War Room: The Complex Realities of AI & the Contact Center

The transformation of the contact center is not only determining the future of customer contact… but helping us understand the future of an entire economy.

Make no mistake. The fight to transform the customer interaction landscape is where the real action is. It could be the most consequential AI-related innovation battlefront of our time. But something else is true too… the effective deployment of AI in the contact center (and across the enterprise) might well be as complex as the technology itself.

As such, that complexity will almost certainly come with no shortage of surprises.

The Guns of August Revisited

I’ve used the battlefield metaphor for the contact center space before. In fact, I’ve been thinking about the role of emerging technology solutions in the contact center in military terms for quite some time. Exactly fifteen years ago, in September, 2010, as an analyst at IDC, I published an IDC Marketscape report on the customer care BPO competitive landscape called, “The Guns of August” (history nerds will appreciate that the First World War started in August, 1914… and that I borrowed my title from historian Barbara Tuchman’s classic account of the war).

The immediate context for my title was a U.S. economy that was suffering the aftershocks of an exploding housing bubble and financial crisis that had brought on the Great Recession. Sitting in my office at IDC, it felt like a new kind of war for survival… and reinvention… had begun across the economy… for individuals and organizations alike.

Apple's iPhone as well as Facebook and Twitter had just recently been entrenched in our daily lives… and despite the American economy’s recent near-death experience (or perhaps, in part, because of it), American capitalism seemed poised to unleash a new wave of technological innovation. It seemed like a “Big Bang” moment for an expanding universe of interactions that would create a new kind of “Experience Economy.” Elver-gleams of innovation were gathering on the horizon.

Our Unfathomable Future

When the iPhone was released in mid-2007, we tossed away our BlackBerrys and their physical buttons and started typing and swiping on glass keyboards. But we hardly could have imagined that by 2022, a company called OpenAI would upend everything with the release of something called ChatGPT… and then soon set about developing an AI-infused device aiming to displace the iPhone with the help of teams of former Apple AI researchers. Nor could we have imagined that by 2025, Facebook would have morphed into something called Meta Platforms, which would release something called “Display” glasses… which we’d soon learn to manipulate by twitching our fingers.

“The Guns of August” Marketscape suggested that as technology rushed forward, the nature and frequency of interactions would keep changing too. And if BPOs were to keep pace, they’d need a strategic approach to “technology’s role in the future of customer care.” The push to innovate would distinguish the winners and losers in tomorrow’s world. As “The Guns of August” insisted—

The utilization of technology in this industry continues to increase and change. Future success depends, in part, upon the ability of providers to adapt to, or develop and implement technology solutions that anticipate and keep pace with continuing change, industry standards, and client preferences. Products, services, pricing practices, and profits can all be affected as technologies decrease the cost or increase the efficiency of services.

Accelerating digitalization was on the horizon… and a slew of complexities would come with it. The report emphasized that, “Providers need to look out over the trenches of today's competitive landscape and be thinking how they might best rearm and innovate so as to reposition themselves for a pitiless new environment.”

Gen AI: Into the Trenches of Implementation

I thought of that report this month when reading a piece in The New York Times, “Tanks Were Just Tanks, Until Drones Made Them Change.” Tanks have been “a mainstay in battle since the early 20th century,” the article tells us. “But in just three years of war in Ukraine, tanks have evolved.” That sounds a lot like what’s been happening to the contact center in the three years since the release of ChatGPT… rapidly shifting circumstances are forcing new configurations of AI few could have imagined not too long ago.

But in 2025, the process of integrating AI into the customer support operations of businesses across industries hasn’t been easy. According to The Wall Street Journal, reality is setting-in for the consulting firms aiming to help “companies deploy the most transformative technology in decades.” Think of pure consulting firms like McKinsey, Bain, and Boston Consulting Group and the Big Four accounting firms Deloitte, PwC, KPMG, and Ernst & Young. It seems these consultants are running into a trench dividing what they claim they can do and what they can actually deliver.

For all their expertise, consulting firms are having trouble coming up with “a playbook for deploying something as cutting edge as generative AI in the enterprise at scale.” AI tech has innovated so quickly, hard-won experience simply doesn’t exist in the field. Some enterprises are in retreat—ending their engagements with consultants and bringing the building of real use cases back in-house.

Enterprises are seeking strategic partners that can help them understand the business context behind specific data requirements and provide real value. But as one chief information and strategy officer put it about the consulting firms, “They overpromised… we discovered that they really also had no idea how to do these things. They were just as good or as bad as what we would have been able to do in-house.”

Rushing into the Learning Gap

The result has been a good amount of chaos across the business battlefront. While consulting firms continue to pull in revenue from generative AI-related work (according to Gartner, global spending on such consulting hit $3.75 billion in 2024, up from $1.34 billion in 2023) “some believe consultants’ contributions to AI deployment could be more effective in four to five years, when the technology is more mature and they have a solidified playbook. Until then, many clients remain frustrated.”

