Every HR leader is being asked the same question: what is AI actually doing to our people, our roles, and the way we work? At the State of HR Summit 2026, more than 400 of the country's HR leaders answered it in real time.
We ran a live poll inside each of the four panels, so the experts on stage had to respond to the room as the numbers came in. What follows is the round-up: the sharpest findings from the day, and the moves worth making because of them.
The conversation has moved past whether to adopt AI
The State of HR Summit 2026 ran under the theme Leading a People-first AI Transformation. Nine minutes into his opening keynote, Cliff Eala, CEO of BS Works, challenged the premise itself, asking whether the goal should be people-first at all. He had watched companies protect their people so carefully that no transformation happened, then lose ground in the market regardless. It was a fitting start to a day that treated AI adoption as settled and turned instead to the harder question of what comes after it.
Across four panels, four moderators and a dozen panelists worked through what a people-first transformation actually demands once the tools are already in the building. A live poll ran inside each session, giving us an unusually candid read on where HR leaders stand today.
The pattern was consistent. Just 12% of attendees named acquiring AI skills as their main struggle. 42% pointed instead to redesigning roles fast enough, and 41% said their adoption is stalled because HR, Operations and Compliance are not aligned. The barrier Philippine HR reports is structural, not technological.
That finding echoes the national picture. Among the State of HR Report's 3,500+ respondents, 78% already use AI in their daily work, yet only 35% receive training for their specific role. The tools have arrived. The structure meant to support them has not caught up.
Top shift reportedRoles changing faster than we can update them 28%
Top adoption blockerHR, Operations & Compliance out of sync 41%
Hardest leadership habitSetting performance standards that keep shifting 35%
Biggest struggleRedesigning roles fast enough 42%
First priorityOperationalizing AI in HR 103 of 244
Opening the State of HR Summit 2026 at Dusit Thani Manila.
Before you read on
Where are you stuck?
Each poll below was answered live and mid-panel, with the experts on stage responding to the results in real time. Select the challenge that sounds most like yours, and see how the room compared, then jump straight to the finding that addresses it.
Why Are Roles Changing Faster Than Teams Can Keep Up?
During the summit's opening panel, two speakers framed the shift from opposite ends of the market. Dannah Majarocon, Managing Director of Jobstreet by SEEK, offered the market-wide view, drawn from Jobstreet's vantage point on both sides of every Philippine hiring transaction. Marvin Zoilo, Group HR Director at Tech Mahindra, supplied the enterprise end, where the company now runs roughly 20,000 AI agents on routine administrative tasks, freeing its human workforce to move into higher-value, insight-led work.
Which shift is showing up most in your company?
n = 297 · final in-session tally
Hover any bar for detail
What this tells us
The most telling result is that no single shift dominates. Attendees are contending with all four at once, and the top two describe the same pressure from different angles: when the work changes but the role description does not, the gap shows up as a widening skills gap from the inside. The implication is clear. Role definitions need a far shorter refresh cycle than the annual review most organizations still rely on.
No option won decisively. Just seven points separate first from last, a spread that suggests attendees are living all four shifts simultaneously rather than choosing between them.
The top two answers likely describe a single underlying condition. A widening skills gap is often simply what a changed role feels like from the inside, once the work has moved and the job description has not caught up.
Cliff Eala had named the mechanism behind this an hour earlier. Within any job, he argued, there is a boundary between the tasks that can be automated and the tasks where a human, assisted by AI, performs better. That boundary is real, and it refuses to hold still.
"Because AI is moving so fast, that boundary is continuously shifting. Your job, or our job as people managers, is to make sure the jobs are redesigned around that boundary. And it's not a one-time thing."
Cliff Eala
CEO, BS Works · Keynote
The gap worth acting on
72% of Philippine businesses now factor AI knowledge into hiring decisions.1 Majarocon then asked for a show of hands: who already has AI embedded in recruitment?
72%
of Philippine employers screen candidates for AI knowledge
32%
have adopted AI within recruitment itself
Ahead of the 29% Asian average, yet well behind what employers now ask of their candidates. Employers screen for a capability the screening function has largely not adopted.
Two vantage points on the same change
28% of attendees report roles changing faster than they can update them, and 26% report a widening skills gap, the same pressure seen from two sides. As one panelist put it, describing how a single role now absorbs work that used to be split across several: one person may be the strategist, the creator, and the reviewer, all of it made possible by AI. The standard of a role is moving faster than the description written for it.
