Four sessions, one room of workforce leaders, and a day that kept arriving at the same answer from different directions.
Hot Topics 2026 ran at Indeed HQ in London on 17 September. The brief was deliberately wide: the labour market, how it is about to be policed, and AI.
Matt Burney, Senior Talent Strategy Advisor at Indeed, opened on the forces reshaping hiring. Matthew Taylor, Chair of the Fair Work Agency, set out what the new agency actually does. Neil Carberry, CEO of the REC, was interviewed on the next era of work. Lucy Adams, CEO of Disruptive HR and formerly HR Director at the BBC, closed on leading through disruption. The sessions were prepared separately and converged anyway. What follows is what was said, not who said it.
The matching is broken, not the supply
58% of employers told Indeed and YouGov that hiring had gotten harder over the past three years. Asked the mirror question, 49% of jobseekers said the same about finding work. If only one side reported difficulty you could fix it with volume. Both sides reporting it at once means the matching is at fault, and more candidates through the same filters makes it worse.
The filters are the part you can move. 63% of jobseekers expect to be judged on what they have actually done rather than how closely they match a traditional profile, and where employers dropped degree requirements, the Burning Glass Institute found those hires stayed in role 20% longer. In flexible hiring that scales: a permanent advert with a proxy in it costs you one hire, while a standing brief with a proxy in it costs you a season, because your agencies match against it on every shift.
AI takes the task, but the job still has to exist
Indeed has scored around 2,900 skills on how far generative AI transforms each one. What comes out is a distribution rather than a cliff, and shift-based work sits furthest from the edge because presence is the job.
The sharper point arrived from the other direction: we keep confusing tasks with jobs. Repetitive work automates first, and repetitive work is how people have always learned a business. Remove it without redesigning the job and you have taken out the bottom rung and kept the ladder. Temporary work is one of the few remaining ways to rebuild that rung, and it is also how you move before you are certain about AI. Redesign a role, staff it on shifts, and find out whether it works without betting a headcount.
Compliance is mostly a question of who supplies you
The Fair Work Agency is deliberately small, and around a tenth of its caseload involves any intent at all. The rest is unclear payslips, deductions, and the real complexity of working time for people who do not work a fixed week. A business that comes forward about its own error will no longer be named, which makes self-reporting cheaper than being found.
For anyone using temporary labour the practical read is narrower than the legislation looks. Deductions, holiday pay and licensing are your supplier’s process rather than your policy, which makes who supplies your workers a compliance decision and not a procurement one. Ask what your agency can evidence before you need it.
Peak is the rehearsal
Filling the same Christmas shifts now takes roughly twice as many candidates as it did ten to fifteen years ago. Nothing about the work changed. The workers did. They hold several agency apps, they can see what each one is offering, and they choose shift by shift: the cafe today, the warehouse tomorrow. That is choice exercised with better information, and it will not reverse.
Which makes how you treat people a fill lever rather than a values statement. The day offered two neighbouring warehouses on comparable pay with very different fill rates. The site that filled better let agency workers use the canteen, and its managers treated them as part of the team. None of that appears on a rate card.
Change lands small, or it does not land
McKinsey puts the failure rate of change programmes at 77%, and the reason given is that they are done to people rather than with them. What works is smaller: understand the pressures on the person you need to move, put the consequence in their language rather than yours, and give them something they can do this week. Then start with the willing, because most of us pour our energy into the loudest objector and get nothing back.
How Indeed Flex answers this
Every takeaway above points the same way. Build the bench before you need it, screen on evidence rather than proxies, and see across your agencies in time to act. That is the job Indeed Flex was built for.
Build the pool early. Workers who learn your site now are the ones covering your busiest weeks. Rate a worker five stars and your shifts reach them before anyone else, so the bench you build this autumn is the one that turns up in December.
Screen on evidence. Every Flexer is interviewed one-to-one and rated by the sites they work, so your brief can ask for what the job needs instead of a proxy for it.
Keep your agencies, lose the blind spots. Hours, fill and spend sit in one view across every agency you use. That is the data the new rules will ask you for, and the data that tells you which site has a treatment problem rather than a rate problem.
Across our network, average shift fulfilment runs at 95%, and 92% of Flexers are rated four stars or more by the sites they work. When a role ends we redeploy that worker elsewhere on the platform rather than lose them, which is how a pool holds its shape between peaks.
The throughline
Separate briefs, prepared separately, all ending in the same place: the change that works is small, concrete and now, not a programme with a name. Pick one hard to fill role and separate its evidence requirements from its proxies. Put a number on the cost of not acting and take it to your board. Audit how agency workers are treated on site before peak. Get hours, cancellations and fill rates into one view.
You will not get certainty. You can get ready, and that turns out to be what thriving through change means.
