence. Inclusion comes from better human conversations, not from AI making decisions.”

AI can also be used for cross-cultural and language support, she adds: “In global teams, AI can help draft clearer messages, translate intent more accurately, and reduce misunderstandings that can occur when the team member’s first language or cultural norms differ from the leader communicating. Used well, it creates more equitable understanding and participation.” For leaders, it can be used as a perspective-checking tool ahead of major communications, or a way to see how that message might land with different groups such as new starters or teams in different regions. “That’s not outsourcing judgement; it’s recognising that every human has blind spots and using AI to broaden perspective before communicating at scale,” says DeAngelis. “As always, it’s worth checking with a human member of the relevant employee group for direct feedback, as well.”

Human oversight

Zeno’s AI Culture Change Center focuses on helping organisations to be transparent on how their AI tools are governed, communicated and embedded. Without transparency, the risks associated with AI can be higher. She adds: “When AI is introduced without explanation, people often assume the worst: that they are being monitored, assessed, or categorised in ways they cannot see or challenge. That perception alone can undermine inclusion.” Furthermore, human insight is non-negotiable – after all, generative AI does not understand power dynamics, cultural nuance or the history behind a situation. “Fostering an inclusive workplace requires empathy, judgement, and accountability, none of which can be automated. The responsibility must always sit with leaders, not algorithms,” insists DeAngelis.

Keeping up with the constant technical updates and changes to generative AI tools may feel like an impossible task, but this is not where the important work lies for D&I professionals. Ensuring strong governance, transparency and that humans stay firmly in the loop are guiding principles for AI’s role in inclusion.