As Fortune’s Jeremy Kahn points out, “Integrating LLMs into enterprise workflows is difficult and potentially expensive. AI models don’t come with instruction manuals and integrating them into corporate workflows—or building entirely new ones around them—requires a ton of work. Some companies are figuring it out and seeing real value. But many are struggling.”

This was best reflected in MIT Media Lab’s NANDA Initiative report, The GenAI Divide: State of AI in Business 2025. As Jeremy Kahn notes—

The biggest problem, the report found, was not that the AI models weren’t capable enough (although execs tended to think that was the problem.) Instead, the researchers discovered a ‘learning gap’—people and organizations simply did not understand how to use the AI tools properly or how to design workflows that could capture the benefits of AI while minimizing downside risks. Large language models seem simple—you can give them instructions in plain language, after all. But it takes expertise and experimentation to embed them in business workflows.

Right on queue, reinforcements are on the way. Startups such as Distyl AI and specialized companies like Cognizant are also scrambling to offer services to help enterprises integrate AI. In fact, Distyl AI recently won a contract to help TMobile build out AI efforts on its customer experience team, helping workers manage workloads from customers.

According to Distyl CEO Arjun Prakash, Fortune 100 companies need to know that using AI inside of the company “isn’t just an upgrade like going from Windows XP to Windows Vista. It’s a little bit of a reorganization and rethinking of your company’s architecture—and that rearchitecturing is something that requires a partner who was born at the time of the models.”

As enterprises scramble to bring Gen AI into the interaction battle space, they’re under fire from unexpected challenges. The realities of the AI Boom are hitting many businesses like mortar rounds. 

The Latest Battlefield Innovation: Agentic AI

Now—amidst this jumble of promise, confusion, and uncertainty—along comes a glamorous new weapon, a nascent technology in the form of agentic AI… which has quickly become a top priority across industries. “Agentic AI is the buzziest new tech tool on the scene,” says McKinsey, “and it offers thrilling capabilities. While Gen AI mostly functions as an adviser by providing well-informed counsel, agentic AI can function more as a direct report by accomplishing tasks.”

AI agents, adds McKinsey, “have the potential to automate complex business processes—combining autonomy, planning, memory, and integration—to shift Gen AI from a reactive tool to a proactive, goal-driven virtual collaborator.” As MIT Sloan Management Review puts it, “Although there is no agreed-upon definition, agentic AI generally refers to AI systems that are capable of pursuing goals autonomously by making decisions, taking actions, and adapting to dynamic environments without constant human oversight.”

According to Salesforce's Agentic Enterprise Index, “the top three most popular use cases for AI agents are customer service, internal business automation, and sales.”

A Leap in Complexity

The technology is indeed full of promise. Business software company Workday just announced new AI agents for human resources and finance. According to Carl Eschenbach, Workday’s chief executive, “We’re taking a very targeted approach: We have the cleanest set of data that’s highly curated for HR and finance.”

In the telecom space, Ericsson offers a vivid example of agentic AI’s potential. The company recently announced an upgrade to its network chatbot with agentic AI, claiming it’s the industry's first 5G agentic AI virtual expert. “Ericsson added a Generative AI chatbot to NetCloud in January, dubbing it the AI-based NetCloud Assistant (ANA). This enhanced version transforms it from a user-prompt-driven tool into one that can handle more complex workflows and can, at an administrator's direction, take automated decision making to the next level.”

Ericsson’s specialized AI agents will be released in stages, tailored for troubleshooting, configuration, deployment, and policy tasks that aim to improve network performance. The intention is to lay the foundation for “fully autonomous, self-optimising 5G enterprise networks that can power the next generation of enterprise innovation," helping Ericsson accelerate digital transformation across industries.

As Nick Wood of Telecoms .com says, “This represents quite a leap in complexity from a network administrator's point of view, which is where the new agentic AI tech should really shine, simplifying some of these management tasks and automating others, and generally making private cellular networking a less daunting prospect.”

A Leap too Far?

That kind of battle plan is all well and good in the context of telecom network optimization, but things get more complicated when customer-facing realities come into play. As with Gen AI writ large, implementing and deploying customer-friendly systems is a significant challenge for agentic AI. The technology is failing to meet expectations due to a range of complex factors, including—unrealistic expectations, poor use-case prioritization, issues around data quality, and governance challenges.

And as Kearney points out, while the potential of agentic AI is compelling, it’s often vexing to get through implementations successfully—

The cost of building and running agentic AI solutions can be daunting. The specialized talent required to build these systems commands premium salaries, software and service costs can be substantial, and the infrastructure required to build, run, and maintain these systems can be expansive…. The technical complexity of agentic solutions can be overwhelming, often requiring the integration of dozens of systems, incorporating multiple AI techniques, and navigating thousands of vendor solutions in the market. Many projects stall due to IT organizations’ inability to navigate these complexities effectively.