Job security perceptionsM = 3.00 · High
1.00 Very Low2.00 Low3.00 High4.00 Very High
State of HR Report respondents say job security perceptions remain stable, consistent across tenure, generation, rank, work setup, and company size.2 Attendees can see the boundary moving. Respondents feel settled. Both are accurate to their vantage point, and the distance between them is where your communication planning belongs.
What to do next
Name an owner for the boundary. Someone's job is to watch where automation ends and augmentation begins, and to flag when it moves. If it is nobody's job, your job design is happening by default.
Audit at task level, not role level. Pick one function. List what people actually do, not what the JD says. Anything that runs on data, patterns, and rules is a candidate.
Check your own function first. If you screen candidates for AI knowledge and your recruitment process has none, start there.
Decide which cohort you are. AI-native or AI-curious. Then recruit for the one you actually are, not the one you would like to be.
Where Sprout fits
Closing the 72/32 gap starts inside recruitment itself. Teamtailor takes the paper-screening load off your team so human time goes into the conversations that need it, and it feeds straight into Sprout's core HR and payroll platform, so a hire becomes an employee record without anyone rekeying anything.
If the Tech Works, Why Is AI Adoption Still Stuck?
To open the session, moderator Firdausi Ismail II Abbas, VP of People and Culture at Acquire Intelligence, grounded the discussion in three external findings. 63% of employees name skills gaps as the single biggest barrier to workforce transformation through 2030.3 Employees are ready to change how they work, yet the metrics and incentives around them still reward the old way.4 And AI-exposed companies have tripled their productivity since 2022, with the clear winners using AI to amplify their people rather than to cut costs.5
"The real question to this panel isn't whether AI works. It's whether or not we've adopted it successfully."
Firdausi Ismail II Abbas
VP People and Culture, Acquire Intelligence · Moderator
Where is your AI adoption stuck?
n = 224 · final in-session tally
Hover any bar for detail
What this tells us
The obstacle is neither the technology nor the pilot. It is that HR, Operations and Compliance are not pointed in the same direction. Stalled pilots and unprovable impact together account for less than a third of responses. Aligning these functions on a single shared roadmap, before any tool is purchased, removes the most common barrier in the room.
29%
The two answers most associated with AI maturity conversations, stalled pilots and unprovable impact, together account for less than a third of responses. The leading answer is organizational alignment.
Abbas attributed the result to competing priorities and the absence of a shared roadmap. Majarocon had made the same case on the earlier panel, and it reframes the headline number: 41% is less a sign of dysfunction than an honest description of scope. HR is being asked to deliver an outcome whose inputs sit across several functions it does not control.
"HR cannot do this alone and HR cannot be the starting point. This needs to be an org-wide conversation. Your CEO needs to champion this. Your COO needs to drive this."
Dannah Majarocon
Managing Director, Jobstreet by SEEK
Three sequencing points from the panel
Tanya Barredo · Booth
Baseline first
Establish an operational or data governance policy before the tools.
"The tendency of a lot of excited companies that want to try AI is that they tend to forget we need to create a baseline."
Booth's first genuine handover was sourcing. Her caveat was honest: the team still has to be equipped to catch the model's errors and hallucinations.
Read more +
Tanya Barredo · Booth
One area at a time
Trying to fix sourcing, CV creation, and interviews at once tends to produce none of them.
"If you try to do all of that, you take too much bite into it and you don't focus on one thing, you will fail in your AI adoption."
Read more +
Annie Caw · CIMB Bank
Not everyone can pilot
In banking, technology is vetted and approved before adoption. Deliberate decisions replace experiments.
"It will always be a provisionary employee subject to performance management, because certain technology will only be as good as your current scenario."
Asked how CIMB balances speed against compliance, her answer was not about AI at all: "The answer is actually not within AI. The answer is with us finding the right people."
Read more +
"It's the everyday grind, the everyday conversations, what you embed in your learning programs, in your performance management system. That is how we bridge the gap."
