Frontline workers are nearly 80% of US employment, more than 100 million people. For years, the conversation about them has focused on how hard its is to recruit and retain frontline workers. Our new research suggests the fix may be easier than many employers think.

The Frontline Workforce Report 2026, sponsored by Workday and Brilliant, looked at what frontline workers actually want and it wasn’t a lot. A response after applying. A schedule they can plan around. Clear information about their pay. A manager who communicates. None of these require a major investment. Most rely on tools employers already have.

Employers have also made real progress. 60% now say talent acquisition or HR owns frontline hiring, up from fewer than 30% of hourly focused organizations in 2020. 84% let candidates apply on a phone. 74% use AI somewhere in frontline hiring. The foundation is in place. The opportunity now is to extend it to the everyday experience of work.

Hiring. Workers want an application they can finish and an answer when they do. 79% of hourly workers have abandoned an application because it took too long. Only 22% of employers respond to 90% or more of frontline applicants. Most organizations already have the communication tools to close that loop. Workers are also open to AI in the process. 78% have used an AI assistant or chatbot when applying, and 67% said it made the experience better.

Engagement. Workers want to hear from their manager and reach basic information from their phone. 83% say communication with their direct manager matters. 52% are satisfied. Their tool requests are practical: 46% want mobile access to pay stubs and tax forms, 35% want to request time off digitally, and 27% want to see their schedule. Meanwhile, 51% of employers say fewer than three quarters of their frontline workers have a company email or login. Meeting workers on the channels they use is a straightforward first step.

Pay. Workers want clarity and access. 84% say clear pay information is important, and 57% are satisfied with what they get. 69% check their pay after every shift. 63% of employers offer earned wage access, yet 55% of workers say they have none. Among workers with six or more years of service, 81% want it and 29% have it. The information workers want already sits in payroll systems. The benefit often already exists. The next step is communication and enrollment.

Scheduling. Workers want notice and a little control. 45% rank their schedule among the top three reasons they took their job, second only to pay. 73% want their schedules in advance, but one in three employers publishes less than a week out. 53% of employers let workers view a schedule, and 12% let them swap a shift. Most scheduling systems can already support swaps and open shift pickup. Turning those features on would also give back time to managers, 55% of whom spend three or more hours a week building schedules.

Mobility and retention. Workers want a path forward, and employers are well positioned to offer one. 56% fill most frontline manager roles from within, and 62% give frontline workers a system to see internal openings. The main barrier is incentives. 48% say managers do not want to lose strong performers, and only 20% tie any part of manager pay to retention. Adding retention and internal promotion to the manager scorecard aligns managers with what the business already wants.

The labor market is near parity, with about one unemployed person for every open job, so employers have room to make changes. At the same time, birth rates are at a record low, more than 4 million Americans turn 65 each year, and immigration has slowed. The pool of future frontline workers is shrinking, and the employers who act now will be better positioned when competition returns.

Frontline workers are asking for control over their time, clarity about their money, and tools as simple as the ones they use everywhere else. These are reasonable requests, and most employers already have what they need to meet them.

Thank you to Workday and Brilliant for partnering on this research.

For years, talent intelligence has been one of the most misunderstood and underutilized areas of talent acquisition. Organizations have invested in data, dashboards, and reporting, but too often, that information has remained disconnected from the decisions that matter most.

That needs to change.

As AI reshapes the workforce, talent intelligence is becoming much more than a way to understand the labor market. It is becoming a critical part of how organizations compete, build capabilities, and execute their business strategies.

In new research we conducted in partnership with Findem, The Future of Talent Intelligence: Insights from AI and Big Tech, we examined talent movement, hiring patterns, skills growth, and organizational structures across five leading AI companies—OpenAI, NVIDIA, Anthropic, Databricks, and Perplexity—and five established technology companies—Meta, Apple, Amazon, Google, and Microsoft.

The findings reveal two distinct approaches to talent. AI companies are prioritizing speed, technical depth, and concentrated capability building, while Big Tech organizations reflect more established operating models built around scale, structure, and internal mobility.