McKinsey agrees. “An agentic enterprise transformation holds the promise of unmatched productivity,” notes the consulting firm. But while “some companies are enjoying early successes with such activities, many more are finding it challenging to see value from their investments. In some cases, they are even retrenching—rehiring people where agents have failed.”

The Salesforce Example

Salesforce offers a prime example. On the one hand, according to Salesforce's Agentic Enterprise Index, “AI agent creation among first-mover companies surged 119% between January and June of 2025, with service organizations leading the adoption of agents, with the average number of customer service conversations led by an agent growing 22 times in the first half of 2025.” Moreover—

The Agentic Enterprise Index highlighted improved customer experiences as more consumers engage with AI agents. On average, 94% of consumers opted into agent interactions. The use of AI agents in customer service is exploding. Customer service conversations with AI agents saw a six-month compound annual growth rate of 2,199% for the average business. Salesforce's survey revealed that nearly 60% of consumers who regularly engage with customer service AI agents feel the tech has become more helpful over the past year. 

On the other hand, The Information claims that Salesforce has struggled to sell Agentforce—the software firm’s new artificial intelligence for automating customer service and other functions like sales outreach and responses to IT helpdesk requests—in part because of the extensive prep work customers need to perform in order to make it work. 

Salesforce has reacted by scrambling its troops. “When customers found the initial price of Agentforce too high, Salesforce earlier this year lowered the cost per task that the AI handled and also allowed customers to pay in bulk at a discount. After agents created using Agentforce gave inaccurate responses to questions, Salesforce made several acquisitions—including $8 billion for a corporate database firm—and other investments to give the agents better data to work with.”

But it seems that “after nine months on the market, fewer than 5% of Salesforce’s more than 150,000 customers are paying for Agentforce…. And more than half of customers that are using Agentforce are still testing it without paying.” Not only that, but “Salesforce also finds itself battling startups like Sierra and Decagon that are zeroing in on its largest businesses: customer service and sales management.”

The contact landscape features operational challenges and relentless competition, a war of attrition not unlike that across today’s Ukrainian plain.

Encountering the Customer: Current Battlefield Realities

Obviously too, when it comes to customer support, once agentic tech is deployed, the reliability of agentic solutions is an absolute must. But consider the dilemma with agentic commerce—agent services that will take actions on behalf of customers. Agents from the likes of OpenAI and Anthropic are not yet reliable in completing all tasks in the real world. Companies like OpenAI, Perplexity, and Amazon—

… are painting visions of AI tools acting as personal shoppers that can seamlessly buy stuff across the internet. But the handful of AI agents that have so far been released have had trouble with online shopping, since they’re easily tripped up by variations in retailers’ product listings and checkouts, investors and founders say. It’s also tough for retailers to distinguish between an agent and a malicious bot, so some merchants are more inclined to block AI tools from checking out rather than make their sites friendlier for AI to navigate.

These early AI agents are crawling across websites trying to make sense of context, like soldiers crawling across No-Man’s Land during the First World War. They encounter logins and pop-ups… or captcha fields and prompts requesting email addresses and phone numbers for marketing. Comparison shopping on product information (pricing rates and sizing, for example) seem beyond AI agents’ comprehension. Many merchants have fraud prevention software that mistakes AI agents for bots and blocks them at checkout.

According to Amias Gerety, a partner at venture capital firm QED Investors, “For as long as there have been payments on the internet, if you were a bot, you were almost certainly a criminal, and so the entire apparatus is built to keep bots out. Everybody in internet payments has built on the assumption that bots are bad.”

Moreover, who can know what actions AI agents might take in various contexts? Rogue agents in possession of a customer’s credit card number is a nightmare scenario. The prospect of rogue agents in industries like retail, finance, government, and travel/hospitality is doubtless scaring many an organization away from deploying them in large numbers. Many enterprises remain wary of using AI agents for critical business functions.

The Fog of War

Still, just like the ongoing war in Ukraine, innovation is relentless. Google just unveiled new protocol for AI agents to make purchases on behalf of users. The protocol is built to collect mandates from users that serve as proof that users have authorized agents to take actions and make purchases on their behalf. “These mandates detail what items and price users have agreed to and allow agents to securely handle users’ payment information to complete the transaction, key issues that merchants, AI companies and payments firms have been trying to sort out in order to allow agents to handle more transactions.”

And due to the fact that LLM responses to prompts are not yet guaranteed to be accurate, Jeremy Kahn notes in Fortune’s “Eye on AI” feature that, “Some are starting to suggest that the solution may lie in neurosymbolic systems, hybrids that try to integrate the best features of neural networks, like LLMs, with those of rules-based, symbolic AI…. It’s just one of several alternative approaches to AI that may start to gain traction if the hype around LLMs dissipates.”

