How Mastercard Put an AI Career Coach in Every Employee's Pocket — And Made External Hiring Optional
By Tim Kreling, Co-Founder, OVI
Every CHRO faces the same expensive math: an open role appears, recruiters scramble to source externally, and weeks later the company pays market-rate premiums for a hire who still needs months to ramp up. Meanwhile, employees with adjacent skills sit two floors away, invisible to the system that is supposed to develop them.
Mastercard decided to break that cycle — not with another training catalog, but with a two-layer AI talent ecosystem that treats internal mobility as the default and external hiring as the exception.
The Architecture: Unlocked + Cai
The first layer is Unlocked, an AI-powered internal talent marketplace built on Gloat's platform and launched globally in 2022. Unlocked matches employees to projects, mentorships, gigs, and full-time internal roles based on their skills, aspirations, and career trajectory — not just their current job title (Charter Works / TIME, 2025).
The second layer is Cai, an AI leadership coach introduced in 2025. Cai uses role-play scenarios to help managers rehearse difficult conversations — performance feedback, conflict resolution, career development dialogues. Employees who have used it report forgetting they were talking to an AI, a signal that the interaction design has crossed a meaningful threshold (HR Leaders Podcast — Lucrecia Borgonovo).
The two layers are distinct but integrated. Unlocked surfaces talent and opportunity; Cai develops the leadership capability to manage that talent well. Together, they create a closed loop: employees find growth paths, and managers are coached to support those transitions.
The Numbers That Matter
The adoption data is striking. Approximately 90% of Mastercard's global workforce — roughly 30,000 employees — have registered on Unlocked, with a 40% monthly active engagement rate. Employees have collectively spent more than one million hours using the platform (Charter Works / TIME, 2025; myHRfuture Digital HR Leaders Podcast).
Those aren't vanity metrics. One-third of active users made an internal career move within a year of joining the platform, and 50% of those moves crossed job families entirely — not lateral shifts, but genuine career pivots enabled by AI-matched skill adjacency (Columbia Business School case study).
In 2024 alone, Unlocked filled 3,000 project roles and facilitated more than 1,300 mentoring relationships. Over 50% of projects and mentoring engagements are cross-functional, breaking down the silos that typically trap talent in narrow career ladders (Charter Works / TIME, 2025).
The financial impact: Mastercard has documented $21 million in cost savings attributed to the platform. That figure comes from a Gloat case study — vendor-sourced data — but the Columbia Business School independently examined the Unlocked deployment and validated the marketplace's structural impact on internal mobility and project staffing (Columbia Business School case study).
Career development survey scores improved by 9 points, a leading indicator that employees perceive real career opportunity inside the company — the exact sentiment that retention strategies are designed to build (myHRfuture Digital HR Leaders Podcast).
When Speed Matters: The Redeployment Use Case
Perhaps the most compelling proof point is what happens under pressure. When Mastercard's fraud detection unit urgently needed AI talent, the traditional approach — post externally, screen, interview, negotiate, onboard — would have taken months. Instead, Unlocked surfaced employees with adjacent data skills across the organization. As Mastercard's leadership described it: "Unlocked helped us redeploy employees with adjacent data skills in record time" (HR Leaders Podcast — Lucrecia Borgonovo).
This is the use case that separates a talent marketplace from a glorified job board. The system didn't just list open roles — it identified transferable capabilities and made rapid redeployment operationally possible.
A New Measurement Framework
Under the leadership of Chief People Officer Lucrecia Borgonovo, Mastercard has deliberately shifted how it measures talent development success. The old model tracked completion rates — courses finished, certifications earned, hours logged. The new framework measures what actually matters: skill growth, internal mobility rates, and business impact (myHRfuture Digital HR Leaders Podcast).
The difference is not semantic. Completion rates reward seat time. Skill growth and mobility rates reward outcomes — faster project delivery, innovation velocity, and reduced dependency on external talent markets. Mastercard's AI-powered platform generates the data exhaust that makes outcome-based measurement possible at scale (Mastercard Global Blog, 2025).
The CHRO Blueprint: Three Actions From Mastercard's Playbook
Other enterprise HR leaders can extract practical lessons from Mastercard's approach:
1. Build a talent marketplace before you need it. Mastercard launched Unlocked in 2022 — before AI talent shortages became acute. When the fraud detection team needed data scientists urgently, the marketplace was already populated with skills data on 90% of the workforce. The lesson: internal mobility infrastructure pays dividends when disruption arrives, but only if it is already in place (AIHR Institute — The Internal Talent Marketplace).
2. Layer AI coaching on top of AI matching. Matching employees to opportunities solves the discovery problem, but not the management problem. Cai addresses the other side: ensuring managers can coach employees through transitions, give effective feedback, and support cross-functional moves. The two layers compound each other (HR Leaders Podcast — Lucrecia Borgonovo).
3. Measure mobility, not completion. If your talent development dashboard still centers on course completion rates, you are measuring activity — not impact. Mastercard's shift to tracking internal career moves, cross-functional engagement, and project delivery speed provides a model for outcome-based HR analytics (myHRfuture Digital HR Leaders Podcast).