Annie Caw
Chief Human Resources Officer, CIMB Bank
On the 30% who have not started
This is a meaningful share, and worth reading alongside the State of HR Report, where 78% of respondents say they already use AI daily. Individual adoption is running well ahead of organizational adoption. Marvin Zoilo described the same pattern from Tech Mahindra: high personal use on mobile devices, far lower use inside formal workflows. For any organization in this group, that existing employee familiarity is an asset already sitting on the balance sheet.
What to do next
Write the governance policy before you buy anything. An operational or data governance baseline is the cheapest insurance available and the most commonly skipped step.
Get the roadmap onto one page with three names on it. If HR, Operations and Compliance have different roadmaps, that is the 41%. Fix the roadmap, not the tool.
Pick one area, finish it, then move to the next. Trying to fix sourcing, screening, and interviews all at once is the most common way AI adoption stalls.
If you are regulated, stop apologizing for not piloting. A deliberate, approval-first approach to new technology is a legitimate operating stance, not a sign of falling behind.
Where Sprout fits
Annie Caw's point is that adoption embeds through learning programs and performance management, not launches. Perform Pro is where that lands, carrying the new AI expectations into goals and reviews, and because it sits on Sprout's core HR and payroll platform, the same record follows the employee.
How Do You Measure "Good" When the Work Keeps Changing?
The panel opened on how much has already changed. Airon Caraullo, Head of Systems, Reports, and Process Excellence at KMC, estimates his own team is now 90% AI users, and Julian Mengual, CEO of Medicard Philippines, explained why that shifts the ground under every leader: information used to be a source of power, but AI equalizes access to it, so the real differentiator becomes the quality of the questions you ask.
Which leadership habit is hardest to update for AI?
n = 202 · final in-session tally
Hover any bar for detail
What this tells us
Leaders can define almost everything except what good now looks like. The moment the standard moves from what you deliver to how you deliver it, everything downstream becomes unstable: hiring bars, calibration, promotion, and pay. Translating that new standard into a handful of clearly named behaviors is what gives managers something concrete to coach against.
Setting performance standards leads by a clear margin, and the reason it matters runs deep. Standards underpin hiring bars, calibration, promotion, compensation, and succession. When the definition of good work shifts, every process downstream of it shifts too.
May Manaog, Head of HR Philippines at IQ-EQ, described precisely what has changed. The old cycle was familiar: set goals at the start of the year, track metrics through it, assign a rating at the end.
"It's not the what anymore right now after two years. It's how you deliver it. The adaptability, your ability to change and drive change, the ability to use the AI technology responsibly."
May Manaog
Head of HR Philippines, IQ-EQ
Cliff Eala's keynote had offered a method for making that assessable. Standards expressed as mindsets cannot be coached or calibrated, but standards expressed as five to seven named micro-behaviors can be modeled, codified, and taught.
On the 10%
Giving feedback when AI shaped the output ranked last. Our finding does not necessarily mean this is a solved problem. It may indicate that the issue is difficult to observe rather than easy to manage.
SOHR respondents: "reliance on AI weakens my thinking skills"M = 2.00 · Low
1.00 Very Low2.00 Low3.00 High4.00 Very High
In the State of HR Report, respondents report that reliance on AI is beginning to weaken their own thinking skills.6 Caraullo, whose team is already 90% AI users, has built a coaching method around exactly this, having found the behavior invisible under standard questioning.
"One thing we should avoid asking our team members is did AI do this. Because 90% of the time AI probably did. Instead: walk me through how you make AI do it."
Airon Caraullo
Head of Systems, Reports & Process Excellence, KMC
Strong performers narrate the problem, the prompt and their review of the output. Others hand over the output unreviewed. Caraullo named the prerequisite: a leader who cannot prompt or iterate cannot evaluate the answer, which makes staying current a coaching requirement rather than a personal interest.
"It's not coaching to decide if someone's using AI or not and catching out if they use AI. Who cares? We want to encourage. It's more about, do you effectively use AI and are you using good decision-making qualities?"
The State of HR Report found that respondents report greater benefits from AI when they see it as part of their own success rather than as one more tool imposed on them. Coaching that reads as encouragement supports that framing. Coaching that reads as detection may work against it.
What to do next
Change one question in your next one-on-one. Not "did AI do this." Ask them to walk you through how they made AI do it. You will sort your team in about ninety seconds.
Write down five to seven micro-behaviors for every standard you claim to want. If nobody can name them, you do not have a standard, you have a slide.