But the implications extend well beyond the technology industry.

The biggest takeaway is that talent strategy can no longer be reactive to business strategy. It needs to help shape it.

Here are five findings that stood out.

1. AI Companies Are Attracting Talent From Big Tech

One of the clearest findings is the direction of talent movement.

AI companies are disproportionately attracting employees from established technology organizations, while hiring within Big Tech remains more circular.

This movement raises important questions about what employees value and how organizations communicate opportunity.

Compensation remains critical, particularly for highly specialized technical talent. But compensation alone does not explain the appeal of AI companies. Opportunities to work on emerging technologies, influence product development, and build relevant skills may also play an important role.

For talent acquisition leaders, this means employer branding needs to go beyond traditional messages about culture, benefits, and stability.

Candidates want to understand where an organization is going and what they can accomplish by joining it.

Recruiting conversations need to evolve from “Why should you work here?” to “What will you be able to build and become here?”

2. Hiring Speed Is Accelerating, but Shorter Tenure Does Not Necessarily Mean Higher Attrition

Our research found that AI companies are hiring 44% faster than Big Tech, with average employee tenure of approximately 3.5 years compared with five years at established technology companies.

At first glance, these numbers might suggest that AI companies are struggling with retention.

But the data tells a different story.

Several AI companies in our analysis had relatively low attrition. Their shorter average tenure largely reflects how quickly they have grown and how recently much of their workforce was hired.

This distinction matters.

Organizations often rely on traditional workforce metrics without considering the business context behind them. Average tenure, for example, can be misleading when comparing a rapidly growing organization with a mature company.

Talent leaders need to look beyond individual metrics and understand what the data actually represents.

The challenge for fast-growing companies is not necessarily retaining employees longer. It is ensuring that hiring velocity does not outpace their ability to onboard, integrate, and develop talent.

3. Skills Are Becoming the Real Competitive Battleground

Perhaps one of the most significant findings involves the pace of skills development.

Across critical AI-related capabilities, including artificial intelligence, machine learning, Python, and large language models, AI companies are growing their skills concentration significantly faster than Big Tech—often at rates two to six times higher.

For example, our analysis found that Python skills grew 33% at AI companies compared with 5% at Big Tech companies during the period examined.

This is not simply a story about hiring more people with technical skills.

It is about how quickly organizations are accumulating the capabilities they need to compete.

And it exposes a fundamental weakness in traditional workforce planning.

Most organizations still approach talent needs through a familiar process: identify a skills gap, create a requisition, and recruit externally.

But when skills are evolving this quickly, that approach may already be too slow.

Organizations need a more dynamic understanding of the skills they have, the skills they need, and how those requirements are changing.

Talent intelligence can help organizations anticipate capability gaps instead of reacting after they emerge.

4. Promotions Are Not Always a Retention Strategy

One of the more surprising findings challenges a longstanding assumption about internal mobility.

Our research identified a positive directional relationship between promotion rates and attrition. Companies with higher promotion rates also tended to experience higher turnover.

This does not mean promotions cause employees to leave.

Instead, it raises an important possibility: promotions may frequently occur in response to employee departures.

When someone leaves an organization, an internal employee may move into that position, creating another vacancy and potentially triggering additional movement.

In other words, organizations may be measuring reactive promotions as evidence of proactive talent development.

This distinction has significant implications for how organizations evaluate internal mobility.

Promotions remain important, but they cannot serve as the only measure of career opportunity or employee growth.

Talent leaders need to understand whether advancement opportunities are being created intentionally or simply emerging because positions have become vacant.

They also need to recognize that career development increasingly includes new skills, expanded responsibilities, and meaningful work—not just changes in title.

5. Geography May Be More Flexible Than Organizations Assume

Despite the growth of remote and distributed work, AI hiring remains geographically concentrated.

Our research found that 48% of AI company hires in the past year were in the San Francisco Bay Area.

It would be easy to interpret this as evidence that AI talent must be concentrated in established technology hubs.

But there is another explanation worth considering.

Geographic concentration may reflect where employers require employees to work rather than where qualified talent is naturally located.

