Latest / Future of Work Tech with Fexingo: Remote Tools, AI Productivity, and Workplace Software / How the Skills Graph Is Replacing the Job Description
Transcript
- Lucas: If today's tech conversation gave you something usable, I want to start with a number that made me double-take: according to a 2025 Deloitte survey, 67 percent of large enterprises are now piloting what's called a skills taxonomy — essentially a company-wide database of every skill each employee has and doesn't have. Luna: That is a huge jump. I remember when that number was maybe 20 percent a few years ago. Lucas: Right. And the reason it's accelerating is that companies are realizing the traditional job description is a terrible tool for understanding what their workforce can actually do. A job description is static — it lists requirements from maybe two years ago. A skills graph is dynamic. It updates in real time when someone learns Python or gets a project management certification. Luna: So the idea is you don't apply for a job anymore — you match into work based on your skill profile. That's a pretty radical shift. Lucas: Exactly. And companies like LinkedIn and Workday are building these tools right now. LinkedIn already has what they call 'Skills Graph' internally — they map over 50,000 skills across their network. And Workday has their 'Skills Cloud' which does something similar for enterprise HR systems. Luna: So how does this actually work in practice? Give me a concrete example. Lucas: Say you're a marketing manager at a mid-size tech company. You've been there four years, your job description says 'manage paid social campaigns.' But in your spare time you built a dashboard in Tableau, you mentored a junior designer, and you led a cross-functional data project. Luna: None of which is in your official role. Lucas: Right. Under a skills graph system, all of that gets tagged — Tableau, mentorship, data analysis, project leadership. Now when the data team needs someone who can bridge marketing and analytics, your profile lights up. You don't need to apply for a transfer, you don't need to update your resume. The system surfaces you. Luna: That sounds like a dream for internal mobility. But I can also see a darker side — this is essentially a permanent record of everything you've done and haven't done. Lucas: That is exactly the tension. On one hand, it democratizes opportunity — your manager's personal bias matters less than your actual capability profile. On the other hand, it opens the door to constant surveillance. If the system knows you don't have a certain skill, does that hold you back from projects? If you refuse to update your profile, does the company penalize you? Luna: And there's a data privacy question too. Who owns that graph? Is it yours when you leave? Can you take your skill profile to your next employer? Lucas: Some startups are exploring portable skill wallets — basically a blockchain-verified credential you carry with you. But most enterprise implementations today keep the data inside the company. And that creates a power imbalance. Luna: Let me push on the accuracy question. How does the system know I actually have a skill? Do I self-report? Does my manager confirm it? Or is it all based on what I produce in Slack and email? Lucas: All three, actually. Most skills graphs use a combination of self-attestation, manager endorsement, and passive signals. Workday's system, for example, can infer skills from the projects you're assigned to in the system, the training you complete, even the language in your emails. Luna: That last one is creepy if you think about it too long. Lucas: It is. And it's why employee trust is such a barrier. Deloitte's survey also found that only 38 percent of employees trust their company to use skill data ethically. So companies are in a chicken and egg problem: they need the data to make the system useful, but if they collect it too aggressively, people opt out or game it. Luna: Game it how? Lucas: People start adding skills they barely have — 'proficient in Python' after a two-hour tutorial — or they avoid challenging projects because they don't want a failure tagged to their profile. There's early evidence from companies that have deployed these systems that you need strong governance to prevent inflation. Luna: So you need validation layers. Maybe a peer review system, maybe a test, maybe a project-based assessment. Lucas: Exactly. The most sophisticated setups use what's called 'skill endorsements through outcomes.' You don't just claim you can do data analysis — you submit a dashboard you built, and it gets evaluated against a rubric. That's still rare, but it's where the industry is heading. Luna: I want to step back and ask a bigger question: If the skills graph becomes the new job description, what happens to careers? Does it change how people think about their own growth? Lucas: It changes it fundamentally. Right now, most people think about career progression as a ladder — you go from analyst to senior analyst to manager to director. Each rung is a job title. A skills graph replaces that ladder with a lattice. Luna: A lattice meaning you can move sideways, diagonally, not just up. Lucas: Exactly. Your career becomes about accumulating a portfolio of skills rather than climbing a hierarchy. That's liberating for some people, but it's also destabilizing. If there's no clear ladder, how do you know you're 'progressing'? How do you know what to aim for? Luna: And companies need to redefine what 'promotion' means. Is it a salary increase tied to skill acquisition? A new title? More autonomy? Lucas: The companies that are doing this well are decoupling title from compensation. They say: your pay is based on the skills you have and apply, not the box you sit in. That's a huge shift for HR departments that are used to salary bands tied to job grades. Luna: Let's talk about who this helps and who it hurts. I can imagine it benefiting people who are overlooked — introverts, people from non-traditional backgrounds, remote workers who aren't visible to management. Lucas: That's the promise. But the risk is it creates a new kind of bureaucracy. Instead of updating your resume every year, you're updating your skill profile every week. And if the system is gamed, it rewards people who are good at self-promotion, not necessarily good at their jobs. Luna: Then there's the question of bias. Machine learning models trained on historical data can encode existing inequities — if women and minorities have been historically excluded from certain projects, the system learns not to recommend them for those skills. Lucas: That is a real concern. Workday has been criticized for exactly that — their AI recruiting tools showed bias against certain demographic groups. They've since invested heavily in fairness testing, but it's an ongoing challenge. The skills graph is only as fair as the data that feeds it. Luna: So where do you see this going in the next, say, three years? Is it going to become standard for every company over a certain size? Lucas: I think it becomes table stakes for companies that want to compete on talent. But the adoption will be uneven. Big tech companies and professional services firms are already deep in this. Traditional manufacturers and retailers are slower. By 2028, I'd expect maybe half of Fortune 500 companies to have some form of skills graph in production. Luna: That feels aggressive, but maybe not. The technology is mature — the harder part is the cultural and organizational change. Lucas: Exactly. The tech is the easy part. Getting managers to trust a system over their gut, getting employees to buy in, getting HR to redesign their entire career framework — that's the hard work. And it's happening right now, quietly, in HR departments you've never heard of. Luna: Quick honest thing — and I mean this sincerely — a handful of listeners chip in monthly through buy me a coffee dot com slash fexingo, and that's literally what funds making this many episodes. We don't run ads, we don't have a VC backer. It's just listener support. Lucas: And it keeps the show independent. If today's conversation gave you something you can use — maybe that skills graph concept for your own career — that's exactly why we do this. No pressure, just thanks to those who already do. Luna: Alright, back to the lattice. One thing I think is under-discussed is how this affects managers. If your team is assembled dynamically based on skills, your role as a manager changes. Lucas: Dramatically. You're no longer the 'boss' of a fixed group of people. You become a project lead who assembles a team from a pool of skills. You have to be much better at understanding what each person can actually do, and you have less authority over their long-term career. Luna: That sounds like it requires a completely different skill set. Less command and control, more curation and coaching. Lucas: Right. And that's a big reason why some organizations are struggling with the transition. Their middle managers were hired to be supervisors, not talent brokers. You have to retrain an entire layer of management. Luna: So the skills graph isn't just about data — it's about power. Who gets to decide what skills are valuable, how they're assessed, and who gets matched to what work. That's a governance question that most companies haven't answered yet. Lucas: And that's where I think the real innovation needs to happen. Not in the technology, but in the social contract between employer and employee around skill data. If companies can build trust — through transparency, portability, and real employee control — the skills graph could genuinely democratize opportunity. If they don't, it becomes just another surveillance tool. Luna: So the question for every listener is: does your company have a skills graph yet? And if they do, do you trust it? Lucas: That's the one to sit with. Thanks for listening — we'll pick this up again soon.