Audit your own capability before you coach. If you cannot prompt or iterate, you are not evaluating your team's work. You are accepting it.
Keep it encouraging, not forensic. The moment the question reads as a trap, you lose the only instrument you had.
Where Sprout fits
If the standard has moved from the what to the how, the system has to carry behaviors, not just targets. Perform Pro handles goals, reviews, and OKRs with the nuance a shifting standard needs, and it connects to Sprout's core HR and payroll platform, so what you assess and what you pay stay in the same system.
What Are Teams Really Struggling With? (It Isn't Skills)
Where is your team struggling most?
n = 100 · final in-session tally
Hover any bar for detail
What this tells us
Only 12% named acquiring AI skills as the hard part. The real struggle is structural: how work is organized once those skills arrive. Investing in training alone, without redesigning the roles and the organization around them, solves the smallest of the four problems on this chart.
Redesigning roles leads at 42%, consistent with Finding 01. The more instructive number may be the last one.
12%
identify sourcing AI skills as their main struggle. Capability acquisition is the smallest of the four concerns. The larger need is structural: how work is organized once the skills are present.
During the closing panel, Fleire Castro, founder of DashoContent, offered the clearest picture of role redesign in practice. She runs a deliberately lean team: four full-timers plus a roster of creators.
"Gone are the days when I can just say okay, your only role is to QA. Doesn't happen anymore. So one person may be the strategist, the creator, the QA, and all of this is possible because of AI."
Fleire Castro
Founder, DashoContent
Scale is doing a lot of work in that answer. What is a conversation at four people becomes a process at four hundred, which is much of what the 42% is measuring. Dr. Raymond Sarmiento, Chief Operating Officer of mWell, supplied the enterprise version: forward-deployed engineers, a role that did not exist a year ago, now carry work that previously required a product owner, a designer, a developer, and a QA.
On well-being at 24%
Well-being surfaced here as a genuine struggle, named by roughly a quarter of the room. Yet it is worth noticing early that difficulty and priority are not the same thing: attendees flag well-being as hard without necessarily treating it as the first thing they will act on.
Sarmiento's answer was infrastructure rather than a perk: a protected 3pm movement habit each day, in-house counselors and psychologists, workload and 360-degree surveys, and wearables feeding a department-level dashboard. Paula Ferrer Cheng of MindYou supplied the number that makes the case urgent. More than 70% of the Philippine workforce is already operating at a psychological deficit before any new tool is introduced.
"The historic view would be, well, you got 20% more time. Brilliant. It's not how it works. Because the energy I've used in the 60% of my time, you expect me to do another 40%?"
Julian Mengual
CEO, Medicard Philippines
An open question worth weighing
Asked whether the 40-hour week is now obsolete, Cliff Eala argued the opposite. All else being equal, he contended, whoever works less will lose to the competitor who works more. Both positions have merit. The trade-off is a real one for HR leaders to weigh rather than a question with a clean answer.
What to do next
Stop treating role redesign as an annual project. At scale it needs an owner and a cadence, because it will not survive as a conversation.
Do not spend the 20% before you have it. The routine work you are removing was partly recovery. Plan for what replaces it.
Measure well-being at department level. Aggregate dashboards give you the signal without the surveillance problem.
Put the trade-off on the table honestly. Burnout is a real cost. So is losing to a competitor who did not blink. Decide with your exec team, not in an HR silo.
Where Sprout fits
Redesigning roles at scale needs somewhere for the org and job structure to actually live. Sprout HR holds the employee database, org design, and job structure alongside payroll, so a redesign becomes a change to a system rather than a memo, keeping the new structure connected to payroll, performance, and everything else downstream of it.
Where Will HR Actually Start? The First Move That Matters
Beyond the four in-session polls, attendees were also surveyed on a single forward-looking question: which priority they would act on first, selecting up to two. Just over half of checked-in attendees responded, making this the one dataset that reflects a full day of discussion rather than a single session.
Which priority will you act on first?
n = 244 · respondents selected up to two, so selections exceed respondents
Hover any bar for detail
What this tells us
Asked what they would tackle first, HR leaders pointed decisively at operationalizing AI and redesigning roles, together drawing nearly double the selections of everything else combined. They are two views of one job: reshaping work around AI. Evolving leadership placing third, in a room that named leadership its hardest problem, is worth a second look before you finalize your own order.