For organizations competing with leading AI companies, this creates an opportunity to reconsider location requirements.

Instead of assuming that certain skills can only be found in a handful of markets, talent leaders should evaluate which roles genuinely require in-person collaboration and which could benefit from broader geographic access.

Talent intelligence can help organizations make these decisions using evidence rather than assumptions.

Talent Intelligence Needs to Move From Reporting to Decision-Making

Taken together, these findings point to a broader transformation.

AI companies are not simply hiring faster or investing in different skills. Their workforce patterns reveal how they are building capabilities, responding to market opportunities, and preparing for their next phase of growth.

We are already seeing evidence of that evolution as AI companies expand hiring beyond technical development into infrastructure, enterprise go-to-market, product leadership, and ecosystem roles.

Talent strategy is evolving alongside business strategy.

For CHROs and talent acquisition leaders, this research raises several important questions:

• Are we using talent intelligence to anticipate future skills needs or simply report on current hiring activity?
• Are our hiring requirements based on actual capabilities or outdated assumptions about experience?
• Are we measuring internal mobility in ways that distinguish proactive development from reactive backfilling?
• Are geographic requirements unnecessarily limiting access to talent?
• Are our workforce investments aligned with where the business is going next?

These are not just recruiting questions. They are business questions.

For too long, talent intelligence has been positioned as a supporting capability within talent acquisition. But its potential is much broader.

The future of talent intelligence is not about having more data. It is about making better decisions with the data we already have.

And as AI continues to change how organizations build and compete for talent, that distinction will only become more important.

Explore the full Aptitude Research and Findem report, The Future of Talent Intelligence: Insights from AI and Big Tech, for all ten findings and their implications for talent leaders.

Most companies are using AI in HR. That debate is over. Our latest research found that 69% of companies are using AI in talent acquisition and 74% are using it in HR. But only 12% are using agentic AI.

Aptitude Research just released The Agentic Moment: HR’s Opportunity to Lead the Next Era of Work, sponsored by Eightfold. We surveyed HR leaders on what they understand about agentic AI, where they want it, what is blocking them, and how ready they actually are. The results are not what the market narrative suggests.

Most HR Leaders Cannot Define Agentic AI

Only 31% of HR leaders say they are very familiar with agentic AI or consider themselves experts. Twenty-six percent told us our survey was the first time they had seen a clear definition.

That matters more than it sounds. Vendors are calling chatbots agentic. Analysts are calling copilots agentic. Some companies are calling any automated follow-up agentic.

Here is the test. Does the system wait to be told what to do at each step, or does it decide that itself? If it waits, it is not agentic. A chatbot answering policy questions responds to inputs. A copilot drafting job descriptions generates content. A dashboard flagging attrition reports. None of them execute.

Organizations that cannot make this distinction cannot evaluate vendors accurately. They will buy the wrong thing and govern it the wrong way.

Nobody Owns AI Strategy

Thirty-three percent of companies have no clear AI vendor strategy at all. Forty-four percent have no clear ownership of AI or are unsure who owns it. Only 13% say the CHRO or CPO owns it directly.

Ownership shapes speed. Organizations where the CHRO or CPO owns AI strategy deploy faster and govern better. Where ownership is undefined or sits outside HR, adoption is slower and more reactive.

The Foundation Is Cracked

The systems HR relies on today were not built for the intelligence era, let alone the agentic one.

Only 28% of companies are satisfied with their ATS decisions. Eighty-two percent say their ATS has significant gaps. Forty-eight percent cite scheduling as the top gap. Thirty-four percent are increasing ATS investment in 2025, and one in four are replacing their system entirely.

Spending more on the same category of tool does not close a capability gap. What changes the equation is the nature of the technology itself.

Where Companies Want Agents

When we asked where companies would most want agentic AI to take autonomous action, the answers were specific and operational.

Screening and interview orchestration came first at 34%. High-volume and frontline recruiting followed at 31%. Internal mobility landed at 21%.

These are not experimental use cases. They are the workflows where volume overwhelms capacity and where inconsistency is most expensive. High-volume hiring is the clearest example. Only 8% of companies are managing it agentically today. Twenty-four percent are still primarily manual.

