(Antonio Marca/Shutterstock)
When DIRECTV sought to elevate its customer experience to the next level, it shifted its approach to AI adoption. Its leadership implemented a targeted AI solution for predictive routing, which delivered measurable results quickly, including a reduction in average call handle time. Building on its early success, DIRECTV expanded its AI capabilities with voice bot and chat bot intent, which decreased agent escalation rates and continued to improve the customer experience while delivering greater efficiency across the organization.
DIRECTV’s approach to AI illustrates what many business leaders are discovering. It’s critical to start with very specific opportunities where AI can deliver measurable gains, then continuously experiment and add use cases to expand its value to adjacent areas.
Increasingly, the push for AI adoption is coming from the top, with CEOs and boards approving significant investments—even before clear use cases are fully mapped out. But early ambition doesn’t always translate into lasting impact. A recent IBM study found that only 25% of AI initiatives have delivered the expected return, and even fewer are successfully scaled.
Deriving value from AI becomes even more important as we move deeper in the experience economy enabled by digital touchpoints—often without direct human involvement—at all hours of the day. Virtual agents are already “conversing” with consumers, bringing a future of always on business nearer. These shifts are fundamentally changing the way enterprises operate and the economics of AI consumption.
The Problem with Traditional Approaches
One of the immediate challenges is that traditional infrastructure wasn’t built for the scale, continuous innovation, or flexibility that AI requires. But organizations are also wrestling with how to pay for AI implementation, especially as these investments consume significant portions of their budgets.
(tete_escape/Shutterstock)
According to Genesys research, leaders say more than a third (33%) of their customer experience related budget will be spent on AI technology in the coming year. Yet despite this anticipated surge in spend, only 1% of companies believe their investments so far have reached maturity according to a McKinsey report– highlighting the gap between ambition and operational reality.
Current pricing models across the industry include license-based, subscription-based, consumption-based, freemium, and outcome-based approaches. While each has its merits, most don’t fully address that businesses need the ability to progress with AI at their own pace and on their terms. Right now, we’re seeing many businesses start small, experiment freely, scale cautiously, and try desperately to tie spend to value.
Some pricing models offer predictability but lack flexibility, locking businesses into long-term contracts or capabilities they might not fully utilize. Others provide flexibility but come with cost volatility, making financial planning difficult. Performance-based models show promise but often introduce complexity around measurement and accountability.
But pricing alone isn’t the issue. Many organizations’ AI investments are stalling because they lack a strategy that enables adoption at scale across the enterprise.
In customer experience, this means going beyond isolated pilots to fully integrating capabilities like virtual agents, copilots, predictive routing, and automated workflows that continually learn and deliver value in concert. These tools work best when they’re connected—not just deployed in silos—underpinned by a platform and economic model that support agility.
The result? Many leaders are delaying AI investments because they struggle to forecast ROI or face constraints on how quickly they can adjust and scale new use cases. To succeed, businesses need more than just the right pricing model—they need the right foundation. That includes a flexible consumption approach and an AI strategy that evolves as fast as their business and customers require.
A More Flexible Approach: Token-Based Models
Flexible AI consumption isn’t just about billing, it can be a strategic advantage for organizations. Token-based models offer organizations a way to consume AI with predictability and flexibility. Instead of paying for products or seats, organizations can allocate tokens toward specific outcomes or actions they prioritize, whether that’s using virtual agents, summarizing conversations with AI assistants, or triggering autonomous workflows.
(NeoLeo/Shutterstock)
This approach will become increasingly relevant as agentic AI-driven customer interactions become more common. AI is quickly becoming the foundation of modern customer and employee experiences—and like any foundation, it needs the right currency to support sustainable growth. Token-based models act as that currency: adaptable, value-aligned and capable of scaling usage without adding complexity. This helps to ensure that businesses can support high volumes of AI usage day and night while maintaining cost control.
Token-based models also encourage experimentation. Leaders can test different capabilities and adjust their AI usage to align with business fluctuations without navigating complex sales cycles. This can removes friction when strategies evolve.
Organizations can begin with uncommitted, pay-as-you-go tokens for maximum flexibility to try new capabilities, adapt to business changes, and implement new features as they become available. This could mean ramping up digital self-service capabilities during peak seasons or adding auto summarization to boost agent productivity. As confidence grows, they can transition to committed token bundles for greater predictability, while still maintaining the ability to reallocate tokens across different capabilities and use cases.
Questions Leaders Should Consider
As AI gains traction (and often working alongside or ahead of human teams) the right economic model can determine whether you’re merely experimenting or creating enterprise-wide impact.
Business leaders should ask:
- Is our AI consumption model designed for constant, autonomous activity, not just human-led tasks?
- Can we evolve our AI use cases without renegotiating contracts or overcommitting resources?
- Does the platform offer transparency, predictability, and adaptability in pricing?
Looking Forward
Advancements in agentic AI are expected to increasingly enable business to happen continuously, often without human intervention. Yet how organizations consume AI matters as much as what they use it for. Businesses need pricing models that allow them to start small, iterate quickly, and scale confidently.
The future of AI adoption belongs to organizations with economic models that balance innovation and ROI. In the AI-driven experience economy, success depends not just on what you deploy, but on how wisely you consume it.
About the author: Olivier Jouve is the Chief Product Officer of Genesys, where he leads the product,
artificial intelligence, and digital teams. Before stepping into this role in 2022, he served as Executive Vice President and General Manager of Genesys Cloud and Head of AI development. Prior to joining Genesys, Olivier held multiple senior executive roles at IBM, including Vice President of Offering Management for IBM Watson IoT. Earlier in his career, Olivier held executive positions at SPSS Inc. and LexiQuest; founded or co-founded Instoria, Portalys, and Voozici.com; and was the Managing Director for Webcarcenter.com. He also served as an Associate Professor in computer science at Leonardo da Vinci University in Paris.
Related Items:
AI ROI Hinges On Addressing New Challenges While Rectifying Old Ones
This Big Data Lesson Applies to AI
How To Keep AI From Making Your Employees Stupid
