Gareth Newbery — People, Talent & Change Management Leader

The policy let associates count up to fifty hours a year of that work towards bonus eligibility, once their billable target had been met. Alignment was built into how it was recorded: the codes mapped to specific strategic priorities, and people had to say what their work supported. That worked in both directions. People could see what the firm was trying to achieve and where their own effort sat against it, and the centre could see what was actually being done on the ground for each priority, including smaller local pieces of work worth pulling into something regional or global.

The idea came out of conversations I'd been having with Talent; the Director of Talent Management took it to the Regional Management Committee. I worked on the wording, reviewed it before release, and revised it after launch on feedback from the people using it, mostly to simplify what was being asked of them. I owned the inclusion strand.

Getting it used meant working across the region: office and practice group leadership, committee calls, and the champions network, where I ran sessions every six weeks and heard what was and wasn't landing well enough to adjust the guidance. I verified what was recorded — the codes, and whether the commentary described work that qualified and connected to the priority it was filed under. Where it didn't I went back to people individually, and a small number had to undo what they'd recorded. Speaking, organising and following through on actions counted; attendance on its own did not. That line went in early, because a loose precedent in year one would have been hard to tighten later.

Associates were in scope. Partners were not.

I built it. That started with the shape of the function itself — which roles, at what level — and phased hiring plans built against each office's own growth plan rather than against whatever vacancies happened to land. Then the operating model: sourcing strategy, employer proposition, assessment framework and ATS governance, with competency-based interviewing, hiring manager training and psychometric assessment brought in to make the quality of a hiring decision consistent across countries.

I hired the recruiters into the function, set their objectives, ran their reviews and coached them as it scaled. I defined the service levels the function would be held to and reported against them: time to hire, cost per hire, shortlist turnaround, interview-to-placement ratios, and hiring manager satisfaction at 30 and 90 days. Being measured on satisfaction three months after someone starts changes what you recruit for.

In twelve months the function delivered 280 hires, more than 85% of them sourced directly. One office launch needed 40-plus hires in two weeks, with sourcing, assessment and onboarding running at once. I led the search for the firm's first General Counsel in the region myself, outside the standard process. Workday Recruiting was deployed across the region during that period, with configuration, training and workflow governance, so the reporting meant something.

I was usually brought in by someone rather than assigned — an office leader, or a colleague from another team who had something they couldn't take anywhere else. The work started with diagnosis: what was happening, who it affected, how much of it was one individual's behaviour and how much was how the office had come to run itself. That stage is mostly listening, and it takes longer than anyone wants it to.

The intervention came after, built for that office and its market rather than lifted from a regional template, and agreed with the leadership who would have to hold it once I'd gone. Some of that was direct work with the leaders themselves. Some was changing how a process ran, so that the outcome people were complaining about stopped being produced.

The measure was the following year's culture review. Offices I worked with showed improvement across successive annual reviews, which is slower and less flattering than a completion rate, and the only measure that tells you whether anything held.

I've kept the specifics out. The situations would be identifiable to the people who were in them, and most of them are still there.

My part was extending that system out to five more offices across Europe, the Middle East and Africa. I worked with the firm's account manager at the vendor to establish how it would be done, then ran the implementation plan through him and through each local office. Most of the effort went into training the recruiters in each one and setting up workflows that recorded a candidate moving through the process, which those offices had never captured anywhere.

They needed more than a system. I also upskilled those recruiters on LinkedIn Recruiter and Boolean search, and on how to use what they had properly. For several offices that was a significant change in how recruitment got done.

It came with a name and a logo that told you nothing about what it did, so rebranding it was the first job. After that it ran as a recruitment business like any other, at a smaller scale and with me doing all of it: advert response, working out who was suitable, and supporting teachers through a move that took them a long way from home.

The clients were schools in China, Vietnam, Thailand and Saudi Arabia, some of them large international brands and some placing one or two people a year. I dealt with them directly, mostly by video call, usually with an HR manager on the other end who had a firm view of what a good teacher looked like.

The difficulty in that market is that a school is hiring someone it has never met, who then moves continent to take the job. I designed and ran an online training course candidates completed before placement, so schools received people who had been prepared rather than only screened.

I also ran the P&L. That was the first time I held full commercial responsibility for a business rather than a recruitment budget inside one.

I sold it when we came back to the UK in 2019.

The business ran its own training on what the tool could do and on prompting, which was new to almost everyone then. Some of the team went, some didn't, and several picked it up gently or left it alone.

What I worked on with them was narrower: what they were putting in, what they were getting back, and how to close the gap between the two. Which tasks were worth starting with — a project plan, a first draft of a presentation — and how to ask for something so that what came back was usable for the business rather than a rewrite. The tooling then ran to a research agent and an analyst agent, with nothing automated.

None of this was part of my role. It's where the AI work I've done since started.