We’re standing at the start of one of the most profound transformations in business history, driven by a new generation of generative and agentic artificial intelligence (AI). These technologies are reshaping how organizations deliver customer and employee experiences by unlocking new levels of automation, augmentation, personalization and optimization.   

AI-Powered Experience Orchestration, once a vision of using AI to coordinate experiences across systems and channels, is now becoming reality. In this article, we explore how this transformation is unfolding, including its future potential, and define the six Levels of Experience Orchestration. This maturity model provides a foundation for organizations to assess where they are today, envision what’s possible and build a strategy for AI-powered growth.1 

Introduction 

The purpose of experience orchestration is to achieve two objectives at the same time: 

  1. Reduce the cost of operations. 
  2. Increase customer loyalty for long-term growth.

Organizations can better balance the tradeoff between operational efficiency and delivering people-centric experiences. The right AI-Powered Experience Orchestration strategy enables both.  

By coordinating data, systems, channels and roles, orchestration creates experiences that are more effective, efficient and emotionally intelligent from the perspective of customers and employees. As new innovations emerge, we’re heading toward universal orchestration — transcending customer-facing activities across the front- and back-office — enabling organizations to reimagine the contact center, customer and employee experiences, and their business overall. 

The following Levels of Experience Orchestration define the maturity curve from fully manual to fully autonomous orchestration. Each level marks a meaningful leap forward in how automation, augmentation, personalization and optimization are applied and unlocks potential new business value in the form of increased efficiency, deeper customer loyalty and stronger employee engagement. 

Level 0 – Zero Orchestration 

Customer interactions are entirely manual, handled through basic telephony systems with no integrated tools or intelligence. Human agents rely on training and static documentation. Every interaction is reactive and inconsistent.  

There is no unified view of the customer, and no orchestration of tasks or insights across systems. Customer service is treated as an operational necessity rather than a strategic function — leading to high effort, high attrition and poor outcomes

  • Automation: None. All tasks — including routine inquiries — require full human agent involvement. 
  • Augmentation: Human agents work without system support. No contextual surfacing of data or task-specific assistance. 
  • Personalization: No system-supported personalization based on customer profile or history. 
  • Optimization: Manual training and static scheduling dominate. No real-time insights, quality management, or workforce planning automation. 

Level 1 – Menu-Based Navigation 

Interactive voice response (IVR) systems provide basic automation with fixed routing logic and limited speech recognition. Customers interact through keypad or voice menus, typically to check status or route to a department. While this reduces call volumes slightly, experiences remain impersonal and voice-bound.  

Human agents are still required for most tasks and rely on limited CRM context. Quality control is manual and retrospective. The system operates, but it doesn’t adapt

  • Automation: IVR handles simple information requests, like checking an account balance or order status, based on keypad input or keyword recognition. Logic is fixed and non-adaptive. 
  • Augmentation: Human agents can view static customer records during interactions but must manually search for relevant insights. 
  • Personalization: Skill-based routing and language preferences are possible, but experiences remain largely uniform. 
  • Optimization: Quality assurance is based on sampled recordings and human agent scheduling is time-consuming and reactive. 

Level 2 – Pre-defined Dialog Automation 

Conversational AI combines automated speech recognition (ASR), natural language processing (NLP) and natural language understanding (NLU) to engage across multiple communication channels. Interactions are governed by pre-defined rules and scripted dialogues. Predictive AI models are applied to specific use cases (like routing or engagement) but have not yet been generalized to determine next-best actions as part of an experience in general. 

  • Automation: Conversational AI enables bots that can automate routine dialogs with customers in digital and voice channels (omnichannel), like order tracking, password resets or identity verification. Bots are rigid and follow predefined flows that are structured around scripted logic and fixed decision trees. 
  • Augmentation: Human agents begin receiving contextual assistance via knowledge surfacing tools and are proposed next steps based on CRM context or keyword triggers. 
  • Personalization: The customer experience remains standardized and lacks adaptability or personalization beyond static inputs. Foundational workforce engagement management capabilities are introduced and help align tasks with employee skills and availability. 
  • Optimization: Experiences are optimized using specialized predictive AI models for routing, engagement, and forecasting. Speech and text analytics power quality assurance processes. 

Level 3 – System-Generated Conversations 

Generative AI uses Large Language Models (LLMs) and transformer-based architectures to produce content within the boundaries of its configuration. AI performs tasks it has been explicitly designed or trained to do no more, no less. It enhances experiences through automation, augmentation, personalization and optimization, while still operating within predefined logic and workflows.  

This level of AI does not reason or make decisions beyond what it has been instructed to do; it simply executes its programming with increasing breadth and fluency.  