One startup, AUI, will soon release a new agentic language model called Apollo-1 that it claims will make AI agents more reliable than those from the likes of OpenAI, Anthropic, and Google. According to Ann Gehan in The Information, Apollo-1, “uses what’s known as ‘neurosymbolic reasoning,’ which combines aspects of the neural networks that power LLMs with older AI tech known as symbolic reasoning, which uses logic to understand the relationship between values and expresses them in code.”

The idea is that “neurosymbolic reasoning will be more appealing to businesses because of the rules and guidelines that models can incorporate into their reasoning process.” Apollo-1 will be available as a foundation model companies and developers can use to build and deploy their own agents directly. 

The tech could be groundbreaking. But will it work on the interaction battlefield as advertised? Yann LeCun of Meta Platforms is among the skeptics.

A Crisis of Confidence (Bubble Chatter Intensifies)

Since September, 2010, the call-to-arms of “The Guns of August” Marketscape report—the ability to innovate and add real business value—has only intensified in relevance. But the concept of the “fog of war” means that any battlefield tends to be complex and unpredictable. It’s often a chaotic, confusing space… with surges and set-backs… innovations and impediments… enveloping combatants like drone swarms.

In September, 2025, the AI Revolution is shrouded in that fog in real-time. “It's fair to say,” says Telecom .com’s Nick Wood, “that a degree of AI hype fatigue has begun to creep in during the second half of this year, amid growing calls for the technology—which has hoovered up hundreds of billions of capital and consumed an unholy amount of energy to develop—to actually prove its practical worth.”

Consider the experience of one contact center leader who had just completed an AI transformation initiative—

On paper, it looked like a success. The vendor promised automation, executives signed off on the budget, and IT handled the deployment. But within weeks of go-live, the cracks appeared. Customers quickly got stuck in endless loops with bots that couldn’t escalate properly. Agents were juggling more systems than before. Supervisors became overwhelmed with rework. What was pitched as innovation quickly became a daily fire drill.

In the words of The Information, “The widespread hype around artificial intelligence is now facing a significant reality check. Many businesses are grappling with the high costs of implementation and the increasingly urgent demand for a clear return on investment…. Some leaders are facing a ‘crisis of confidence,’ with data showing that corporate AI adoption is actually slowing down among the largest companies.”

High costs and unproven ROI are slowing adoption of Microsoft’s Office 365 Copilot, the company’s flagship Office software that comes with artificial intelligence features. According to a Financial Times analysis of hundreds of corporate filings and executive transcripts at S&P 500 companies last year, “the biggest US-listed companies keep talking about artificial intelligence. But other than the ‘fear of missing out,’ few appear to be able to describe how the technology is changing their businesses for the better.” It’s not yet clear consumers will be any more willing to spend on AI than businesses have been.

Or as a piece in the Harvard Business Review recently put it, troops across the AI battlefront are getting stuck in “workslop”—“Employees are using AI tools to create low-effort, passable looking work that ends up creating more work for their coworkers. On social media, which is increasingly clogged with low-quality AI-generated posts, this content is often referred to as ‘AI slop.’ In the context of work, we refer to this phenomenon as ‘workslop.’”

In effect, the diffusion of AI into enterprises and throughout the economy resembles the beginning of the First World War in August, 1914, when Germany unleashed the Schlieffen Plan, a blueprint for quick victory that would end up running into difficulties born of vast complexity. As Germany poured troops into Belgium and Luxembourg to execute a quick strike on Paris that would win the war…. its ambitions faltered in the face of reality. German, French, and British troops would soon find themselves bogged down in a new kind of trench warfare all along the western front.

Restructuring Everywhere

Amidst this fog of war, enterprises are scrambling to figure out what to do, and “everyone, everywhere, is restructuring.” Which means businesses are “reimagining workforces around skills and capabilities rather than mere head count” and “tech leadership at the top is overseeing a broader swath of responsibilities than ever before as they attempt to integrate artificial intelligence across all work functions.”

The words of Sastry Durvasula, chief operating, information and digital officer at TIAA, are apt. “We’re going to have to rewire the whole company,” he says. “Every role, what do they do? What’s the workforce of the future? I believe that 80% of the jobs will change at least 20% by AI. And 20% of the jobs will change as much as 80%.” 

Citi is piloting new agentic capabilities inside the proprietary AI platform it’s been developing over the last two years. With the new update, users will be able to direct an AI tool to complete multiple tasks, accessing multiple company systems with a single prompt. “Does it mean that we need less people?” asks Chief Technology Officer David Griffiths. “I don’t know. It certainly means that we would get a lot more done. And we’ll see how the workforce evolves with that massive boost of capacity that we’re getting here.”

Workforce adaptability will be everything. Technology and HR departments are collaborating like never before, pursuing a strategy of integrated work planning, “reworking teams not just around the capabilities and skills of the humans on staff, but also around what can be completed by AI tools.” Already today, organizations have “People come together on an ad hoc basis to work on projects and problems and move to other ad hoc groups. That’s not the same thing as a group of people reporting to a boss.”

