Successful AI implementation should be as human as humanly possible

As a tech professional, the last three years have been fascinating – I started out in IT 40 years ago, and this change is unmatched. We face the steepest transformation of a generation. It is on us to ensure we manage this in a way that prioritizes people and drives the world economy forward.

The IMF states that 60% of jobs are exposed to AI-driven change, and the World Economic Forum indicates that 44% of core skills will change in the next five years. The challenge for all of us as leaders is to ensure AI is adopted in ways that protect worker integrity and inclusion, ensuring no one is left behind.

This is a defining moment. Success requires strategic collaboration, a redesign of workflows to account for new digital workers, and guardrails to ensure we implement technology responsibly. Success depends on human-centric implementation.

How I look at it is that the CEO, CHRO (chief human resources officer) and CDIO (chief digital information officer) need to join forces. We read earlier in the year about a combined CIO and CHRO role for the first time – this shows how organizations are shifting. But, rather than necessarily combining the roles, it is more about fostering a coherent collaboration between IT and personnel. Our Business Leaders research of 2,500 C-suite professionals showed that only 53% had an aligned AI strategy. From our work with around 100,000 organizations, we know that the tech implementation strategy is only half the story. Ensuring that organizations have a clear plan for upskilling and change management is of equal importance. Workers demand support to navigate this disruption.

It is on all leaders to become deeply curious about AI – not just those of us in IT and tech. It is about meeting people where they are. For our people, we have developed more than 760 hours of AI training from the basics up to training for AI experts. And we have partnered with Microsoft’s philanthropic arm to offer free AI training for all our candidates.

Successful implementation requires us to unlearn how we have implemented tech before and consider how we will interact with AI as part of new workflows. AI requires experimentation, iteration and getting used to solutions proposed by technology. This workplace reinvention is a team sport – our IT teams will need to get even deeper into workflows.

So how do we make this AI-human interaction work? GenAI needs human commands, and with agentic AI we can delegate outcomes and it will iterate towards them. For this to work, we need to design workflows with clearly defined roles – where AI complements human strengths. The question is: What can only humans do and what can be enhanced by AI? And, to answer this, you need the deep domain experts to be part of the design process.

We all read about the impact of AI on the recruiting process. We want to turn this into an opportunity. Agentic AI can free up recruiters to do what brought them into this career in the first place: engage with candidates. There is a lot of heavy lifting to match people to roles at scale – millions every day.

So, we worked with our recruiters to work out what are the repetitive and more administrative steps in the workflows and have designed agents accordingly. Given that each job opening attracts an average of 250 increasingly AI-generated applications, we are building up capabilities that allow for AI-enabled recruitment processes that effectively match the right candidates for the right job. We have rolled out pre-screening, onboarding and redeployment agents. Agents source, score, follow up and shortlist ... And even match candidates to open roles when an assignment is ending.

The results are strong: Candidates rate the experience around 4.5/5 and are engaging 50% of the time outside the normal working hours kept by human recruiters. And, as their time is freed up, our recruiters are happy to be free to support candidates in richer human-to-human contact. It is about designing the workflows based on deep understanding of the desired outcome and business impact, while ensuring the technology is augmenting human experience.

As autonomy rises, so must oversight. With AI becoming agentic, the stakes rise – you cannot audit every step, so you need to audit intent, constraints, ethics and values. Without strict governance, we will lose trust. We need to ensure our stakeholders have confidence in the decisions we are taking. Our research showed only 34% of companies had a clear AI policy. Companies need to adopt responsible AI policies that ensure strong human-in-the-loop oversight, ethics and data protection are in place.

We have a defining leadership moment – we must act now to introduce AI for the good of humanity. I have always been an optimist, and I am excited about the opportunities this intelligent era brings. But the burden on global leaders is high. We are making decisions that will impact the future of humanity. We must join forces and be rigorous on governance. The future of AI will be shaped not by what machines can do, but by what people can become.