Can Linkedin survive AI?
AI and the modern economy are already challenging how we perceive the labour market. Can Linkedin survive it, asks Paul Armstrong
AI is busy dismantling the professional network model and forcing companies to rethink how they measure capability, find talent and more.
Linkedin still behaves as if work is stable and identity can be summarised in a neat feed of credentials. Microsoft’s Q4 FY25 results show Linkedin revenue up roughly nine per cent year on year, reaching more than 1.2bn members worldwide. Scale masks stagnation. The network still sells the illusion of permanence in a labour market now moving at the speed of automation.
Artificial intelligence isn’t only changing who gets hired; it’s redefining what hiring means. Machines can already filter, assess and perform faster than any recruiter. The social graph of professionals is becoming irrelevant in an economy where performance is verified continuously rather than endorsed socially.
Linkedin has not died but it feels fossilised. The platform’s feed has become a slopification tsunami: endless synthetic authenticity and algorithmic filler. Every post looks the same because the system rewards engagement over insight. Professional virtue signalling has, or perhaps more importantly feels like it has, replaced professional substance in order to train LLMs. Attention has become the currency because ability can’t be measured. Real work now happens elsewhere in environments that record results, not sentiment. Coders prove worth on Github, analysts demonstrate skill through dashboards, designers show competence in prototypes and builders display progress in public logs. The professional graph is morphing into live proof rather than static profiles.
Linkedin in the AI world
Linkedin was built for linear careers, durable credentials and the company as the primary container of labour. None of those assumptions remotely survive any amount of contact with the modern economy even before you add AI into the mix. Skills depreciate quickly, roles fragment and organisations leak capability into networks of freelancers, tools and AI systems. Competence moves between projects and ecosystems faster than corporate bureaucracy can follow. Professional identity has become dynamic and contextual, yet Linkedin still treats it as fixed with ever dwindling organic reach.
Artificial intelligence dismantles this structure further by turning work into data. Skills should be verified continuously as systems analyse quality, speed and adaptability. Right now I could say I was a top brain surgeon at St Bart’s and Linkedin wouldn’t bat an eyelid. Every deliverable becomes a datapoint in an expanding graph of competence for AI agents to feed on while acting as intermediaries that source work, negotiate contracts and maintain reputational portfolios on behalf of individuals.
The professional network no longer connects people to each other. Capability now connects to demand in real time.
What comes next?
The next Linkedin won’t resemble a social platform. Instead, an invisible infrastructure will sit beneath the surface of work, verifying skill and matching it to need instantly. The unit of value will no longer be the job title but the measurable contribution. HR departments will need a completely new skillset, and set of tools.
Competitors have already sensed the opportunity. Naukri dominates India’s job-tech market, Xing remains strong in Germany, Jack and Jill searches 10,000 new jobs every hour for you, and Welcome to the Jungle makes you swipe for your next role. Then there’s Toptal, Upwork, and AngelList cater to project-based and start-up talent. Each is experimenting with AI for matching, scoring, and verification. Regional clones will multiply as economies seek control over local workforce data. The next evolution pairs platforms with personal AI agents that handle introductions, contracts and reputation autonomously. Once agents communicate directly, platform loyalty disappears. Reputation, skill data, and behavioural history become portable. The question shifts from where a profile lives to which agent represents capability.
Businesses can’t afford to wait for the new architecture to stabilise. Visibility into workforce capabilities has to become a strategic asset rather than an HR exercise. Job descriptions and annual reviews now describe the past, not the present. Data already exists inside project tools, communication channels and workflow software. The challenge is to interpret it as organisational intelligence rather than administrative record (or reasons to replace and automate).
Work needs to be designed with data in mind. Every project should reveal which skills are being used, how effectively they perform, and where automation might extend capacity. Firms that understand their skill graph as clearly as their financial one will adapt faster than rivals. Insight into how talent flows through projects could soon matter more than insight into how capital flows through markets.
Identity will have to become portable. Employees who can prove competence through verified credentials will command opportunities wherever they go, and firms supporting that transparency will attract better talent. The instinct to hoard skills behind corporate walls will become an expensive form of blindness. Credibility and mobility become what’s most valuable.
A larger question overshadows all of this: how can companies treat work as data without dehumanising labour? Best not to look to Amazon as a north star. When algorithms measure output and price contribution, creativity and intent are often found on the cutting room floor in favour of metrics. Leaders have to design systems that measure precisely yet preserve humanity or even they’ll be replaced eventually. Machines calculate but they do not care, and as Target in the US now knows, real smiles matter.
Linkedin will continue to generate revenue and traffic long after its purpose expires because institutions degrade slower than technology. The future of work will unfold elsewhere; inside systems that prove competence, broker intelligence and price contribution at machine speed. Firms capable of reading organisational capability as fluently as financial statements will dominate. Winners won’t be the most connected but the most comprehensible to machines. Those building new data infrastructures for work will control the next labour market. Work is fast becoming data, and companies that fail to interpret it correctly will soon have little left to trade but people.
Paul Armstrong is founder of TBD Group and author of Disruptive Technologies