Digital labor: how AI and human agents can work side by side
The major contact center platform providers made one thing clear at Enterprise Connect 2026: so-called agentic AI has become a core component of customer experience. The idea is to use AI agents for both smart routing and direct handling of simple, routine inquiries. The common view among these executives is that digital labor is not just about cutting costs, but about improving human work and customer interaction at the same time.
This shift raises some important questions: what changes in the daily routine of frontline workers when AI-based automation becomes the default? And beyond that, how do these professionals start interacting with task-specific or function-specific artificial agents that operate in parallel with the human team?
To understand this relationship in practice, it makes sense to look at the companies building the platforms that power a large share of today’s contact centers. Five executives who live this reality up close shared how they see the partnership between people and AI in customer service:
- Neville Letzerich, CMO at Talkdesk
- John Finch, Global VP of Product Marketing at RingCentral
- Jasen Williams, SVP of Corporate Marketing at Verint
- Craig Walker, CEO at Dialpad
- Gautam Vasudev, SVP of Product Management for Agentforce Service at Salesforce
Below is an organized overview of these leaders’ perspectives, staying true to what they originally said while adding extra context for those closely following the rise of AI in CX.
Specialized AI to reduce queues and scale expertise
Neville Letzerich from Talkdesk explains that the company launched its CXA platform about a year ago. The core idea is straightforward: create an environment where different AI agents can operate either standalone or together, like a coordinated team of digital agents.
Letzerich highlights a point any contact center leader knows well: many human agents have highly specialized skills. In traditional operations, this means putting the customer on hold to wait for the one person who knows a specific system, a more complex product, or a critical business rule.
With agentic AI, Talkdesk is working to train AI agents on those same niche skills. Instead of routing nearly all critical demand to the same small group of human experts, the system starts to absorb part of those interactions, especially repetitive cases that follow a clear pattern. According to Letzerich, when this is done carefully, two things happen at the same time:
- Wait times drop, because fewer customers are competing for the limited number of human specialists.
- Customer satisfaction rises, since issues are resolved faster.
The key is not to replace the specialist, but to emulate their expertise and make it available at scale. Knowledge that previously lived only in the heads of a few agents gets codified into workflows, structured knowledge bases, and AI models trained with real-world examples. That changes the scale game without erasing the human role.
Letzerich does not ignore the impact: some jobs are likely to be eliminated. But he points to another side of the story: the savings generated in the contact center can be reinvested in areas like sales, marketing, and product development, accelerating the company’s growth. In his view, the contact center stops being just a cost center and becomes more directly connected to revenue generation.
Redistributing roles instead of mass layoffs
John Finch from RingCentral follows a similar line, but emphasizes redeployment of talent. In his view, each organization will seek a different kind of ROI with AI. Some want straight cost reductions; others prefer to use automation to free people up and move them into more strategic roles.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Finch points to a very common example: agents who grew with the business and understand how the company works end to end. Instead of simply letting these people go when AI starts taking over a large portion of customer service, he suggests moving them into roles where that accumulated knowledge creates even more value.
In a tech company, for instance, a veteran agent often knows more about recurring issues, frequent questions, and product usage bottlenecks than many people in other departments. In logistics or shipping companies, frontline staff who talk to customers daily have a clear sense of where the biggest operational pain points are. From Finch’s perspective, it makes much more sense to put these people into internal roles focused on improving processes, products, and communication instead of just cutting them from the headcount.
Another strategy he highlights is gradually shrinking the team through attrition, meaning not backfilling every role that opens up due to natural turnover. This way, part of the workforce transitions into new roles within the organization while AI takes over more front-line interactions. As a result, companies can:
- Handle more contacts with the same number of people.
- Move experienced agents into higher-impact roles.
- Contain more interactions right at the entry point, without always needing a human.
When well orchestrated, the outcome is a healthier balance between automation and human talent, avoiding unnecessary internal shocks and making better use of people who know the business from the inside.
Rebalancing capacity and quality at scale
Jasen Williams from Verint brings in another key data point: according to him, there still is no aggressive, widespread wave of agent layoffs driven by AI. What stands out much more is a rebalancing of how capacity is used.
In practice, the question customers bring to Verint is:
What do I do with the capacity I’m freeing up through automation?
The answers vary:
- Extend service hours without growing the team.
- Offer more robust support to premium customers.
- Reallocate part of the team to sales and retention roles.
Williams also points to a strong impact at the management level. Traditionally, quality teams could review only a small fraction of interactions — around 1% to 3% of calls, depending on the size of the operation. With AI, a new scenario emerges: models can analyze virtually 100% of calls, from chat to voice.