  • Automation: AI-driven virtual agents automate broader and more complex interactions, like troubleshooting, order status or product inquiries. These virtual agents appear autonomous but operate strictly within configured workflows and rules. They don’t reason or infer beyond defined patterns. Capabilities like intent recognition or FAQ handling enable them to manage more nuanced scenarios, but only to the extent their training and configuration permit. Virtual supervisor features help automate operational monitoring, alerting stakeholders based on pre-set thresholds or behavioral signals. 
  • Augmentation: Agent copilots enhance human performance by surfacing relevant insights, like suggesting the next best action, generating summaries or retrieving knowledge articles, but within the constraints of pre-defined rules and models. Agent copilots respond to recognized cues or signals in the conversation but don’t adapt or plan beyond those parameters. Supervisor and admin copilots provide guidance and recommendations based on configured criteria, helping scale knowledge without overstepping decision boundaries. 
  • Personalization: Generative AI can tailor responses using structured segmentation, intent classification and business-defined attributes. It personalizes based on what it’s told drawing from CRM data, known preferences or prior interactions to generate output that aligns with specific business goals or segments. While the content feels custom, it’s generated within the guardrails of predefined logic and configured behavior. At this level, personalization is powerful but still bound by what has been structured. 
  • Optimization: Journey orchestration and experience management are improved by AI’s ability to execute pre-planned optimization strategies. Forecasting, scheduling and workload balancing benefit from predictive models that continually refine recommendations based on historical data. Even here, though, the AI is not adaptive. It executes trained behaviors and is periodically retrained by humans to maintain relevance. Orchestration of tasks, alerts and workflows across the front and back office remains reactive to defined conditions, not proactive reasoning. 

Level 4 – Agentic Experience Generation 

AI evolves from simple execution to intelligent problem-solving. Systems are configured for specific objectives and use reasoning, planning and memory to determine how to best accomplish goals while still operating within clearly defined boundaries.  

This level introduces agentic AI that interprets context, plans across steps, and adjusts actions based on dynamic inputs. However, all execution remains semi-autonomous. Human input, approval and oversight are still integral, enabling alignment with intent and preventing overreach. 

  • Automation: Virtual agents, supervisors and admins now carry out complex transactional tasks and decision sequences across more demanding domains like sales, renewals and retention. They determine optimal steps within a configured objective, guided by defined guardrails and approval requirements. These systems can reference and follow structured content such as standard operating procedures (SOPs), knowledge articles or instruction documents to execute tasks accurately and consistently. Asynchronous execution becomes more common, allowing tasks to progress in the background while customers or employees attend to other work. The virtual agent will notify the user when tasks are complete or need input, maintaining transparency and human control. 
  • Augmentation: Copilots are increasingly proactive, surfacing intelligent suggestions to agents, supervisors and administrators and offering to execute them once approved. This includes updating records, identifying risks, streamlining processes and translating communication in real time. These copilots also provide real-time signals to help coach human users; highlighting missed steps; suggesting compliance tips; or reminding them of key context in a supportive, non-intrusive manner. Rather than taking control, they help people perform better through subtle, contextual nudges. While they analyze complex inputs and adapt their suggestions, they never act autonomously, preserving human decision authority. 
  • Personalization: Personalization becomes more strategic and data-driven. AI systems use internal memory, customer profiles, prior interactions and contextual cues to determine which responses or workflows best align with the customer’s profile. This includes drawing from business-defined segments, transactional history and configured rules. Human agents receive support that adapts to the complexity of the scenario, with suggestions that reflect personalized playbooks rather than generalized workflows. However, all personalization continues to operate within the boundaries of business configurations, without improvising outside defined limits. 
  • Optimization: Orchestration now leverages dynamic context to improve experience flows across systems. AI components work semi-autonomously to identify better paths and more efficient resolutions, requiring less manual setup but operating within predefined constraints. Capabilities like anomaly detection, pattern recognition and memory-based decision-making help identify process gaps or escalations. In cases that require discretion or policy interpretation, like mortgage approvals or financial adjustments, AI supports your workforce by preparing the decision context — but the final action remains with a human. 

Level 5 – Universal Agentic Orchestration 

AI reaches a state of goal-driven autonomy, capable of independently planning, deciding and executing based on objectives defined by human stakeholders. Virtual agents, supervisors and administrators are no longer constrained by fixed workflows or linear task execution. They dynamically generate new strategies and adaptively coordinate actions in pursuit of business outcomes, guided by overarching goals rather than rigid instruction sets.   

This is the apex of orchestration maturity — where AI transitions from reactive automation to self-directed, collaborative experience management. 

AI systems combine LLMs with memory, planning and reasoning, enhanced by continuous feedback loops. Experiences are no longer siloed or transactional, but instead become fluid, adaptive and intelligent across entire ecosystems.  

 

AI entities interact directly with one another, sharing goals, exchanging context and delegating responsibilities, enabling distributed orchestration across both internal systems and external partners. Human involvement becomes strategic and intentional, focused on oversight, governance and complex decisions that benefit from empathy, creativity or judgment. 

