The Hiring Intelligence Race

Google / The selection process was rejecting people it should have hired

Google once placed great weight on academic pedigree, difficult questions and long interview sequences. Former people-operations chief Laszlo Bock described how analysis changed that view. The company reduced its interviewing burden and deliberately hired some applicants its conventional screening process had rejected but whom a separate set of questions identified as promising. Bock reported that these hires performed as well as—or better than—people selected through the traditional route.37

The discovery was not simply that better assessment could improve matching. Google's familiar process had created a false negative: people who could succeed were screened out. The business consequence was a needlessly restricted talent pool and costly recruiting effort. Bock's account is first-hand evidence of management changing its hiring logic, not a published controlled experiment establishing an exact effect size.

Google / The working environment changed the performance equation

Google's Project Oxygen began in a culture skeptical that managers contributed much to technical work. Using employee surveys, performance information and qualitative research, the company found that manager behavior was associated with important team outcomes. Google identified effective management practices, trained managers and tracked subsequent improvements. Harvard Business Review documented higher manager-feedback scores and changes in how Google developed its leaders.39

Hiring capable employees had not removed the importance of the conditions in which they worked. That does not prove that every disappointing hire is a management failure. It does show why employee outcomes alone cannot tell the organization whether to revise selection standards or change the environment that determines performance.

Harvard Business School / The talent shortage partly created by hiring rules

Harvard Business School and Accenture's Hidden Workers research documented employers filtering out potentially productive candidates through rigid job descriptions, employment-gap rules and automated exclusions. Their survey found that employers using overlooked talent pools were less likely to report talent shortages; they also reported favorable performance and retention outcomes for those employees. These are associations and employer reports, not proof that changing a specific filter caused the reported gains.38

Still, the underlying contradiction matters: organizations described difficulty finding qualified candidates while acknowledging that their own recruiting machinery could eliminate some of them before serious consideration.

Across employers / Management misunderstood why people were leaving

McKinsey's 2021 surveys found a meaningful disconnect between employers' explanations of voluntary departures and employees' stated reasons. Employers emphasized transactional factors such as pay; employees placed greater weight on feeling valued by managers and organizations and on belonging. The surveys illuminate the risk of treating an employee's exit as a failed recruitment outcome without investigating the working experience. They are not a controlled study of individual hiring decisions.40

Micron / Hiring judged by the business it produces, not the requisition it closes

Jackson Teng, Micron’s Senior Director of Asia Talent Acquisition, Workforce Development & Pathways, describes learning from the people who stayed, thrived, progressed and left. Micron evaluates contribution through performance, time to productivity, retention and future advancement. When retention falls, Teng asks whether the failure reflects a hiring decision, a misleading job preview or onboarding—and argues that the resulting learning must change the next requisition and interview design.43

The consequential move is not another hiring metric. It is giving talent acquisition and operating leaders joint responsibility for revising the assumptions that govern the next decision. This account is an executive interview published by a workforce-intelligence vendor, not independently measured proof of an attributable financial gain.

A hiring-intelligence layer without a mandatory rip-and-replace

In September 2026, Brilliant AI launched from Crosschq, ApplicantX and Traitify, combining reference intelligence, assessments, candidate verification and quality-of-hire analytics. It says it works with existing recruiting systems and will link its products more tightly over time. Its reported scale includes more than 500 organizations, over 40 integrations and more than 200 million “hiring outcomes.”16

The proposition is unusually close to the emerging G2G question: use information from before and after the hire to improve future decisions. This is more than a feature inventory. It is an attempt to name and lead a category.

The test: What counts as one of those outcomes? Are they verified post-hire performance, assessments, interactions, retention, or a mixture? How much is unique and reusable across customers with appropriate permissions? And does the evidence show a lift in prediction or business results versus the next-best approach? The public scale claim is not independent proof of a compounding advantage.

