A People-First Blueprint for AI Implementation

An estimated nine out of ten AI pilot projects fail, according to MIT's Media Lab.

It’s a sobering statistic, but the reason often has little to do with technology. The core issue, and the central topic of a recent Welcome to the Jungle roundtable, is human. How do you lead a cultural transformation that ensures AI is not just implemented, but truly embraced?

To answer this, we brought together AI and HR expert and trainer Igaël Derrida with our own senior leaders: Kevin Le Roy (VP of AI & Transformation), Alice Hagger (Chief Brand & Marketing Officer), and Wendy Georges (VP of People). The consensus was clear: the most powerful AI tool is useless if your teams don’t fully adopt it.

Their insights form a blueprint for how HR and People leaders can architect this change successfully, leveraging AI as a performance lever rather than an expensive nice-to-have. 

The art of creating tools that you love to use

The single biggest barrier to AI adoption can be indifference. To overcome it, we have to shift our thinking from deploying a ‘Minimum Viable Product’ to creating a ‘Minimum Lovable Product’, according to Igaël Derrida.

This means looking beyond pure function. “Functionality is essential, but it’s no longer enough,” he explains. “We need tools that people enjoy using, tools that create an emotional connection from the very first use.” The best way to do that isn't with a top-down mandate, but with empathy.

“Before you start, I always recommend identifying a ‘pain’ – a specific problem – and focusing on solving only that problem, with care.”
Igaël Derrida, AI & HR Expert

As a People leader, your unique advantage is your deep understanding of your organisation's human operating system. Before any tool is commissioned, HR should be leading the discovery process, identifying the genuine pain points and daily frictions that AI could solve. True adoption always begins with genuine utility.

Driving cultural change: a look inside AI adoption at Welcome to the Jungle

Simply providing access to AI isn’t much of a strategy. At the roundtable, Kevin Le Roy, our VP of AI & Transformation, explained how this philosophy is put into practice at Welcome to the Jungle.

“An AI transformation isn’t just about having access to great AI tools… It’s about a cultural shift,” he says. While our teams can use a wide range of AIs, the real work is in fostering adoption and turning access into a genuine business transformation. To do this, we've implemented a professionalised, five-pillar approach that includes:

1. An AI Transformation team,

2. An AI Champions programme with managerial commitment and representation from each team,

3. An AI training plan,

4. Rigorous monitoring of the adoption process, and

5. The development of tailor-made solutions

A key part of this, Kevin stresses, is empowering our internal ambassadors. “You have to rely on your early adopters and accept that it may take time,” he explains. To create momentum, our teams are invited to a monthly, hour-long “AI-Power Up session” where anyone can share tests, agent creations, and new use cases.

The AI Champions are central to this. They aren't in the role indefinitely, ensuring more people can participate in the transformation from across the business. But most importantly, their primary skill isn't technical.

“These champions aren't necessarily innovators, but rather educators who understand business use cases. The ultimate goal is to get other teams on board.”
Kevin Le Roy, VP of AI & Transformation

A core part of driving this change is measuring what matters. Our approach to measurement has evolved as our own adoption has matured. We began with a simple, accessible KPI: defining monthly usage as any team member generating at least 5 messages in our primary AI tool.

As the organisation advanced, we needed more nuanced metrics to truly understand engagement. In 2025, we expanded our KPIs to track deeper, more active usage:

Weekly usage: Defined as activity on at least 8 days with more than 30 messages a month.

Daily usage: Our benchmark for true integration is set at a minimum of 12 active days and over 100 messages a month.

Creator culture: We also track the number of AI agents created per person and identify "intensive creators" who have built more than five agents.

Against these clear benchmarks, the results are compelling:
- 82% of our team achieves monthly usage.
- 56% are now classified as active weekly users.
- We have identified 27 intensive agent creators, with 48 people in total having created at least one agent.

The ambition with these five pillars is to make AI a daily reflex rather than just another tool (met with mixed enthusiasm).

Identify where you stand

Once you get a sense of your strategy, it helps to know your starting point. It’s not always easy to step back and get a clear view of your company's position on the AI adoption curve.

A practical tool we’ve found useful is the AI fluency scale from Zapier. It offers a well-designed framework to help you understand your organisation's current level of maturity.

According to this scale, there are four distinct levels:

Unacceptable: No consistent strategy or AI usage.

Capable: Pockets of AI use exist within certain teams for specific tasks.

Adoptive: AI is integrated into the standard workflows of multiple teams.

Transformative: AI is fundamentally changing how the business operates and creates value.

What makes this framework particularly useful is that it can be applied function by function, allowing for a true overview of AI applications across the entire organisation.
Once you’ve assessed your level, you can deep-dive into what is preventing you from moving to the next stage.

With great power comes great responsibility

AI as a mirror to your culture

Once you have a framework for implementation, the conversation should consider a more profound question: what is the impact of this technology on your brand, your culture, and your candidates?

At the roundtable, this was a central theme for Alice Hagger, our Chief Brand & Marketing Officer. She urged leaders to move beyond the technical and ask two foundational questions:

1. How will AI be used in our business in a way that is unique to us?
2. What will be our point of differentiation?

“The issue isn't AI itself. It's how we use it,” Hagger asserts. For a company whose mission is recruitment, this carries a profound responsibility. AI, she explained, will inevitably reveal what works, what doesn’t, and – most critically – a company's inherent biases.

“A poorly designed AI applied to recruitment doesn't just speed up the hiring process. It can reinforce exclusion, standardise profiles, and erase the very notion of human potential.”
Alice Hagger, Chief Brand & Marketing Officer

The uncomfortable truth, as Alice pointed out, is that many current automated sorting systems are not neutral. They favour conforming and already visible profiles. Unintentionally, they risk creating a world where career changers are filtered out, unconventional paths become invisible, and a candidate’s potential is buried under algorithmic assumptions.