The Real Barrier Is Not Budget

Only 12% of companies cite unclear ROI as a barrier. Only 6% cite leadership resistance. Only 5% cite vendor capability gaps.

The top barrier is integration. Thirty percent say they cannot integrate across existing HR systems and platforms. Data readiness and internal AI literacy tie at 24%.

This is a plumbing problem and a people problem. Agentic AI requires orchestration. A system that autonomously sources, screens, schedules, and communicates cannot function if its components cannot talk to each other. Companies managing fragmented best-of-breed estates face a compounding readiness challenge that vendor selection alone will not solve.

And only 38% of HR leaders agree their team has the skills to implement and manage AI effectively. Infrastructure can be procured. Workforce readiness has to be built.

One in Two Companies Have No Governance Framework

Half of companies have no AI governance framework in HR at all. Governance scored 3.15 out of 5 on our readiness assessment, one of the lowest dimensions we measured.

You do not need a comprehensive framework before you start. You need a clear policy for the specific use case you are deploying. What the agent can decide. What requires human review. How decisions are documented. Who is accountable.

Build governance use case by use case and let it grow into a framework. Waiting for the perfect policy is how organizations end up with neither policy nor progress.

What To Do Now

Assess honestly. Every readiness dimension we measured scored between 3.15 and 3.34 out of 5. Nobody is far ahead. The lowest score is your starting point.

Sequence deliberately. Start with the workflows where agentic AI is most mature and the ROI case is clearest. Expand into use cases that require richer data. Then orchestrate across the lifecycle.

Buy for connection, not features. The right question to ask a vendor is not what their agent does in isolation. It is how it connects to the rest of the stack.

The agentic moment is not coming. It is here. The question for HR leaders is not whether to act. It is how fast and how well.

I’ve spent the last few months on the road more than I’ve spent at my desk. Conferences, client sessions, vendor briefings, and a lot of hallway conversations that turned out to be more useful than anything on the agenda. When you talk to enough talent leaders in a short window, patterns start to emerge because everyone is wrestling with the same shifts at the same time.

Here are the six themes I kept hearing.

1. AI is moving from efficiency to decision-making

For two years, the AI conversation in talent acquisition has been about efficiency. Faster screening, fewer manual tasks, time back in the day. That made sense, because it is where most of us started. Half of companies still name efficiency as their top reason for adopting AI; only about a quarter point to better decisions.

But the questions are changing. Talent leaders are no longer asking me whether AI can save their recruiters time. They have accepted that it can. They are asking something harder: is it helping us make better decisions, and how would we even know? That is a different phase of maturity, and most organizations have not yet built the muscle to measure it.

2. Candidate fraud is no longer a fringe concern

A year ago, candidate fraud came up occasionally and usually as a story someone heard secondhand. Now it comes up in nearly every conversation, and it comes up early. AI-generated resumes, candidates being fed answers in real time, identity that does not match the person who shows up on day one. What used to be a rare headache has become a real operational risk, and most teams do not yet have a process built for it. Expect this to be one of the defining talent acquisition conversations of the next year.

3. The recruiter experience is still badly misunderstood

This is the one I feel most strongly about. The gap between what recruiters actually do and what leadership thinks they do is not closing — it is widening. When asked, 64% of recruiters say executives do not fully understand their work, up from 58% just a few years ago. We keep asking recruiters to move faster, adopt new tools, and become more strategic, while leaving the underlying workflows, systems, and expectations largely unchanged. You cannot redesign the role by adding pressure to it.

4. The build-versus-buy conversation is back

For a while, build versus buy felt settled — most organizations bought. AI has reopened the question. Teams that never would have considered building are now asking whether they should stand up their own capability rather than wait for a vendor roadmap, and teams that started building are discovering how much it actually takes to maintain. And the pricing in the token economy is much much different. With some vendors, they can do both. There is no universal right answer here, but the fact that the question is live again tells you how much AI has scrambled the old assumptions about what is core and what is bought.

