The BPO sector lies at the frontier of AI driven business transformation. It has always been a labour-intensive industry that has, more recently, increased revenue margins by offshoring much of its workforce since the 1990s. A subsequent nearshoring trend was exacerbated by the Covid-19 pandemic and then quickly followed by the early introduction of AI driven automation.
Just as the industry had faced the labour arbitrage of offshoring, the BPO industry faces a new kind of labour arbitrage, namely AI vs human.
Most estimates put the number of contact centre agents at up to 13 million globally, or tens of millions including wider BPO roles. Hiring data is mixed and varies significantly by geography and where a market is in the AI implementation cycle. For example, US contact hiring is down by 5% in 2026, according Gitnux, while global BPO hubs like India and the Philippines are seeing significant headcount growth.
For Debra Maxwell, CEO of BPO services company ArvatoConnect, the binary proposition of AI versus human headcount simply doesn’t ring true. It’s less about replacing humans, she says, and more about how the work itself is evolving. And taking employees on that journey is a key component of her leadership strategy.
ArvatoConnect is a digital transformation partner to enterprises and the public sector, helping them redesign how they interact with customers and run back-office operations. The BPO industry has a longstanding problem with notoriously high attrition rates. Automating low value tasks and upskilling staff to take on more complex and rewarding work may address this problem. Furthermore, Maxwell says a better employee experience leads directly to a better customer experience.
“Our leadership approach to AI has been one of inclusion rather than exclusion,” she says noting that companies deploying AI for straightforward labour arbitrage are missing the opportunity to create enhanced outcomes for their customers.
AI versus human is the wrong question
Maxwell says the most common mistake she sees in AI deployments within her industry is starting with the wrong question: what can we take out that humans do at the moment? The right approach would be looking at what customers need at every point in the journey, and where can AI augment or make it easier.
“Customer journeys are super complex, and people think that everything is predictable, and it’s not,” says Maxwell. Assuming that AI can address complex customer problems leads to frustration and problems.
“We use AI tools to help our agents do their jobs better, finding information faster, rather than automatically just replacing them, and assuming that the AI can simply take over all of these very complex problems.
The key is deciding where to automate and where to augment. For less complex tasks, ArvatoConnect customers are currently automating up to 70% of customer interactions, and Maxwell envisages a point where very simple, scripted processes could reach 100% automation. “But overall, as AI tools are being developed, we think the optimum model is likely to be around 70% digital and 30% handled by people, and that’s particularly where emotional intelligence or sensitive conversations need to happen,” says Maxwell.
As more basic tasks and roles become displaced, tasks that require human interaction to get a good outcome for a client will become more effective. “We certainly are seeing an uptick in more skilled roles, either in developing AI products, and less hiring of entry level type roles,” says Maxwell.
Bringing staff into AI transformation
Maxwell describes contact centre work as a “tough job” that could become easier and more rewarding if basic tasks were automated. And she recognises that employees are more likely to support the deployment of AI if they’re involved in decision making around when to switch from an AI customer interaction to a human one.
“I think everybody assumes, that our customer service agents are all terrified of AI. I think that there’s a way more acceptance than people realise. They see it as an advantage, certainly if it’s communicated well and managed properly,” says Maxwell.
Once a year, Maxwell visits all ArvatoConnect’s sites to hold town halls and speak directly to employees. And she’s having quite a big conversation around AI. The feedback has been more excitement than anything, something she puts down to the company providing opportunities for staff to get involved.
Maxwell has also created a working group she calls The Big Debate which includes nominated employees from across the business meeting with Maxwell once a quarter to share anonymised employee concerns. “And AI has been driving that conversation for at least the last two years,” she says. The most common concerns are around how AI will affect jobs, what training will be offered and its environmental impact.
Employees designing their own AI workflows
But perhaps the best demonstration of employee contribution to the company’s AI journey is an initiative called MyCreationStation, which has seen some of the company’s staff actually build their own chatbots in copilot. “Because we give everybody access to copilot, they are building their own AI bots themselves in copilot,” says Maxwell who has seen the initiative create a lot of excitement among employees about AI.
These are not senior managers. They include some of the most junior staff. Maxwell cites an example of work done on automating Financial Viability Risk Assessments for Crown Commercial Service. An employee submitted an idea through MyCreationStation about automating slower manual processes which led to the deployment of an AI-enabled tool that analyses financial statements and generates reports, freeing assessors to focus on professional judgement.
The tool has already delivered 97% accuracy in assessments, significant reductions in handling times, higher employee engagement, as colleagues now use their expertise to review and interpret outputs rather than input data. The project was such a success that the model is now being expanded to other areas of the business.
Maxwell’s view is that members of staff are better placed to think up really creative ways of using AI every day in their jobs, right. “That’s just logical, I don’t understand companies that hold back from letting their employees experiment with AI,” says Maxwell. And Arvato Connect’s attrition rate is lower than that of the general BPO market. Employees benefit as they are continually being told that AI will not take your job but someone who knows how to use it will.
This is certainly true of contact centre jobs. According to research and analysis company GlobalData, nearly one in five (and rising) new contact centre hires globally now focus on AI expertise. AI roles are growing at over 40 times the rate of overall contact centre hiring. And this proportion is substantially higher in major outsourcing hubs like India and the Philippines, and in digital-native companies, where it can exceed one-third of all new postings.
In light of this, Maxwell’s approach of bringing employees in on the AI transformation journey will be a net benefit to employees, the business and ultimately customer outcomes.
“ArvatoConnect CEO Debra Maxwell on AI transformation in high-churn BPO operations” was originally created and published by Verdict, a GlobalData owned brand.
The information on this site has been included in good faith for general informational purposes only. It is not intended to amount to advice on which you should rely, and we give no representation, warranty or guarantee, whether express or implied as to its accuracy or completeness. You must obtain professional or specialist advice before taking, or refraining from, any action on the basis of the content on our site.