The risk is that AI narrows the field of possibilities. But she is convinced that when implemented with intention, it can and should broaden it.

To counter this, Welcome to the Jungle developed a Code of Honour: seven principles that guide every internal decision regarding AI. This isn’t just a policy; it's our commitment.

Our AI Code of Honour:

Always prioritise the human. We're dealing with people, not profiles.

Create connections, not noise. Be there at the right moment.

Show what's real, even if it's imperfect. An honest story beats a polished lie.

Take technology seriously, without losing our soul. AI assists, but it never speaks for us.

Respect those we support. Every word and feature must be honestly reviewed.

Be useful. And remain free. Don’t follow trends that lead nowhere.

Never forget your mission. We believe in the power of human connection.

She concluded by urging every leader in the room to ask three essential questions of their own AI strategy:

- Do your tools help you find the right people, or just process applications faster?
Are you using AI to expand possibilities or to reinforce existing models?
- When you automate, are you creating more human connection, or less?
- The key, she argued, is to be intentional. To constantly question not just what the technology can do, but what it should do.

AI has a bias. And it's probably yours.

This conversation on ethics led directly to one of the most criticised aspects of AI: bias. Igaël picked up this thread, posing a challenging question to the room: if AI is biased, isn't it just reflecting our own biases back at us?

The answer, largely, is yes. AI learns from the data we give it. Igaël cited the well-known example from Amazon, where in 2018 it was discovered that their recruitment AI had become deeply biased, despite never being programmed to be so.

What happened? The model was trained on ten years' worth of CVs submitted to the company - a dataset that reflected the historic male dominance of the tech industry. From this data, the AI "learned" that male candidates were preferable. It began to automatically penalise CVs that included the word "women's" (as in "captain of the women's chess club") and downgraded graduates from two all-women's universities. It had taught itself to reject perfectly qualified candidates simply because they were women.

This isn't just an ethical issue. As Igaël explained, there are real performance, legal, and reputational consequences.

“Biased recruitment is economically irrational because it deprives you of the best talent.”
Igaël Derrida, AI & HR Expert

The culprit, he insisted, isn't the AI itself, but the way we design and govern it. It is our responsibility to proactively hunt down and mitigate bias. How? He outlined a four-step audit process:

1. Exploratory audit: Before you even begin, question the ethical relevance of the project. Should you be using AI for this at all?

2.  Data audit: Analyse the data you're using for inherent imbalances and underrepresentation.

3. Performance audit: Measure the model's accuracy across different demographic groups, not just as a whole.

4.  Post-deployment audit: Continuously monitor the model's behaviour to detect any drift or deviation once it's live.

This responsibility doesn't fall on a single team. It must be shared across IT, Legal, HR, and your DEI function. Finding the right balance isn't a magic formula; it's a direct reflection of your company's values and your strategy. It's up to you to define it.

People teams are in the driving seat

The conversation naturally turned to the function at the heart of this transformation: Human Resources / People. Wendy Georges, our VP of People, argued that HR’s role is to lead the charge in augmenting human capability, not replacing it.

For her, the entire strategy hinges on one core element:

“The key to successful AI adoption is team training. As an employer, we have a duty to build the Future of Work for our own teams. Enhancing their skills is a driver of growth, a powerful retention tool, and essential in a world where the half-life of a technical skill is now less than two years.”
Wendy Georges, VP of People

At Welcome to the Jungle, this is guided by our value that asks everyone, “What would it take to learn faster?” with a formal training programme built on three pillars:

1. A common foundation: This includes masterclasses on the fundamentals of AI, understanding the AI Act, and team-specific training on the art of prompting to develop advanced, role-specific skills.

2. AI Champions programme: An advanced curriculum designed to train champions in mastering agent creation, empowering them to drive transformation within their own teams.

3. Manager engagement programme: Focused on supporting leaders to manage change, foster a culture of experimentation, and rethink their leadership in the age of AI.

This commitment to continuous learning has a direct and powerful impact on our employer brand. As Wendy explained, “Nine out of ten candidates now ask about our use of AI. We need to convince them that we genuinely provide training and support. Ultimately, it’s a competitive advantage in an increasingly demanding recruitment market.”

Putting it into practice: how our own People team uses AI

This isn't just about enabling other teams. Wendy also shared several examples of how our own People team uses bespoke AI agents to augment their work:

TA Digest: An agent that creates a weekly newsletter summarising the latest recruitment market news, helping our Talent Acquisition team refine their sourcing strategies.

Interview scheduling: An assistant that automates interview organisation by creating communication templates and calendar events. 

My People Mate: An internal agent that allows any employee to ask questions and get instant answers based on our HR policies.

Development plan builder: An assistant that helps managers structure effective development or performance improvement plans for their team members.

Your blueprint for action

If the roundtable left us with one clear message, it's this: AI should be used to expand possibilities, not restrict them.

Here are five actions you can take right now based on the experiences and points of view shared in the discussion to bring back to your team and make the most of the opportunity AI presents.

◻️ Assess your adoption level: Get an honest view of where you are on the transformation curve.

◻️ Identify your AI champions: Look for the natural educators in each team, not just tech and AI enthusiasts.

◻️ Audit your current AI tools: Proactively hunt for and mitigate bias.

◻️ Define your own AI Code of Honour: Establish clear principles to guide your strategy.

◻️ Plan your 6-month training programme: Build a curriculum with a common foundation for everyone, an advanced track for champions, and dedicated support for managers.

If you're serious about investing in a successful AI adoption programme, don't forget to consider local regulations, including the EU AI Act.