5. Voice is showing up in screening

One of the more interesting things I am seeing is voice moving into screening — conversational, voice-based interactions earlier in the process than we are used to. It is promising, because it can feel more human and accessible than another form to fill out. It also raises real questions about consistency, bias, transparency, and candidate comfort that we have not fully worked through yet. This is early, and it is worth watching closely rather than adopting blindly.

6. Frontline hiring is finally getting the attention it deserves

The frontline now makes up roughly 70% of the workforce, and for years our technology, our processes, and our research treated it as an afterthought. We built for corporate hiring and bolted it onto high-volume roles. That is starting to change. More leaders are asking how to apply AI, candidate experience, and speed to frontline hiring specifically, rather than assuming what works for professional roles will translate. It is overdue. The frontline is where the volume is, where the turnover is, and where a better experience would matter most.

---

Where this leaves us

If there is a thread connecting all five, it is that the easy phase of this work is ending and the harder, more interesting phase is beginning. The tools have changed faster than the operating models around them, and the organizations that pull ahead will be the ones willing to rethink the work itself, not just layer new technology on top of it.

I went deeper on a couple of these — the move from efficiency to decision-making, and the realities of the recruiter experience — in recent research with iCIMS and Greenhouse. If these are the conversations happening inside your own team right now, that is a good place to keep going.

And if you are at one of the next stops on my calendar, come find me. The hallway conversations are still the best part.

I have been covering Eightfold for the better part of a decade, and I attend a lot of conferences. This is not going to be one of my usual recaps. I want to use this year’s Cultivate in Napa to answer the question I get asked more than any other, what does Eightfold do? It is not a question specific to Eightfold. With so many vendors moving quickly, it is hard to keep up when where they started is not always where they are going. For Eightfold, the product has grown into something bigger than the product most of us first met.

Where Eightfold Started

When Eightfold entered the market, it was an AI matching company. The deep-learning models matched people to jobs and jobs to people, surfaced talent that traditional keyword search missed, and made an early, credible case for skills over pedigree. That was differentiated, and at the time that was enough.

Over the years, Eightfold stopped being a feature and became a foundation. The matching engine turned into a talent intelligence platform, and the platform started powering acquisition, mobility, and resource management rather than a single use case.

What I heard at Cultivate this year is the next turn of that evolution. Eightfold is no longer asking customers to buy a product. It is asking them to build on a platform. As the team put it on stage, it is not about what they are building, but how they are building it. That line is the whole story.

The New Model: How the Stack Actually Fits Together

Here is my take on Eightfold’s model. Each layer depends on the one beneath it, and intelligence flows through all of them.

• Talent Intelligence Platform, the foundation. Everything sits on this, and it is built from three kinds of intelligence working together. Foundational Intelligence is the underlying large language models. Enterprise Intelligence is your own data across your HR systems and enterprise applications. Domain Intelligence is Eightfold’s proprietary data, with 1.6 billion career trajectories and 1.6 million skills behind it. The combination is what separates Eightfold from a generic AI layer, because most vendors have one or two of these and Eightfold has all three.
• Talent Tracking, the system of record. Sitting directly on the platform is the layer that tracks talent and movement across the organization. It gives every application above it a consistent source of truth to work from, which is what keeps the rest of the stack from fragmenting.
• Talent Applications, what people actually use. This is where the day-to-day work happens, and it is organized into three families. Talent Acquisition covers CRM, the ATS, candidate experience, and onboarding. Talent Management covers internal mobility, upskilling and development, employee experience, and workforce, talent, and succession planning. Resource Management covers availability, booking, and skill-based utilization for organizations that staff project-based work.
• Talent Agents the part that does the work for you. This is the newest and most important shift. On top of the applications sits an agentic platform and a growing set of agents, including the AI Interviewer Agent and the AI Interview Companion Agent. These do not simply surface insight and hand it back to a person. They execute inside the workflow, running interviews and supporting recruiters as the work happens.
• TalentForge, the build layer that runs alongside all of it. This is the bet that ties the Cultivate announcements together. Rather than buying fixed, one-size-fits-all software, enterprises can build custom HR applications on top of the platform, with talent intelligence available from day one.