McKinsey emphasizes that ignoring fast-moving developments like agentic AI could lead to competitive disadvantage. “To avoid this, companies need to foster cultures of flexibility, experimentation, and adaptability. They should stay constantly informed about new tools and platforms entering the market, conduct regular build-versus-buy assessments, and maintain a willingness to pause or pivot from projects when necessary…. Successful adoption and scaling of agentic AI will require rethinking and redesigning underlying business processes so that the organization can sustain new ways of working.”

As the intelligence explosions of the AI Boom come in bursts, providers in the contact center space are positioning themselves to gain advantage. All across the contact line, an assortment of diverse types of competitors—from software providers like Twilio and Genesys to business process services companies like TP and ibex—are looking to not only keep pace with technological innovation… but shape their services to the complex realities of today’s chaotic interaction battle space.

Twilio: Adapting to a Chaotic AI Battlefield

Twilio is one of the more intriguing software companies to consider because it’s undergone a massive restructuring of late, realigning its business units into a single functional organization and reducing staff to operate more efficiently. I’ve been interested in Twilio’s fortunes for some time (for example, five years ago I penned, “Twilio & Electric Imp: Testing Business Models in the Wild”).

Today, the company integrates various communication channels (messaging, voice, email, and video) directly into software applications through a flexible set of APIs, enabling developers and businesses to build and operate real-time, personalized communications with their customers. The business is organized into two segments: Twilio Communications, which houses the core communication APIs… and Twilio Segment, which offers a customer data platform (CDP) that pulls customer data from many tools into one profile so messages and offers can be more relevant.

Twilio is partnering to integrate with the likes of Microsoft, OpenAI, Snowflake, and Databricks so as to adapt to all AI models and grow as the future AI battlefield morphs and expands. In September, 2025, notes Seeking Alpha, AI adoption seems to be “accelerating growth in (Twilio’s) Voice and Messaging segments, with high-margin AI workloads boosting overall profitability and customer retention.”

So “over time,” Seeking Alpha theorizes, “AI will make Twilio’s services even more useful by enabling smarter, more personalized customer interactions…. As these high-margin AI use cases become a larger portion of the revenue mix, they will directly lift Twilio’s overall profitability.” Such is the path to the company’s stock price regaining some of its former glory.

For that to happen, the narrative will need to change. Recently, at Goldman Sachs’ annual “Communacopia + Technology Conference” at the Palace Hotel in San Francisco, software makers were under pressure to explain their progress in AI… or the lack of it. After all, as Valida Pau of The Information points out, on today’s fast-moving battle space where AI is rumored to be poised to kill software, software makers must answer the question—“how are customers using AI and are they paying for it, or will they soon?”

To which a Twilio spokesperson emphasized that, “As of last quarter, our ten largest AI startup customers that were founded in the last three years are all spending at a six-figure revenue run rate, and a few have eclipsed a seven-figure revenue run rate.” According to Matt Lucas, director of Goldman Sachs’ tech, media and telecom investment banking group, “It’s clear that corporates are still figuring out how AI is going to change their businesses and at what pace. That's creating some strategic tension within software companies as they consider opportunities to transform.”

Genesys: Humans in the Loop

Another software company I’ve been keeping an eye on is Genesys (for example, almost six years ago I penned, “Genesys: A Cloudburst of Microservices.”). It too continues to change with the times. On September 9, 2025, Genesys CEO Tony Bates opened the Genesys Xperience Summit (the first global livestream in the company’s history), with a bold assertion. “Agentic AI,” he said, “is going to be huge for our industry… from automation to orchestration for experience… the experience you deliver to your customer.”

In other words, the potential for agentic AI to impact customer support is considerable. But winning the fight for seamless experiences will not be all about automation… it’s going to be about orchestrating technology’s rhythms with the pulse of human agency. Augmenting human abilities with AI will elevate employees to scale human discernment, empowering people where they’re strongest and letting AI support them everywhere else.

For Genesys then, the future of an emerging Experience Economy is about effectively orchestrating interactions. It will be about compounding capabilities, a hybrid of human and AI, each reinforcing the next, resulting in an adaptive, emotionally intelligent system of interaction. AI agents will keep humans in the loop, working alongside humans instead of replacing them.

To realize its vision, Genesys will double-down on strategic partnerships. Because the future is also going to be about increased collaboration across a highly complex contact ecosystem. And in the case of Genesys, it’s going to be about aligning capabilities with the likes of Salesforce and ServiceNow. As such, experience orchestration across platforms from Genesys, Salesforce, and ServiceNow positions each for compounding benefits.

Think, for example, “Agent-to-Agent” orchestration between Genesys and ServiceNow, converging CCaaS/Customer Service with ITSM (ie, closing the gaps between front-office workflows through Genesys and back-office workflows through ServiceNow)… with the objective being to extend “the value of Unified Experience to help desk agents.”