  • Automation: Virtual agents, virtual supervisors and virtual admins autonomously initiate, execute and complete tasks end-to-end. Systems interpret organizational goals and contextual data to determine the optimal path forward without relying on predefined scripts or manual intervention. Task ownership and handoff occur dynamically between intelligent agents, with decisions made cooperatively across roles and domains. As systems pursue shared goals, they align decisions across departments, channels and even partner networks, executing actions at scale and in harmony. Most operational needs, whether customer-facing or back-office, are resolved automatically through intelligent, multi-actor collaboration. 
  • Augmentation: While AI handles most tasks independently, humans remain essential for oversight, policy and strategic judgment. Copilots proactively surface results, summarize actions taken, and present them for audit or intervention. In other scenarios, copilots anticipate needs and offer to complete tasks, learning from approval patterns and expanding their scope of support. Importantly, copilots and autonomous agents work together, passing insights and intermediate decisions fluidly between each other to assist human stakeholders in order to maximize efficiency. Employees benefit from orchestrated intelligence that adapts to their roles, context and workflow, elevating human contribution to high-impact, decision-oriented work. 
  • Personalization: Experiences are orchestrated by virtual admins, supervisors and agents, each contributing unique perspectives, context and functions. These intelligent systems draw on prior interactions, enterprise knowledge and evolving behavioral signals to tailor experiences in real time. The personalization is dynamic and distributed, not just driven by one system — but refined collectively across AI-powered actors that coordinate their understanding of the customer’s goals, preferences and journey state. Whether inside a single brand or across ecosystems, virtual agents synchronize their responses and decisions to help deliver continuity, relevance and intent alignment at each touchpoint. 
  • Optimization: Optimization becomes autonomous, distributed and goal-focused. Each AI-driven system contributes to performance improvement not in isolation, but as part of a continuous, collaborative learning network. They refine workflows and decision models based on shared feedback loops, performance data and goal achievement metrics. Orchestration logic adapts fluidly to changing organizational priorities, and AI-driven agents work together to reallocate effort, rebalance strategies and improve outcomes at scale. This creates a self-sustaining intelligence layer where orchestration evolves with the environment without relying on manual configuration or retraining. 

The Levels of Experience Orchestration maturity model illustrates how businesses can evolve from fully manual operations to intelligent, AI-driven systems capable of independently managing and improving customer and employee experiences. Each level reflects a significant leap in AI capability and the potential value it can deliver — starting with isolated task automation and culminating in systems that can plan, reason and act in pursuit of business goals.  

As organizations evolve their experience orchestration maturity, they will often operate across multiple levels at once, depending on business priorities, customer segments, operational constraints and risk considerations. Some experiences will remain highly structured and human-supervised, while others will benefit from increasing autonomy and self-direction. 

Critically, the path to maturity also involves the growing collaboration between intelligent systems — AI-driven agents coordinating with one another to resolve complex tasks, share context and dynamically adapt across journeys. These interconnected agents, whether supporting customers, supervisors, or administrators, form the foundation for scalable, adaptive orchestration. In this model, intelligence is no longer isolated to a single system or interaction. It becomes a distributed capability, capable of continuously learning, sharing, and improving across the entire experience landscape. 

Conclusion 

The Levels of Experience Orchestration provide a structured maturity model to help organizations navigate their transformation journey. We believe most enterprises operate today at Levels 1 or 2. But the path forward is clear — and accelerating.  

Organizations that invest in agentic orchestration will be well-positioned to unlock potentially exponential value: greater automation and scale, more empowered employees and deeper customer loyalty. 

Where does your organization operate today? What would it take to move up a level?  

Genesys is here to help you define that path—and take the next step toward orchestrated, emotionally intelligent experiences at scale. 

 

1 This is a discussion paper, not a product roadmap. Genesys does not commit to delivering any capabilities described in this document. 

* This article was originally published on May 14, 2024 and has been updated.  

Authors:

Tony Bates is the Chairman and Chief Executive Officer of Genesys. He leads the company’s strategy, direction and operations in more than 100 countries and oversees a global team of more than 6,000 employees.

Tony has decades of experience steering business-to-business and business-to-consumer companies through major market transitions and rapid scaling. A passionate technologist at heart, Tony began his career in network operations and internet infrastructure, teaching himself to code during his daily train commute. He swiftly gained the business acumen to advance into trusted executive roles at some of the world’s most respected global SaaS companies.

Career highlights include leading Cisco’s Service Provider business, growing its Enterprise and Commercial division to more than $20 billion in annual revenue and serving as CEO of Skype, where he was responsible for expanding the business to over 170 million connected users. Once Skype was acquired by Microsoft, Tony became president where he was responsible for unified communications before serving as executive vice president of business development and developers. In addition to his role at Genesys, Tony serves on the board of directors at VMWare.


Dr. Peter Graf is the SVP of Strategy at Genesys. In his role, he is responsible for developing, communicating, sustaining the Genesys strategy.

Prior to joining Genesys in 2017, Peter held a variety of executive leadership positions in strategy, development, and marketing throughout his more than 25 years in the global enterprise software industry, most notably as an Executive Vice President at multinational software corporation SAP. Peter earned a doctorate in artificial intelligence from Saarland University and a master’s degree in computer science and economics from Technical University of Kaiserslautern in Germany.