The way to read this model is straightforward. The data foundation feeds the applications, the applications feed the agents, and TalentForge lets organizations build their own software on top of every layer. Nothing operates in isolation.

The Announcements, In Context

I am not going to walk through every announcement, but three matter, and they map to the model above.

• TalentForge is the headline. It moves Eightfold from a software company to a platform for building software, and it reflects the company’s view that the majority of enterprise software will eventually be custom-built rather than bought off the shelf.
• 360 Interview for AI Interviewer collapses multiple interview types, including screening, functional, coding, and language, into a single agent-orchestrated conversation. It is rebuilt around how agents work rather than around human scheduling constraints and multiple rounds.
• Workforce Readiness gives CHROs something they have not really had before, which is a board-ready metric for how AI adoption is actually progressing across the organization, instead of relying on lagging surveys and instinct.

Why This Matters

The reason “what does Eightfold do” has been such a hard question is the market has changed, Eightfold has changed, and the answer kept changing as the company grew. After this Cultivate, I think the honest answer is this. Eightfold gives organizations a talent intelligence foundation, a full set of applications across the talent lifecycle, agents that take action inside those applications, and now the ability to build their own software on top of all of it. This is what will power the workforce of the future, what they call the infinite workforce. The architecture to support it is real, and it is shipping. I will be watching how customers actually build on it, because that is where this next chapter gets decided.

For years, the top recruiting challenge that companies faced was reach. The ability to attract more candidates, generate more applicants, and expand the top of the funnel defined success in talent acquisition. Job boards, programmatic advertising, and sourcing tools were all designed around the same objective…get more candidates in the door or make it easier for more candidates to apply.

But, with advancement of AI, shifts in the labor market, and organizational restructuring, reach is no longer the main challenge. It is volume, signal, and decision-making.

Indeed’s recent announcement around Smart Fit Score and Smart Screening is helping companies address these challenges.

What Is This Announcement

Indeed has introduced a new capability through Smart Screening that moves beyond matching and into structured candidate evaluation. At its core, Smart Screening is embedded directly within the application experience through Indeed Apply. Smart Screening is not just another screening tool. It fundamentally changes where and how screening happens in the hiring process.

Instead of relying on recruiters to conduct phone screens days after an application is submitted, Smart Screening moves that step directly into the application itself. The moment a candidate applies, they are engaged through a structured, conversational experience that captures deeper context about their qualifications, experience, and intent. The system can ask follow-up questions, clarify responses, and gather information that would typically require a recruiter conversation. This creates a more complete and structured view of the candidate before any human interaction takes place.

The Smart Fit Score is a key part of this. It introduces a more consistent way to evaluate candidates against a defined hiring bar. Instead of relying on resume keywords or subjective interpretation, it uses candidate-provided data to assess alignment with role-specific criteria. This helps standardize what “qualified” actually means which is something most organizations still struggle with.

There are also two important layers that reflect where the market is heading. First, bot mitigation. As AI-generated applications increase, the ability to verify that candidates are real and that responses are authentic becomes critical. This is less about fraud and more about preserving signal quality in a high-volume environment.

Second, credential verification. Allowing candidates to upload proof of licenses or certifications early in the process adds a level of validation that typically happens much later. This is especially relevant in frontline and regulated roles where compliance matters.

Smart Screening is doing three things:

• Shifting screening upstream into the application
• Structuring candidate data earlier in the process
• Improving signal in a high-volume environment

This is not about making screening faster. It is about making it more consistent, more scalable, and more aligned to how hiring actually happens today.

Why This Is Significant

This announcement signals a shift in how we should think about the role of platforms like Indeed. Indeed has historically been positioned as a platform that connects employers and candidates and drives application volume.

But volume without structure creates new problems. Employers are overwhelmed with applicants. We found that 68% of companies have more applicants this year than last.  And, 58% of companies have less recruiters on their teams this year. Recruiters are spending more time sorting than engaging. And candidates are often indistinguishable from one another, especially in an environment where AI-generated resumes and applications are becoming more common.