Everything is changing because “AI has flipped the script. Now, there’s an opportunity for AI agents to fulfill end-to-end resolutions, grabbing data from and triggering actions within various enterprise systems. IT must work more closely with customer service to bring these resolution flows to life and pull the contact center into the broader enterprise. System convergence can aid these efforts” (as Nvidia’s Jensen Huang has noted, ServiceNow seems destined to be the best platform, the operating system of enterprise AI agents).

Listening to the Xperience Summit presentations, it became clear that today’s fog of war can also be a force multiplier when it rushes over human beings in the form of cloud computing. AI-powered real-time orchestration will aim to put human beings at the center of the action, with the cloud as the great enabler, the ambition being to transform human agents into super agents capable of high levels of productivity. As Genesys Chief Product Officer Olivier Jouve put it, “We are living through an extraordinary moment in technology… the agentic AI revolution is a seismic leap forward.”

And indeed, Genesys just announced its second-largest Genesys Cloud deal to date with a top ten global bank. It’s the second eight-figure annual contract value (ACV) agreement secured within the past 12 months. In fact, the Genesys Cloud platform has grown 35 percent year-over-year (YoY), and Genesys becoming the first tech provider to exceed $2 billion in CCaaS annual recurring revenue (ARR) earlier this year. Genesys Cloud AI exceeded $250 million in ARR last quarter, growing at twice the pace of Genesys Cloud ARR.

TP: Orchestrated Intelligence, Human Expertise

Next up, TP … for a business process services view of today’s interaction landscape. On August 27, 2025, I attended the TP.AI Industry Analyst Summit at their Innovation Experience Center (TIEC) in Silicon Valley (almost six years ago I wrote about my first visit there). It quickly became clear that the company is redefining its value proposition in an AI-driven world. In fact, TP is also leaning into the idea of agentic, AI-enabled orchestration. Global Head of AI Danny Kuivenhoven talked at length of orchestrated intelligence.

In June, TP unveiled its new “Future Forward” strategic plan to become a next-generation, AI-enabled company, including the launch of an AI partnership program with planned investments of up to €100 million in 2025. Strategic partnerships with AI companies will be key, including the likes of Parloa (agentic AI), Ema (horizontal agentic AI), and Sanas (real-time speech understanding). TP also has acquired Agents Only, an AI-enabled crowdsourcing platform.

The company also recently announced “TP.ai FAB,” a proprietary technology integration platform to safely orchestrate AI, human experts, and technology at scale. Next-generation, AI-enabled solutions and future-ready workforces will be essential in the new world emerging. That animating vision is also reflected in the fact that TP has joined forces with Carnegie Mellon University’s School of Computer Science with the aim of accelerating AI research for AI-human augmentation and innovation.

On July 31, the company released its first-half 2025 results, with group revenue of €5,116 million, up +1.5% like-for-like (LFL), which was supported by an acceleration in Core Services (H1 revenue growth of +2.9% LFL). In Q2 2025, revenue growth was up +3.5% LFL (vs. +2.3% in Q1 2025). The volatile exchange rate environment and macroeconomic situation… particularly stemming from the American political economy… has impacted TP, most clearly Specialized Services (visa application management), and its current stock price has reflected those headwinds.

But then, these are volatile times (as I discussed earlier this year in “Our Multi-nodal Hall of Mirrors”). While surprises are everywhere, TP’s operating model is wasting no time in adapting to the shifting battlefront, in part by leveraging its capabilities to seize new opportunities from new AI related value streams—AI-powered data services, data analytics, and back-office services are all growing strongly as contracts ramp-up.

Meantime, according to TP, not only are controllability and reliability the two biggest challenges when it comes to agentic AI, but the complexities of servicing the customer in industries such as healthcare, BFS, retail, and travel are often underestimated. For instance, enterprises need to understand that a lack of centralized data ownership across a particular industry ecosystem can drag on network effects and feedback loops that would otherwise accelerate AI performance. 

Intriguingly, at their Silicon Valley event, TP emphasized that the surge in AI is also creating a talent and capability gap. As such, “organizations are looking to partners to help embed AI into workflows and redesign operating models that integrate human and digital labor.” And “there’s demand for partners that can design operating models, map transformation roadmaps, and support execution.” Chief Solutions Officer for U.S. Markets Himadri Sarkar doubled-down on the point that today’s business processes demand verticalization. As such, the TP.AI orchestration platform embeds agentic AI, experts, and tools into vertical-specific solutions, working with clients to develop detailed use cases for highly complex scenario planning.

The ability of providers to work on AI use cases with the internal teams of client companies to realize business value is emerging as a particularly important competitive differentiator. As McKinsey points out, “Investing in internal use cases can have multiple benefits. For instance, risk reduction: When mistakes occur in the early application of a new technology, they’re far less costly if confined to internal use cases. Also, building muscle memory: Piloting technologies in internal workflows can help develop organizational familiarity and technical readiness that accelerates the rollout of future customer-facing innovations.”

