The first HiredScore product was not born from a recruiter staring into an inbox. It began with Athena Karp staring at a restroom line. During her first week in investment banking, the queue for the men's room was long and the women's room was not. In finance, this was the unusual version of an ordinary sight. Karp saw a workforce imbalance hiding in plumbing choreography. If a company said it wanted more representative talent but kept producing the same lopsided room, perhaps the hiring machinery needed help.
She founded HiredScore in New York in 2013. The company would spend the next 11 years making an unfashionable bet: the big enterprise systems were not going away, recruiters were not going away, and hiring judgment should not go away. The useful opportunity sat between them. HiredScore could read the company's jobs and talent data, rank the work, revive forgotten candidates, and tell a human what deserved attention next.
That sounds modest beside the usual promise that artificial intelligence will reinvent an industry before lunch. It was also commercially sharp. In March 2024, Workday paid $530 million in cash for all of HiredScore's outstanding stock. The buyer already owned the system where many customers stored jobs, candidates, skills, and employees. HiredScore supplied the layer that could make those records act like a live queue instead of a digital attic.
01 / What it actually doesA score is only useful if it moves somebody
HiredScore calls its category talent orchestration. Strip away the concert-hall language and it is a next-best-action system for recruiting and internal mobility. Spotlight compares applicants with stated job requirements and places them into explainable priority grades. Fetch searches old applicants, prospects, and other talent pools for people who may fit a new opening. Reverse Fetch goes the other direction, matching a person to open jobs. Automation recipes can move a highly rated candidate into a manager-review lane, remind a manager about a stalled task, or invite an old lead to apply.
The internal-mobility product applies the same logic to existing employees. Rather than expecting a worker to browse a neglected careers portal, it can deliver tailored opportunities through email or Microsoft Teams. Managers get coaching and prompts in their normal flow of work. Recruiters see current employees alongside external possibilities. HiredScore's promise is not that every match becomes a hire. It is that qualified people stop disappearing because the relevant human never saw them.
That final box matters. Public product materials describe HiredScore as decision support, with bias auditing, transparency, and human review. The software can decide who rises on a screen or receives an invitation, which is meaningful power, but the hiring call remains with people. In a market crowded by Eightfold AI, Beamery, Phenom, Paradox, SAP, and Oracle, HiredScore's distinction has been its combination of ranking, rediscovery, automation, and deep connection to existing systems rather than a standalone talent marketplace.
“It's almost like a digital exoskeleton that helps recruiters, HR professionals, and the business use the right technology at the right time.”Athena Karp
02 / Who pays and whyThe customer is a company with too many candidates
HiredScore is enterprise software, sold by subscription and priced by quote. Workday documentation treats AI for Recruiting and AI for Talent Mobility as separate licensed SKUs, with implementation and integration work around them. The economic buyer is usually a talent, recruiting, or HR technology leader. Daily users include recruiters, sourcers, hiring managers, employees, and the administrators who worry about access controls, retention rules, audits, and whether the data will arrive on time.
The ideal customer already has an applicant tracking system and a large pool of records. HiredScore publicly lists integrations that reach beyond Workday Recruiting to SAP SuccessFactors, Taleo, SmartRecruiters, and Oracle Cloud. It can also surface leads from recruiting CRMs. That is a useful sales posture: do not ask a global employer to uproot the system of record; make the old investment perform better.
JLL provides the cleanest picture of the pain. The property-services company says about 1.5 million applications arrived each year. Recruiters were struggling to handle ten open roles at a time, and average time to fill had reached 52 days. After adopting HiredScore, JLL reported a 70 percent decrease in screening time, a fivefold increase in recruiter capacity, and quarterly hiring that rose from 3,500 to 5,500. These are customer-reported outcomes, not a universal warranty, but they show what the product is priced against: labor, delay, vacancies, and agency spend.
Vanderlande, the logistics-automation company, reports that screening time for high-volume jobs fell by half. AdventHealth said it hired more than 1,000 clinical candidates from its existing database in the first year and reduced manager decision time by 40 percent. Public materials also name Southwest Airlines, Moody's, Cummins, Charter Communications, Domino's, and JLL. HiredScore's LinkedIn page says the software is live in 150 countries and available in 70 languages.
