LinkedIn Steps Up Its AI Play, Turns Its Contacts App Into A New Connected App
With 40% of LinkedIn's 300 million users accessing the social network for professionals via mobile devices, the company has been stepping up a gear on its mobile business, adding new apps around specific experiences like job hunting.
Context & Ripple Effects
LinkedIn's summer of 2014 has been an exercise in taking itself apart and putting the pieces back on your phone. In June it shipped its first standalone product outside the main app, the job search app built with privacy guarantees so employers cannot see who is looking, and days earlier it rebuilt search infrastructure around its homegrown Galene system after ditching a stack of open source extensions. With 40% of its 300 million members arriving via mobile, the company is now converting the Contacts app into Connected — a dedicated app whose pitch is an 'AI play': machine-surfaced insights about the people in your network.
The pickup was broad — The Next Web, VentureBeat, ZDNet, PC World, ReadWrite and Apple-focused outlets all carried the story on the day — reflecting how closely watchers treat every LinkedIn unbundling move as a test of a live argument: analysts writing in early July 2014 made the case against breaking LinkedIn up even while acknowledging it faces ongoing competition, describing it as one of the most durable companies sitting at the intersection of SaaS and social.
First-order effects
- LinkedIn members who used Contacts get a repurposed app whose core loop is algorithmic — Connected prompts relationship maintenance rather than waiting for users to look people up, directly extending the mobile share that already stands at 40% of members.
- The move confirms the portfolio strategy begun with June's job search app: LinkedIn now ships single-purpose mobile products per experience instead of folding each feature into the flagship app.
Second-order effects
- Recruiting-adjacent competitors and CRM-style contact managers face a free rival backed by LinkedIn's exclusive relationship graph — data no standalone contacts app can replicate, and the same graph advantage Galene was built to exploit in search.
- Each unbundled app becomes a new telemetry source feeding the core profile and graph data, tightening the feedback loop between mobile usage and LinkedIn's paid recruiting and subscriptions products.
Third-order effects
- If the pattern holds, professional networking shifts from one destination site toward a constellation of AI-assisted utilities — jobs, contacts, content — each monetized separately but anchored by the same identity graph, which is exactly the structure the mid-2014 anti-unbundling analysts were arguing LinkedIn should resist.
- Relationship-graph data positioned as the raw material for machine-generated suggestions foreshadows the model where the platform's proprietary social data, not its interface, becomes the defensible asset competitors must match.
The trend: Professional networks are unbundling into purpose-built mobile apps where machine intelligence over a proprietary relationship graph, rather than the destination site itself, becomes the product — and LinkedIn is running the earliest large-scale version of that experiment.