This would explain why some, but certainly not all, of the company’s fastest growing opportunities come from existing clients. Himadri and Chief Operating Officer, Global Technology Services, Siva Pothi described everything from outcome-based simulations to customer journey/workflow mapping. It’s clear that many organizations not only need global scale across their support operations, they also require guidance-mapping and the optimization of customer journeys across digital and human touchpoints.

That could be a real opportunity for TP, as few entities understand the inner workings of business process services as they relate to the contact center on a global scale as deeply as does the Paris headquartered firm.

ibex: Tech Powered Humans

There’s another event happening soon I’ll be attending to gain additional perspectives on the business process services space…. ibex will host analysts in Scottsdale, Arizona at the end of this month with some real momentum behind it. On September 11, 2025, the company reported record numbers for the 2025 fourth quarter and fiscal year (ended June 30).

The Washington, D.C.-headquartered company delivered record fiscal year 2025 revenue of $558.3 million, up 10% from the year before. Q4 revenue jumped 18% year-over-year to $147 million, driven by strong growth in its top three markets: retail and eCommerce (up 25%), healthcare (up 19%), and travel, transportation, and logistics (up 10%). Meantime, the revenue mix in ibex’s higher-margin digital and omnichannel services also looked strong. In Q4, digital and omnichannel delivery represented 82% of its total revenue, an increase from 77% in the prior year’s quarter.

Adjusted EBITDA was a record $72 million for the fiscal year, up more than 10% from the year before. The company also had record free cash flow. Fiscal 2025 also saw ibex’s entry into India. And earlier this month, Ibex said it is now targeting the government sector.

It’s clear ibex is "transforming into a digital-first business" by leveraging AI through its Wave iX platform, which uses generative AI to improve customer experiences. As CEO Bob Dechant put it, "Importantly, this quarter marked the shift from proof of concept for our AI solutions to full scale deployments, setting the table for future growth.” The recent surge in the company’s stock price is proving his point.

For Dechant, the animating force for these numbers is “an extremely engaged employee, powered with great technology and analytics…. What we have found is that having a seamless integrated solution from AI agent to human agent uniquely positions us to support customers along the entire customer journey…. When we IPO-ed the company in August of 2020, we were early in our strategy and a work-in-progress company. We believed in ourselves and our strategy and what an amazing journey this has been.”

Key Takeaway #1: Complexities of the Human Interface

In September, 2025, the interaction battle space is endlessly fascinating in its complexities. As I suggested in The Widening Turn, we’re in the midst of a nonlinear transformation of industrial society. But while automation and robotics appear poised to enable full scale automation of considerable parts of manufacturing (particularly in China), customer support presents specific challenges to full scale automation. As I suggested over four years ago, when it comes to the service economy specifically, one might wonder about the strange possibility of “Automation’s Future Fictions.”

Because when it comes to customer service, replicating the nuances of the human interface in pursuit of a seamless customer experience continues to be a daunting challenge. Computer scientist and podcaster Lex Fridman appears to see LLMs as powerful but limited—excellent at surface-level conversational behavior but prone to errors, hallucinations, and lacking the full situated cognition of human interlocutors.

He also has explicitly acknowledged that while the verbal/linguistic side (communication) of the AI may seem advanced, something essential is still missing—“digital embodiment” and interface features, the non-linguistic/corporeal/contextual stuff that make human communication more than just strings of words (for example, podcast #368 with Eliezer Yudkowsky).

This is all to say that in September, 2025, while AI can help streamline the service experience, there are still limits. Ultimately, it’s important to remember that winning the fight for great customer support means conquering an immensely complex battlefield that is both fluid and somewhat mired in its own complexities. As the MIT Media Lab’s NANDA Initiative report, The GenAI Divide: State of AI in Business 2025, noted—“many companies are deploying AI in marketing and sales, when the tools might have a much bigger impact if used to take costs out of back-end processes and procedures.”

Key Takeaway #2: Our Nonlinear Reality

At the same time, it also strikes me that this is a nonlinear game we’re all participating in. In the fifteen years since I published “The Guns of August” Marketscape that September of 2010… business process outsourcing has spun from a world of resource-based pricing through transaction-based pricing… accelerating into the outcome-based pricing world of business process services. As I’ve long suspected and talked about, in the business process context, the term “outsourcing” may die a slow death on the field of battle.

But with every death, something new is born. As all kinds of CRM and CCaaS vendors work overtime to position their AI platforms as “agentic,” Actionary predicts that within two years, Gartner will retire the CCaaS MQ (Magic Quadrant) and launch a new Customer Service MQ that blends CRM CEC (Customer Engagement Center), CCaaS, and WEM into one. “Nirvana isn’t ‘better CCaaS,” they suggest. “It’s total customer service orchestration.”