03 / What failed firstHuman attention, then the promise to call back
The first thing to fail in high-volume recruiting is not judgment. It is attention. A recruiter cannot thoughtfully inspect an unlimited pile of resumes. Managers postpone reviews because hiring is one item among many. Good applicants arrive after the first shortlist. Strong runners-up from last year's search vanish inside the ATS. A company pays again to source candidates it may already know.
HiredScore's wager was that these failures could be represented as queues. Which application has not been viewed? Which requisition is aging? Which past applicant fits a new role? Which manager has candidates waiting? Once the work becomes a queue, software can rank it, explain the ranking, and send a nudge. The recruiter gets leverage without pretending the algorithm understands a career better than a person does.
What changed Karp's mind was not a failed startup pivot that can be packaged into a neat morality tale. Public accounts show a longer conversion: an early observation about gender imbalance stayed with her, then became a belief that data science could expand fair access to work. The product followed the business problem. By 2022, HiredScore said its systems had learned from more than 500 million hiring decisions. Its Blueprint tool used that experience to show how requirements such as location, education, industry, or years of experience could shrink the diversity of a prospect pool.
04 / The catchA faster queue can still point the wrong way
Explainable AI does not make the underlying job definition correct. If a company demands a degree that the work does not require, favors a narrow set of prior employers, or feeds incomplete employee profiles into the system, a consistent ranking can consistently reinforce the mistake. HiredScore's grading can show reasoning and its tools can test for bias, but governance starts before the grade. Somebody must challenge the requirements, monitor outcomes, handle accommodations, and preserve a real appeal path for candidates.
Nor is this product a natural fit for every employer. A small company hiring a dozen people a year has little queue to orchestrate and limited historical data to rediscover. A large company with inconsistent job architecture may need to fix that foundation first. Integration requires subscriptions, security work, role mapping, data policies, and implementation. Automation also disappoints when recruiters ignore the alerts or managers keep the same slow habits. The software works best when volume is high, workflows are measurable, the data is rich, and leadership is willing to change operating routines.
Low hiring volume, vague job requirements, thin historical data, poor manager adoption, or an organization unwilling to audit outcomes. A ranking system can organize a process; it cannot rescue a process nobody trusts.
05 / The founder playbookFour things worth stealing
The most portable HiredScore lessons are architectural and commercial, not algorithmic. Karp's company entered a crowded HR stack by choosing a layer the incumbents had underbuilt. It treated integration as product, not post-sale plumbing. It turned old customer data into new value. And it attached every grand claim about responsible AI to prosaic features: a visible rationale, an audit, a notification, a human review.
Find expensive waiting: unread applications, dormant leads, delayed approvals. Promise measurable movement, not synthetic intelligence.
HiredScore pushed actions into the ATS, email, and Teams. Adoption improves when the user does not need another daily pilgrimage.
Fetch makes a customer's neglected candidate history useful again. Before buying more inputs, look for inventory trapped in an old system.
Explain the recommendation, preserve human review, test outcomes, and document controls. In sensitive software, governance is a feature.
The acquisition makes those choices easier to see. Workday already had Talent Management, Recruiting, Skills Cloud, and a vast installed base. HiredScore brought matching, grading, rediscovery, alerts, and internal-mobility guidance. It was an existing certified partner with a bi-directional integration, so the buyer had evidence that the technology and the teams could work together. The deal closed on March 29, 2024. By August, Workday was selling HiredScore AI for Recruiting and HiredScore AI for Talent Mobility through its own channel.
HiredScore now fits inside a broader Workday recruiting strategy, including conversational scheduling from Paradox and the platform's own job and skills data. The brand may become less distinct as capabilities are described as recruiting and talent-mobility agents. The underlying thesis remains crisp. Enterprise databases remember everything and notice nothing. HiredScore built a business by noticing what the database was trying to say, then handing the decision back to a human before the queue formed again.