Diagnose what changes—and activate the organization

Visier was linking recruiting to employee performance well before the current AI cycle. In October 2026, it sharpened the case: the Workforce Context Engine unifies governed people data; Vee Diagnostics, launched in early access, promises stronger explanation of workforce changes; and newly announced Talent Agents aim to guide managers and connect program activity to outcomes.22192021

The company now describes an understand → plan → activate operating model spanning talent management. That is more than a passive analytics dashboard. It could help distinguish a selection problem from one involving onboarding, managerial action or organizational conditions. Its strength is breadth of workforce context rather than ownership of the ATS. Its hiring-specific solution already describes following people from initial contact into eventual performance.1832

The test: Can Visier make this specific hiring decision cycle more effective and demonstrate that its diagnostic claims are valid? Its published proposition is compelling but broader than self-improving hiring. Leadership in workforce intelligence would not automatically establish leadership in hiring decisions.

Make the economics of each hiring cohort better

Journeyfront builds role- and business-line-specific selection models and feedback loops around high-volume recruiting. Its Everise case reports 22,186 healthcare-agent hires, a reduction in time to hire from 11.5 to six days, and a fall in 90-day turnover from 59% to 29% for US healthcare-agent cohorts (Q3 2024 baseline versus Q3–Q4 2025). The reported 22,186 hires cover a broader population, including 5,514 US licensed agents; the time-to-hire average runs through February 2026. The case describes changes to assessment length and thresholds as outcome data accumulated. It also says Journeyfront worked alongside the existing Workday system rather than requiring its replacement.23

Those are vendor-published, observational outcomes. Differences in hiring mix, operations and management could also contribute; they are not independent proof of a causal product effect. Nevertheless, this is tangible evidence of an economically relevant learning process, not merely a conceptual platform announcement.

The test: Does success persist across roles, industries and economic cycles with different success measures? Does Journeyfront control something competitors would struggle to reproduce, or does a well-run deployment still require extensive customer-specific work?

What if the talent shortage is partly a specification failure?

Inherited belief. Employers define what a role requires. The recruiting system makes access, assessment and matching more precise against that specification.

Emerging business reality. Google's former people-operations chief Laszlo Bock described hiring applicants its traditional process would have rejected and finding that they performed as well as, or better than, conventionally selected hires. Harvard Business School and Accenture documented employers using recycled requirements and automated filters that made potentially productive people invisible. The HBS findings are survey-based, and Bock's observations are a first-hand account rather than a controlled trial.3738

Candidate reframe. Don't start with the ideal job description. Start by testing what success in the actual work requires—and continuously revise the criteria. This is more specific than generic skills-based hiring: assess demonstrated requirements using job-relevant evidence from people doing the work, while checking for misleading historical patterns.

Economic job and buyer. The target is not only lower recruiting cost; it is fewer qualified candidates unnecessarily excluded, less vacancy time, stronger contribution and fewer costly false hires. The line executive sharing these costs could become a buyer alongside talent acquisition.

Who could credibly pursue it. Cangrade, Journeyfront and established assessment providers already connect job-relevant evidence to performance. Workday itself discusses employee-performance-based benchmarks. A new story about success profiles without a demonstrably better model or application would not constitute an opening.2346

Evidence that would establish an opening. Locate employers with demonstrable false-negative screening rules. Measure how revised, validated criteria affect access, job performance, ramp time and retention using comparable hiring cohorts. Compare the result and total cost with what capable assessment specialists already deliver; audit job relevance and unfair exclusion.

What if the wrong hiring decision is actually the wrong diagnosis?

Inherited belief. Improving candidate selection improves employee performance and retention; a disappointing hire is evidence that the hiring process should change.

Emerging business reality. Micron executive Jackson Teng explicitly distinguishes a poor job preview, unsuccessful selection and deficient onboarding when examining retention, while tracking productivity and advancement as well as whether the hire remained employed. Sony Sports Technology's talent-acquisition leader described using early quality-of-hire signals to identify a window for supporting struggling recruits, not simply judging them. Across employers, McKinsey found that management often misunderstood why employees left. Micron and Sony accounts are vendor-hosted interviews; none proves the prevalence of an attribution error across the entire category.431240

Candidate reframe. Don't just measure whether you hired the right person. Learn which decisions and working conditions cause people to succeed—and change those. The distinction is causal and organizational: a selection score, manager rating and 90-day retention result can all be accurate while prescribing the wrong remedy.

Economic job and buyer. The value may lie in faster time to productivity, lower avoidable early turnover, improved manager capacity and fewer repeat hiring cycles. Talent acquisition alone cannot implement every remedy. The business-unit leader, CHRO and operations leader may have to share responsibility.