He cites the case of a large fintech that uses Verint’s so-called quality bots. According to Williams, those bots do the equivalent work of roughly 1,200 quality managers. In other words, it would be impractical to hire and maintain a structure of that size just to listen to and evaluate calls, but AI makes that level of coverage possible.
The direct effect:
- Compliance costs drop, because the company can monitor everything, not just a tiny sample.
- Service quality goes up, since deviations and issues are identified and corrected much faster.
In Williams’s view, the recurring pattern is this: a company starts using AI in one business unit, one region, or one product line. Once it proves value there, it expands to more use cases, spinning the learning flywheel in an incremental way.
When 100% automation works (and when it doesn’t)
Craig Walker, CEO of Dialpad, gets straight to the point about a common hype: companies that promise fully automated service, with no humans in the loop. He acknowledges that this model can work, but only for a small slice of real-world use cases.
According to Walker, in the overwhelming majority of scenarios, customer service requires a mix:
- Part of the flow is handled by AI agents acting on their own.
- Another part needs a handoff to a human agent when the complexity goes beyond a certain threshold.
In this transition, AI has a decisive role: setting the stage for the human. Instead of just transferring the call, the system gathers context, organizes information, summarizes the conversation history, and delivers that package to the agent. So when the person comes on the line, they don’t need to ask for everything all over again. That speeds up resolution and significantly improves the customer’s perception of care.
Walker specifically calls out voice-based use cases. A strong example is healthcare. The age group that consumes the most healthcare services tends to be people over 65, who prefer to call, talk, and ask questions over the phone. For this audience, a text-only chatbot experience is usually not enough.
In that scenario, the ideal setup is for AI agents to handle most of the voice interaction in a natural way, and then hand the case smoothly to, say, a doctor’s assistant, with all the context already summarized. In Walker’s view, that is where the real magic of combining agentic AI and human work happens: each side does what it does best, without friction for the user.
Supervisors, observability, and the importance of the human path
Rounding out the group, Gautam Vasudev from Salesforce sums up the moment in one line: the near future belongs to humans and agents working together, not one completely replacing the other.
He highlights AI’s role especially in two areas:
- Real-time supervision
- Smarter workforce management models
In Salesforce’s architecture, AI agents can call on a supervisor in the same way a human agent would. If the system notices that the conversation is going off track — for example, rising friction, negative sentiment, or deviation from the expected flow — it can flag it and request an immediate escalation. The supervisor, in turn, can jump in and out of the conversation, taking over when needed and handing control back afterward.
Vasudev also stresses another essential point: when deploying AI agents in customer service, a company must ensure from the start that there is a clear escalation path to humans. Leaving users stuck in an automated loop with no way out is seen as a poor and often unacceptable experience.
In the company’s philosophy, both human agents and supervisors are treated as personas within the engagement flow, with routing rules, observability, and intervention logic designed for each. AI is not outside that ecosystem; it is embedded in it as another digital persona, with its own behavior, limits, and escalation triggers.
The common thread: customer experience, not automation for its own sake
What ties these five executives’ perspectives together is a mature view: agentic AI and digital labor are not ends in themselves, but means to build a better experience for customers and workers. When done right, they reduce queues, extend service hours, improve quality control, and create room for human agents to move into more specialized, strategic roles.
At the same time, they all recognize a few basic principles that don’t change:
- Not everything can or should be fully automated.
- The escalation path to humans has to be simple and transparent.
- AI should learn continuously from human experts, not operate in the dark.
- Managers gain new visibility by analyzing 100% of interactions, not just samples.
- The ultimate focus remains combining efficiency with respect for the customer journey.
At the end of the day, the most interesting form of digital labor is not the one that boasts about removing every person from the front line, but the one that can balance AI and human work to deliver something that feels simple, fast, and reliable from the customer’s perspective. In that scenario, human agents stop being seen as a cost to be eliminated and start being treated as AI partners — essential for teaching, supervising, correcting, and, above all, representing the brand in moments when empathy and nuanced judgment make all the difference.
About the authors of the original content
The article that inspired this analysis was originally written by Lisa Schmeiser and Matt Vartabedian, both with long careers covering technology and enterprise communications.
Lisa is editor at No Jitter and Workspace Connect, with more than two decades in technology journalism, including stints at ITPro Today, InfoWorld, and Macworld. Over the course of her career, she has been nominated for major industry awards for her in-depth tech reporting and continues to appear on podcasts focused on this space.
Matt, a senior editor at No Jitter, covers AI topics — including predictive, generative, and agentic AI — always linked to unified communications, contact centers, and the digital workplace. Before returning to journalism, he spent about 20 years as a wireless industry analyst, producing reports, presentations, and articles grounded in research and data analysis.
The perspectives organized here closely follow the points raised by these authors and the executives they interviewed, with additional context and unpacking for readers closely tracking how AI is reshaping customer service.
