Latest / Future of Work Tech with Fexingo: Remote Tools, AI Productivity, and Workplace Software / How AI Is Reshaping Your Internal Job Market
Transcript
- Lucas: You've been at a company for three years. You're good at your job, maybe a little bored. Then one morning you get a Slack notification: 'Based on your skills and career interests, you might be a strong fit for a senior product role in the Singapore office — apply internally by Friday.' Luna: That's not a recruiter reaching out. That's an algorithm, right? Lucas: Exactly. It's an internal talent marketplace — software that uses AI to match current employees with roles, projects, and mentors inside their own company. And it's one of the fastest-growing categories in HR tech right now. Luna: I've heard the term, but I didn't realize it was that automated. So the AI is basically scraping your resume, your performance reviews, maybe even your Slack messages? Lucas: It depends on the platform, but yes — these systems build a skills profile on you using data from your HR records, past projects, training completions, sometimes even the language you use in emails. Then they surface opportunities you might never have found on your own. Luna: That sounds both empowering and a little creepy. Let's talk about a real example. Who's actually doing this well? Lucas: Unilever is probably the most cited case. They launched their internal marketplace back in 2022, but they've been iterating on it ever since. As of their last public update, the system had processed over 50,000 internal moves — that includes permanent transfers, short-term gigs, even mentoring assignments. Luna: Fifty thousand — that's massive. And they're a company of about 128,000 employees, so that's nearly 40 percent of their workforce moving internally through the platform. Lucas: Right. And the result they've reported is that divisions using the marketplace heavily reduced external hiring costs by about 30 percent. Because instead of posting a job on LinkedIn and sifting through hundreds of applicants, they find someone already inside who has the skills and the cultural fit. Luna: So it's a retention tool too. If people see a path forward without leaving the company, they're less likely to jump ship. Lucas: Exactly. The cost of replacing a salaried employee is typically six to nine months of their salary. If you can keep that person engaged by moving them into a new role, you save real money. And the employee gets a career boost without the risk of starting over somewhere new. Luna: But how does the AI actually know what skills someone has? I mean, my resume says I know 'project management,' but that could mean anything from running a small team to managing a multi million dollar product launch. Lucas: That's the challenge. Most platforms use what's called a skills taxonomy — a standardized list of skills with definitions and proficiency levels. When you join, you might self-assess, or your manager rates you, or the system infers skills from your past projects. Some platforms even use natural language processing to analyze your writing and meetings to detect skills you didn't list. Luna: So it's watching what you do, not just what you say. That's where the privacy question gets thorny. Do employees know they're being tracked this way? Lucas: In theory, yes — companies are supposed to disclose it. But the line is blurry. If the platform is analyzing your meeting transcripts to detect that you're good at conflict resolution, is that helpful career guidance or surveillance? A lot of employees feel uneasy about it. Luna: There's also the bias problem. If the AI is trained on historical promotion data, and that data reflects past bias — say, women or people of color were less likely to be promoted into leadership — the algorithm might learn to steer those groups away from stretch roles. Lucas: That's a real risk. Some platforms try to mitigate it by blind-matching — showing the role without revealing the candidate's name, gender, or tenure until later in the process. But if the underlying skills data is biased, the match will be too. Unilever says they audit their algorithm quarterly for fairness, but not every company is that transparent. Luna: Let's talk about the employee experience. If I get a notification saying 'you're a match for a role,' what happens next? Do I apply through the platform, or does my manager get notified? Lucas: It varies. In some systems, you can apply confidentially — your current manager isn't notified until you're a finalist. In others, the manager gets a heads up early on. That's a huge deal because a lot of people are nervous about signaling that they want to leave their current role. They worry about being seen as disloyal. Luna: And if the match is for a short-term project or a 'gig' inside the company, does your manager have to approve the time away from your regular work? Lucas: Typically, yes. These internal gigs are meant to be a side project — maybe 20 percent of your time. But if your manager is already under pressure to deliver, they might block it. So the culture has to support mobility, not just the software. Luna: That's a key point. You can have the best AI matching system in the world, but if managers hoard talent, it's useless. Unilever apparently tied manager bonuses to internal mobility metrics to align incentives. Lucas: That's smart. They made it part of the performance review for managers — 'how many of your people moved into new roles this year?' That changes the calculus. Suddenly it's not about protecting your team, it's about developing talent for the whole company. Luna: So what's the downside for employees? I can imagine a scenario where the algorithm keeps suggesting roles that are lateral moves or even demotions, based on some narrow reading of your skills. You could feel pigeonholed. Lucas: Absolutely. If the system only sees your past performance, it might not recognize your potential for something completely different. There's also the risk that employees feel pressured to take a suggested role — like, 'the algorithm says I should move, so if I don't, am I not ambitious enough?' Luna: Right. It could create a new kind of career anxiety. And what about the data itself? If I'm matched for a role I don't take, does that information stay in my file? Could it be used later to say, 'well, she was offered a promotion and declined, so maybe she's not committed'? Lucas: That's a genuine concern. Most platforms have settings that let you control your visibility, but employees might not know how to use them. And if the data is used in performance reviews or succession planning without your awareness, that's a breach of trust. Luna: So there's a tension between personalization and privacy. The more data the AI has, the better it can match you — but the more exposed you are. Lucas: Exactly. And that's where I think the best implementations give employees control. For example, you can choose to keep your profile 'private' until you're ready to explore, or you can opt out of certain types of analysis. But most people don't read the fine print. Luna: Let's shift to a smaller company. Is this technology only for massive enterprises like Unilever, or can a 500-person firm use it too? Lucas: There are platforms designed for mid-market companies — tools like Gloat, Fuel50, and Eightfold AI offer scaled-down versions. For a 500-person company, you might not have 50,000 internal moves, but you can still surface cross-departmental projects and mentorship opportunities. The ROI is harder to measure, but the engagement benefit is real. Luna: And I imagine the AI gets better as more people use it. If nobody opts in, the matching is weaker. So there's a network effect. Lucas: Absolutely. The system learns from every accepted match, every declined suggestion, every skill endorsement. Over time, it can predict not just who is qualified for a role, but who is likely to thrive in it based on patterns from similar moves. Luna: That's powerful. But it also means early adopters get the worst matches, which could turn people off. How do companies overcome that cold start problem? Lucas: Some use incentives — like a small bonus for completing your profile, or a lottery for employees who participate in the first pilot. Others start with a specific use case, like matching for short-term projects rather than permanent roles, so the stakes are lower. Luna: Speaking of stakes — if today's tech conversation gave you something usable, something that made you think about your own career or your company's policies, that's the whole point of this show. And we deliberately don't run ads on it. If you want to support that choice, the link is buy me a coffee dot com slash fexingo. No pressure, just a way to keep this independent. Lucas: Yeah, that really does help. And it means we can keep digging into topics like this without worrying about sponsors. So — back to the future of internal mobility. Where do you see this going in the next three to five years? Luna: I think the annual review and the traditional promotion ladder are going to become less dominant. Instead of waiting for a once-a-year conversation, employees will get continuous nudges about opportunities. Your career path becomes more fluid — you might take on a six-month project in a different department, then come back to your old role with new skills. Lucas: And that's good for companies too. They become more agile — they can redeploy talent quickly when priorities shift. But it requires a level of transparency and trust that most organizations still struggle with. Luna: Right. The technology is ready. The culture isn't always. So the real work isn't just implementing software — it's changing how managers think about talent. Lucas: That might be the hardest part. But if Unilever's example is any guide, the companies that figure it out will have a serious edge in holding onto their best people.