Who could credibly pursue it. Visier already describes this broader diagnosis through its Micron discussion and connected workforce context. Workday, One Model and workforce-performance specialists hold related information; Journeyfront addresses parts of the loop in defined roles. We must distinguish a vendor's access to data from a credible mechanism for identifying and changing the responsible decision.184323

Evidence that would establish an opening. Find a business that initially blamed selection, then isolated onboarding, manager behavior, territory, job design or compensation as the actionable driver. Document what changed, the counterfactual or comparison, and financial outcomes. Test whether a specific provider made that diagnosis reliably enough to be chosen for it.

What if connecting the hiring journey still doesn't make the organization learn?

Inherited belief. Integrate Before, During and After, gather more outcome data, run predictive models and better future hiring decisions will follow.

Emerging business reality. SHRM's 2025 recruiting benchmark found only 20% of its respondent organizations used quality-of-hire measures, despite relevant performance information often being recorded elsewhere. Micron's executive describes the core problem as keeping retention information in a postmortem report rather than feeding it back into role requirements and interview design. Journeyfront's Everise case provides a counterexample: it describes changing predictive profiles and thresholds as new outcome data arrived, without replacing Workday. That case reports substantial operational improvement but does not independently establish causality.444323

Candidate reframe. Don't just connect hiring systems. Govern how evidence changes the next hiring decision. An operating discipline would record what criteria were changed, who authorized the change, what business outcome was predicted, what subsequent cohorts actually showed, and whether the policy should remain in force.

Economic job and buyer. The value is repeated, documented improvement in contribution per hiring cohort—not a one-time integration project. An overlay could fit enterprises unwilling to replace ATS, assessment and HCM systems. But integration cost, missing identifiers, low outcome quality, confidentiality and authority across HR and line functions can weaken the business case.

Who could credibly pursue it. Brilliant explicitly claims a continuous hiring-data flywheel across existing systems. Journeyfront documents an implemented customer loop. Visier connects the workforce context; iCIMS and Greenhouse are making different orchestration and intelligence claims. The attractive phrase “continuous learning” is already competitive territory, so the distinguishing proposition must be an executive governance system and better documented results—not the diagram of a loop.16231311

Evidence that would establish an opening. Follow at least three consecutive hiring cohorts from initial hypothesis through changed requirements, screening, selection and employee contribution. Identify the executive who approved the change, compare performance after the intervention, and show that customers pay for this management discipline rather than only software integrations.

Is accumulated hiring history a moat—or an increasingly precise blind spot?

Brilliant points to more than 200 million historical hiring outcomes; iCIMS claims decades of recruiting context. Those numbers are company claims and do not tell us whether the records measure genuine subsequent contribution, whether longitudinal identities are linked, or whether a customer's data can lawfully improve a vendor's models for other employers.1613

There is a deeper evidence constraint: employers generally observe workplace performance for those they selected, not for rejected applicants. An academic study titled Hiring as Exploration investigates that selective-label problem and assesses alternative screening approaches; its observed outcome concerns interview and hiring potential, not proven post-hire productivity. Historical training data can also embed exclusions. Industrial-organizational validation guidance cautions that performance measures must be job-relevant and suitable for the predictive task.4245

Our judgment. This is a competitive stress test for Seams A–C rather than a fourth fully established business reframe. It could become a standalone seam if customers demonstrably prefer a provider that learns from uncertainty and excluded potential, rather than merely generating increasingly confident scores.

The test. Ask for held-out future-cohort performance, change over time compared with a simpler benchmark, controls for job/manager context, lawful cross-client learning rights, and evidence that previously rejected profiles are not automatically treated as failures. Without that, “more data” describes scale, not superior intelligence.

From candidate seam to credible strategic opening

For each path, G2G should reconstruct actual employer decisions: what management expected, the contrary outcome, the decision that was revised, the economic difference and who bore the consequences. Then compare that operating story with at least three vendors, separating capability, distinctive executive POV and validated evidence or asymmetric data advantage.

Do not assume a winner—or force all three into one integrated platform. A specialist might own one economically compelling seam. A workforce-intelligence company might own the diagnostic layer. An overlay may win where replacement costs defeat an all-in-one suite. If the buyers and comparative outcomes do not support the proposed alternative, the seam remains a hypothesis.