‘An entirely fluid workforce’: WPP plans AI-allocated talent across 120,000 workforce plus external partners – with skills to trump job spec per holdco’s global AI chief
WPP is using AI to allocate agency staff based on skills, not job specs – and is working on a rollout across not just its 120,000-strong workforce, but to external partners too. The holding company's AI chief Daniel Hulme modelled bumblebee brains as a PHD student. Now he's applying it to the workforce, after honing it on distribution for Tesco trucks and PwC's auditors in the UK. He thinks AI can make a much smarter fist of workforce optimisation than humans, provided the humans set the right parameters. Optimising to one metric, he warns, risks a major backfire.
What you need to know:
- WPP AI chief Daniel Hulme is applying technology developed by his firm Satalia to optimise fleets for the likes of Tesco and workflows for 5,000 auditors at PwC to WPPs talent pool.
- The ultimate plan is to allocate all of WPP’s 120,000 global staff via AI – with people deployed based on skills, not job spec – and then extend that approach to the third parties with whom the holdco works.
- Hulme, who modelled bumblebee brains as part of his PhD, thinks a hive mind can change the way the multinationals operate, creating a mass, fluid talent pool that gives each agency brand infinitely more firepower.
- He acknowledges there are risks in getting the approach wrong. Optimising to metrics such as productivity, for example, will likely lead to burnout and people leaving.
- But Hulme says the evidence coming out of PwC is that applied via multiple metrics, the opposite is true.
It's been deployed in a number of agencies … but also there's also an initiative to be joined up at the top across the board. So we can optimise individual agencies, but the Holy Grail is to be able to have this fluid workforce that essentially can be used across all of the different [WPP] agencies in a much more utilised way.
Mightier hive
WPP has embarked on a mission to allocate its 120,000 staff via AI and create a fluid workforce that transcends its agency brands. It's aiming to put talent-pooling initiatives by rival advertising multinationals in the shade – and AI chief Daniel Hulme reckons the machines can help staff avoid burnout while giving people broader scope to do what they are good at, not just what's listed on their job spec.
The template has been built by AI specialist Satalia, which WPP acquired in 2021, wrapping it into Wunderman Thompson. Hulme, the firm’s founder and CEO, is now WPP’s AI chief – and he’s applying the technology initially licensed by the likes of the UK’s largest retailer, Tesco, to optimise delivery routes to WPP’s own army of staff. Hulme calls it AI-driven “workforce allocation” and the tech is already being used by PwC in the UK to organise workflows for 5,000 auditors, or “allocate them to demand”, per Hulme.
Hulme, whose PhD spanned modelling bumblebee brains and mathematical optimisation, is now applying that theory to construct a WPP hive mind.
He told Mi3: “We do it in a way that is significantly better than any human being. It enables the business, obviously, to increase utilisation, but it enables auditors to have better career development and clients to have better continuity.”
But there are risks. Just as social platforms and broader society are realising the fallout from optimising to keeping people glued to the screen as long as possible – polarisation, free-falling mental health indicators, confirmation bias bubbles – the perils of optimising to single metrics present a risk to those embarking on AI-powered workforce allocation.
“If you're focusing on one KPI, then you can move the needle significantly. But that may have an adverse effect on other KPIs,” said Hulme. “If I focused on … maximising utilisation, that might compromise on client continuity.” It would probably also lead to burnout, as Hulme acknowledged.
“It might compromise on people having to drive longer distances. It might compromise on people having time to do training. So if I just focused on utilisation, I might end up seeing adverse effect, which is people leaving the company.”
Hence the firm “focusing broadly on five flavours of KPI”, per Hulme. “One is the business, which usually means money. The second is employee wellbeing, third is the client wellbeing, fourth is environment, and then your ecosystem – your partners and suppliers. You need to take all of those things into account and make sure that you're solving it holistically across all of them.”
Most companies are not set up to manage that kind of change management across multiple KPIs, Hume accepted. But he’s backing AI to ultimately succeed in making best use of human resources.
Instead of being allocated by role, people are allocated based on skills. Instead of you [the worker] being a [given a single] role, you are now a plethora of different skills so that you can be allocated in a much more fluid, granular way.
Corporate body fluid
Hulme said the shift toward “fluid” AI-powered workforce allocation for WPP is a project without an end-date, though said it was likely “a five-year journey” to be fully operational across the group. But he said the mission is already well underway.
“It's been deployed in a number of agencies … but also there's also an initiative to be joined up at the top across the board,” said Hulme.
“So we can optimise individual agencies, but the Holy Grail is to be able to have this fluid workforce that essentially can be used across all of the different agencies in a much more utilised way.”
There's a few hurdles to clear yet. “As you can imagine, in an organisation like WPP there's challenges around how things are structured, earnouts and all that kind of stuff. So you have to navigate all of that. But that’s the aspiration.”
Can resources at a multi-disciplined holding company that spans communications, creative, media, digital, CX, commerce, digital transformation and the rest be deployed in the same way as PwC’s auditors, which specialisms aside, largely do the same thing? Hulme thinks so, backing digital twinning of each staff member to map their skillsets – which are signed off or amended by the individual worker – and apply them accordingly.
“So instead of being allocated by role [or job title], people are allocated based on skills,” said Hulme. “Instead of you [the worker] being a [given a single] role you are now a plethora of different skills so that you can be allocated in a much more fluid, granular way.”
My goal will be an entire fluid workforce, not only across WPP but even to try and make some of the boundaries more porous. So the consultants that we work with outside of WPP, I want to be able to leverage that capacity and capability in a much smarter, more fluid way.
Swarm up act
Hulme's looking to take that model beyond WPPs own walls.
“My goal will be an entire fluid workforce, not only across WPP but even to try and make some of the boundaries more porous. So the consultants that we work with outside of WPP, I want to be able to leverage that capacity and capability in a much smarter, more fluid way.”
What implications does that have for business units as they stand today?
“I think organisations are moving towards a matrix structure where they have their offerings, but then they have this fluid capability that swarms across those offerings. And so what you see [as a client] is the different offerings that are presented to you. But what sits underneath that and how that's delivered is much more fluid and organic.”
How far away is that?
“Ultimately it's up to [WPP CEO] Mark Read and the rest of the leadership to understand how to optimise for that offering. Which is what they're doing.”