The top two selections describe a single agenda approached from two directions. Operationalizing AI and redesigning roles are the technology side and the people side of the same work, and together they drew nearly twice the selections of everything below them combined.
Two entries invite reflection
Evolving leadership placed third, at 52 selections, just over half the weight given to role redesign, in a room that had spent the afternoon naming leadership habits as its hardest problem.
Performance management transformation placed fifth, at 44 selections, even though 35% of attendees had named performance standards the single hardest habit to update.
A sequencing consideration
Both patterns suggest a preference for addressing structure first. Given that 41% report their adoption is stalled by cross-functional misalignment, sequencing leadership alignment earlier may be worth considering.
What to do next
Treat the top two as one plan, not two. Operationalizing AI and redesigning roles fail separately and work together.
Sanity-check your own order. If leadership sits third on your list too, ask whether that is a decision or a default.
Start with the finding that stung. Four findings, four sets of moves. Pick the one you recognised, not the one that is easiest.
Where Sprout fits
Operationalizing AI in HR is the top priority in this dataset, and the fastest route is through work your team already does daily. Sidekick puts an AI assistant into everyday HR and payroll, and HR Analytics adds headcount forecasting and attrition prediction, so the role redesign conversation runs on your own data rather than instinct.
Closing a session on stage at Dusit Thani Manila, July 8, 2026.
Next Steps
Where This Goes Next
Cliff Eala closed his keynote with a sequence that maps cleanly onto the AI Maturity Journey in the State of HR Report.
"Automate the predictable, upskill the unpredictable, and sharpen human judgment."
Step 01
Automate the predictable
Tasks built on data, patterns, and rules are the first candidates.
The boundary between automatable and augmentable work moves, so it benefits from a named owner rather than an annual review. Eala's closing instruction: identify the boundary, then redesign roles around it.
Read more +
Step 02
Upskill the unpredictable
AI tends to raise the floor. Training designed around that pattern reaches further than training that starts at the top.
Research on roughly 5,000 customer support agents found a 14% average productivity gain, with about 34% for novice and lower-skilled workers and minimal impact on the most experienced.7
Pairing matters as much as the skill. Research analyzing 962 skills found a skill's value is largely determined by complementarity, and that AI skills carry roughly a 21% wage premium because they combine with so many others.8 This aligns with SOHR Key Finding 02, where only 35% of respondents receive role-specific AI training.
Where Sprout fits:Disprz handles the learning and skills side, and connects back to Sprout's core HR and payroll platform so skills sit against the employee record rather than in a separate catalog.
Read more +
Step 03
Sharpen human judgment
Judgment is required where information is incomplete, inaccurate or untimely, which describes most real decisions.
Four questions from the keynote support this in practice: What assumptions is this answer making? What data would change this recommendation? What is the strongest opposing argument? Is this data representative of what I am trying to understand?
Read more +
The skills the closing panel named
Asked which human skills are non-negotiable as AI absorbs routine work, four panelists answered without conferring. No two picked the same thing.
One inside each panel, answered in the moment rather than after the fact.
0
Priority survey responses
About half of all checked-in attendees. Up to two selections each, so selections exceed respondents.
524
check-ins
Who was in the hall
Attendees spanned every level, from individual contributors to the executive team.
Officer / Specialist43.1%
Manager24.4%
Director / VP / C-Level22.7%
Supervisor / Lead9.7%
That mix matters when reading these findings. This is not purely a leadership view, and not purely a practitioner view. It is both, in the same hall, on the same day.
Four ways we captured the day
Live polls
Four Mentimeter polls, one inside each panel, answered in the moment. Final in-session tallies (n = 297 / 224 / 202 / 100).
Priority survey
An attendee survey on first-move priorities, n = 244 (about half of those checked in). Up to two selections each, so selections exceed respondents.
National survey
The State of HR Report: 3,500+ respondents across 13 industries, fielded May 8–29, 2026. Medians use the report's interpretation bands. Self-reported and correlational.
Session recordings
Every quote is drawn from the recordings, lightly edited for readability. Figures attributed to speakers are their own citations.
References
Sources
Every external figure in this round-up is sourced below. The live polls and priority survey are our own, collected on the day.
State of HR Summit 2026 · Expert Round-Up