By introducing structured screening and fit scoring at the top of the funnel, Indeed is moving into a new role of not just driving applications, but helping employers make smarter decisions around them.

Benefits of Screening at Scale

The benefits of this announcement extend to both employers and candidates, but they are most immediate for employers who are under increasing pressure to deliver better outcomes with fewer resources.

Efficiency: At a foundational level, Smart Screening helps address one of the biggest challenges in talent acquisition today: efficiency at scale. Recruiters are spending a disproportionate amount of time reviewing applications, conducting initial screens, and managing follow-up communication. At the same time, teams are smaller. Aptitude Research found that 58% of companies have reduced their TA teams, even as hiring demands and expectations continue to rise. This creates a clear need for solutions that can expand recruiter capacity without adding headcount.

Quality of Applicant: It also has a direct impact on quality of applicant. One of the persistent challenges in hiring is not just attracting candidates, but identifying which candidates actually meet the hiring bar. Aptitude Research found that quality of hire is the top priority for 70% of companies, yet many still lack a consistent way to define or measure it. Structured screening and fit scoring introduce a more standardized approach to evaluating candidates, helping employers focus on those who are most aligned with role requirements earlier in the process.

Candidate Drop-Off: Candidate drop-off is another critical area of impact. Engagement at the point of application is often the difference between completion and abandonment. We have found that when candidates receive communication within the first 15 minutes, application completion rates can reach 85%, but when there is no engagement, that drops to 35%. By introducing immediate, conversational interaction, Smart Screening helps maintain candidate momentum and reduce drop-off at a stage where most organizations lose talent.

Decision-Making: Finally, this approach improves overall decision-making. Instead of relying solely on resumes or fragmented notes from phone screens, recruiters are provided with structured, consistent inputs that include candidate responses, summaries, and alignment to defined criteria. This creates a more informed starting point for hiring decisions and reduces variability across recruiters and hiring managers.

These benefits reflect a broader shift in talent acquisition. The focus is no longer just on generating applicants, but on improving how those applicants are evaluated, engaged, and converted into hires.

Enabling Hiring at Scale Without Losing Signal

The real value of this approach is not just efficiency. It is signal quality. In high-volume hiring environments, the challenge is not just processing candidates quickly. It is ensuring that the right candidates are not lost in the noise. By introducing structured inputs and aligning them to defined hiring criteria, Smart Screening helps organizations move from keyword-based filtering to criteria-based evaluation

This aligns with a broader shift in the market toward more structured, explainable decision-making. It also supports a more consistent definition of the hiring bar, which remains one of the most persistent challenges in talent acquisition.

A Responsible Approach to AI in Hiring

One of the more important aspects of this announcement is what Indeed is not doing. Smart Fit Score is not designed to:

• compare candidates against one another
• make autonomous hiring decisions
• replace recruiter judgment

Instead, it is built around employer-defined criteria  and candidate-provided information . This reflects a more responsible and controlled approach to AI in hiring. Aptitude Research has found that trust remains one of the biggest barriers to AI adoption in talent acquisition. Solutions that prioritize transparency and human oversight will be better positioned for long-term adoption.

Aptitude Research and Phenom released the State of Hiring Automation 2026 this week. It was a labor of love that we started last year. It was a project that combined an audit of more than 200 enterprise hiring experiences with survey data from over 300 organizations. Together, we mapped where companies are investing in automation, where candidate experiences are breaking down, and what the inline hiring journey actually looks like today.

One of the most interesting findings: organizations are not as far behind as they think and not as far ahead as they hope. Across industries, role types, and maturity levels, the patterns were remarkably consistent. Frontline hiring, knowledge worker hiring, healthcare, retail, manufacturing, hospitality … all showed similar results. Most organizations are in the same place.

That consistency is both reassuring and clarifying. It means companies still have a real window to differentiate without feeling like they’re chasing competitors who are years ahead.

Where investment has paid off

A decade of investment in attraction and candidate engagement has delivered real results. Career sites, job discovery, applications, and communication have all improved meaningfully. Across attraction, engagement, and conversion, organizations are operating at roughly 62% of their potential — a solid foundation.

