top of page

No-Bull Q&As 🚫💩

Candid interviews with BPO & CX industry insiders and experts...without the bull!

Bryce Maddock: Industry Headwinds and the Future for BPO

Bryce Maddock

We're excited to welcome back our good friend Bryce Maddock from TaskUs for BPO Bullhorn's final No-Bull Q&A of 2025.


Bryce is spilling the beans on:


📈 Growth challenges in BPO and how to adapt

🤖 The true impact of generative AI on the industry

🎓 Advice for training and upskilling the agents of tomorrow

💡 ...and much more.


Let's get into it!


---


Q: BPO appears to be slowing down this year. How is this impacting the industry, and what can other providers do about it?

A: The industry’s noticeable slowdown in growth has led many providers to cut prices to irrational levels and created some margin pressure. This also raises concerns about a potential race to the bottom. Companies reducing rates will often soon face the need to lay off staff and cut operational investments, making it even more difficult for them to deliver operational excellence and real value.


At TaskUs, we are not immune to these challenges. In 2023, we experienced a 4% revenue decline, which was the first time in our company’s history, not only that we hadn’t grown but that we hadn’t grown by double digits.


We fought hard for 18 months to turn this around. We invested in specialized service expertise to increase the value of our services to our clients, we invested in technology to automate processes and drive efficiencies, and we increased spending on sales and marketing. Fortunately, these investments have paid off, and in Q3 2023, we returned to accelerating double-digit revenue growth of 13.2%. Q3 2024 was the best quarter in TaskUs history.


The turnaround, I believe, is a signal for our industry. We have to focus on differentiation, invest in operational excellence, and develop deep expertise in specific service lines and industries. At TaskUs, we strive for all of this, but we’re also positioning ourselves as a premium provider and avoiding the well-trod path of price-cutting. We believe this combined approach will help us sustain long-term success.


Q: What’s been causing this industry slowdown? How much has the rise of generative AI been playing into it?

A: In 2022, we saw headlines about the "year of efficiency" and significant layoffs across major tech companies. That restructuring impacted both their managed service providers and internal teams as everyone was forced to innovate and become more efficient. This aggressive drive for efficiency led to better product design within the digital customer experience space, eliminating the need for direct human interaction in thousands of cases.


However, we’ve also seen growth in trust and safety services in areas like content moderation, financial crimes, and compliance, and demand for AI services is growing, driven by the wave of funding for LLMs (large language models) and the broader generative AI market.


While I don’t think the current slowdown among major players is due to generative AI, declining stock prices across the BPO space suggest that investors fear generative AI could significantly reduce volumes in the next few years. That uncertainty is creating a lot of tension in the market, but the big question about the long-term impact of AI remains.


Q: To what extent has TaskUs seen AI impact BPO staffing and service delivery so far? What’s the real-world impact?

A: Nothing substantial so far, but we’re certain we will see noticeable impact in the next few years. Although 75% of enterprises have either deployed or are planning to deploy generative AI across various functions, including customer service, most are in an estimation phase, where they’re predicting they can automate between 10% and 20% of contact volumes using this technology—and I think that’s likely accurate—but implementing these solutions is far from simple.


Using LLMs and generative AI to address customer issues isn’t as easy as training an LLM on your data and letting it handle problems autonomously. You need an agentic AI framework that integrates these tools with critical systems such as CRM platforms, internal order systems, and payment systems. This complexity is why we haven’t seen these impacts materialize as quickly as some of the hype might suggest.


That said, we’re fully anticipating a 20% reduction in volumes across traditional customer service. The companies that will weather this transformation are those positioning themselves in higher-value services that are less prone to automation. We see trust and safety services, as well as AI-related services, as being significantly less automatable than traditional CX.


Within our CX business, the vast majority of the work we do is complex. So we don’t believe it will be exposed to GenAI automation risk in the foreseeable future.


Q: If you were leading a smaller BPO, how would you look at providing services for AI companies and preparing agents for these roles?

A: One opportunity we’ve seen is that LLMs have run out of public training data from the internet, so to advance AI capabilities further, companies are investing heavily in building private datasets. That’s not a skill you can train a teammate for—it requires recruiting people with master’s or PhD-level expertise in various fields to answer complex chemistry questions or interpret Shakespeare in ways that go beyond the open internet. It’s an entirely different business model.


Another involves combining trust and safety expertise with the growing demand for adversarial testing, also known as “red teaming,” within AI services. Here, we’re focused on upskilling existing teammates and training them to handle subjective processes that require critical thinking, such as intentionally testing the model’s boundaries and weaknesses.


As more simple tasks are automated, the key question is whether there will be enough work for medium-skilled workers. No one knows the answer yet. If the only demand ends up being for PhD-level expertise to train and maintain AI models, it would be a negative outcome for our industry—and, frankly, for the world.


I believe we can take medium-skilled talent, traditionally engaged in customer service roles, and upskill them for future opportunities, like adversarial testing. I also think we haven’t yet identified all the roles that will emerge as AI adoption continues. There’s likely to be significant new demand as AI becomes more deeply integrated across industries globally.


---


Connect with Bryce on LinkedIn here to continue the discussion.


Don't miss future Q&As. Join our mailing list ---> here

bottom of page