As such, it’s my belief that the future is not zero-sum. For those thinking spherically, we appear to be moving ever deeper into a world of massive interconnection, the flywheels turning, the gains of interaction compounding, spinning us forward into an unpredictable but potentially abundant future. Or as Danny Kuivenhoven put it at the TP event in Silicon Valley, “AI makes humans better in the moment. Humans make AI better over time.”

Key Takeaway #3: Immersive Human Experiences

It strikes me in turn that if that thesis is correct, the birth pangs of agentic AI are not unrelated to the complicated task of building a metaverse, a spatial Internet, with AI being the carbon as seedbed for increasingly immersive environments wherein interactions expand exponentially, sparking increased commerce, driven by new kinds of AI-infused human agency.

And yet. Under fire in the trenches of the day-to-day battle to satisfy customers, too many participants continue to think in linear fashion. I’m reminded of the words of psychologist Rollo May… “Technology,” he asserted, “is the knack of so arranging the world that we do not experience it.” It seems May wasn’t imagining the possibilities of an emerging technology like agentic AI to help enable whole new realms of immersive human experiences… a cross-over economy between digital and physical worlds. Or as Maximus might someday tell his troops arrayed for battle in a future remake of “Gladiator”… “What we do in real life echoes in the metaverse… what we do in the metaverse echoes in real life.”

But there’s a hitch, of course. As I explained three years ago in “Metaverse Rising,” any such vision will take lots of investment and considerable time to make real. The development of new technologies has been exponential… their implementation and diffusion across the economy is another matter altogether. That fact continues to confuse countless observers of today’s interaction battlefield.

Key Takeaway #4: AI in the Contact Center... the Equivalent of Airpower in War

Thinking more deeply about today’s business battlefield, it also struck me that… to extend my martial metaphor and bring it full circle… AI in the contact center is playing a similar role as airpower does in war….

Think about it. For all the hype about AI’s impacts on customer support, deploying AI in the contact center is like unleashing airpower in war—powerful when applied skillfully in specific contexts no doubt, but… so far at least… it’s an effective complementary strategic weapon at best. When managing the customer, it’s too often true that realities on the ground are simply far too complicated to win wars with airpower alone.

Recall when the media was abuzz with gaudy headlines that Klarna was replacing contact center agents with AI’s airpower-like capabilities (our LinkedIN feeds were blitzkrieged with the news). Over time, amidst the fog of war, Klarna (and we ourselves) learned more about the realities on the ground… servicing Klarna customers was more confounding than Klarna perhaps realized. A reassessment was in order. Human beings weren’t so easily replaced within Klarna’s customer support operations. Airpower alone wasn’t going to cut it. A change in battle plans was quickly drawn up. Humans would be brought back into the loop.

It sounds familiar. In his magnum opus, The Tragedy of Great Power Politics, Professor of International Relations John Mearsheimer points out that airpower has never been enough in war. “Strategic bombing,” he writes, “is unlikely to work (in winning at war)…. modern industrial economies are not fragile structures that can easily be destroyed even by massive bombing attacks…. Land power is the most formidable kind of conventional military power available to states…. wars are won on the ground.”

Conclusion: Armor of the Titans

Maybe we’ll soon see seamless implementations of complex technologies across enterprises. Maybe too we’ll soon discover the Master Algorithm that solves the mind-bogglingly complex nature of communicating via the human interface. Maybe human genius level AGI will soon become reality. Maybe all questions will soon be answerable with a little black box. Or, put another way, maybe troops from Silicon Valley will soon find the Holy Grail… or discover the lost city of El Dorado.

For now, however, the fog of war is all too real. Paul Allen’s “complexity brakes” come to mind—the more we learn, the more we realize there is to know, and the more we have to go back and revise our earlier understandings. The potential deceleratory effects of complexity are quite real. Not only that… as the AI Boom intensifies and more intelligence explosions blow our minds… AI’s concentration among the biggest tech firms also grows… along with all the attendant risks, lying in wait across our surreal tech-littered battlefront.

In the words of Torsten Slok, chief economist at Apollo Global Management, “equity investors are dramatically overexposed to AI.” Consider yourself warned… future casualties could be significant.

Regardless of the scenario that plays out from here, it’s a good bet we’re going to continue to need ground troops. Because in September, 2025, there’s business to be done and customers to be supported. And winning the war for successful customer engagement requires ground forces in the form of human beings… sometimes in large numbers… a hybrid workforce where humans and AI collaborate to co-create real business value.

Ultimately, for any particular vendor to rise from today’s customer contact competitive landscape as a great power (whether in software or in services) it will need to possess a nuanced appreciation for the difference-making nature of human agency. Or as Ernst Jünger, author of the masterful World War One memoir Storm of Steel, put it— “Technology is just the clothes, the armour of the Titans."

Image credit: Paul von Hindenburg, Wilhelm II, and Erich Ludendorff, January, 1917 (Wikipedia)