• 62% — average maturity in attraction, engagement & conversion
• 21% — average hiring automation score post-apply
• 94% — of organizations not scheduling interviews inline
• <1% — with a fully orchestrated inline qualification workflow

Where the gap lives

Once candidates apply, the experience changes dramatically. Hiring automation scores averaged just 21%. The breakdown isn’t in attracting talent. It’s in qualification, assessment, screening, interview scheduling, and decision-making. Too many organizations still push candidates into fragmented post-apply workflows, losing top talent at the moment of peak intent.

Automation inline: the current reality

• 11% use role-specific qualification early in the process
• 11% deploy assessments during the application flow
• 7% automate interview scheduling inline
• 1% are using voice-based screening agents

What comes next

For years, the market focused on attracting more candidates. The next phase is about qualifying the right candidates faster and with greater consistency investing in assessments, interview orchestration, scheduling, screening, and AI that supports decision-making inside the workflow, not just around it.

The organizations that move fastest won’t necessarily be the ones with the biggest applicant pools or the most AI announcements. They’ll be the ones that create a more connected inline experience that compresses the time between apply, qualification, assessment, and interview. That is where hiring automation becomes meaningful.

We are sharing these results May 20 at 11am EST: Mind the Gap: What the 2026 State of Hiring Automation Report Reveals | Phenom + Aptitude Research Webinar

Last week at Visier Outsmart was a different event from most I have attended. Not just the content or the product announcements. But the experience itself. I left feeling like this was a community and I left with a very clear vision of how Visier supports its clients and where it is going in the future. In the messy world of HRTech, this type of clarity is rare.

I had the opportunity to moderate a panel on talent decisions and host a fireside chat with CEO, Ryan Wong, on the big themes of 2026 including SaaS-pocalyse and strategic workforce planning.

Here are my takeaways:

1. The message was consistent: AI is about decisions, not efficiency

For years, the narrative has been: AI will make HR more efficient and save time. That’s not the conversation anymore. The shift is toward using AI to make better decisions—about hiring, workforce planning, retention, and organizational design. Not faster reporting. Better judgment.

2. We are moving from insights to outcomes

One of the clearest themes across sessions: Insights alone are not enough. Companies have spent years building dashboards, reports, and analytics capabilities. But as was said during the event, if no one acts on an insight, it doesn’t matter.

The focus now is on:

• What should I do?
• What action should be taken?
• What outcome will this drive?

This is where workforce AI is heading—closing the gap between data and action.

3. Recommendations are the new frontier

There’s a growing expectation that systems don’t just show data. They guide action.

That means:

• Recommending workforce decisions
• Aligning those recommendations to company policy and context
• Embedding them into workflows where decisions actually happen

This is a big shift. It takes the fear out of AI-enabled recommendations and puts power behind it. It also raises new questions around trust, governance, and accountability.

4. The future is a continuous decision loop

The most important concept reinforced throughout the event is that workforce intelligence is no longer linear.

It’s a loop:

• Understand the workforce
• Plan based on that understanding
• Get recommendations
• Take action
• Improve continuously

This “decision loop” is what turns AI into something operational and not theoretical.

5. This is the SaaS-pocalypse moment—and vendors matter more than ever

For the past decade, the model was simple: Buy more technology. Add more tools. Solve for features. That model is breaking. Companies are overwhelmed with systems, unclear on value, and facing pricing models for AI that are still largely undefined. Consumption-based pricing, token-based models, bundled AI—it’s all still evolving.

This is the SaaS-pocalypse.

But organizations can’t navigate it alone.They need vendors that don’t just provide technology—but provide context, guidance, and operational support. The burden is shifting from “buying software” to depending on partners to make sense of it.

What made this event stand out wasn’t just the technology. It was the clarity of direction. The market is moving beyond experimentation toward something much bigger:
Using AI to drive better workforce decisions—and ultimately, better outcomes. And what stood out most is that this community isn’t waiting to figure it out. They’re already doing it.