AI agents have changed what consumers know about and expect from brands—often before they connect with one. This shift raises the bar for customer experience and the urgency for brands and organizations to innovate. Learn how an agent-enabled, people-powered front-office ecosystem can help.
Mesh with the new customer experience reality
AI agents are increasingly shaping what customers see, consider, buy and trust—often before they reach a brand-owned channel. They play a greater role in shepherding brand discovery, acting as search engines, filters and evaluators. In fact, Generative AI referral traffic to retail websites increased 693% year over year during the 2025 holiday season.¹ Yet 58% of business-to-consumer brands in 2026 have not fully operationalized generative engine optimization.²
To compete, brands and organizations must be discoverable and credible to AI systems while responding to live customer context. In other words: Prove more, do more and know more—while still feeling human.
An agent-enabled, people-powered front office helps organizations meet these demands through three connected layers: an experience layer where customers engage; an agentic mesh that connects external signals and internal capabilities; and a front-office operations layer that enables the enterprise to act. Together, these layers help organizations sense demand, intent and preference in real time and coordinate action across functions to deliver a seamless customer experience.
The layers of connection and value
Companies that move quickly to rethink how people and AI work together can unlock new growth and improve how work gets done. Understanding the roles and dependencies of the three layers provides a clear way to plan transformation. The sections below show how signals move and value is created across the front office—and what each layer contributes to an effective, agent-powered front-office operating model.
At the top is the experience layer, the visible face of the enterprise. This is where customers, and agents on behalf of customers, engage across channels and touchpoints—from web and mobile to chat, live interactions, wearables, smart-home devices, AI search and embedded agents. It’s where customers expect experiences to feel personalized, contextual and continuous.
Beneath the experience layer is the connective tissue that translates external signals and internal capabilities into a unified experience, integrating data, knowledge and context across agents, functions and touchpoints. Through learning and feedback, it captures outcomes, human and agent input, and quality and performance data to continuously optimize agents and improve the system. Human oversight and accountability reinforce feedback loops, helping teams refine workflows, agent behavior and cross-functional operations. Organizations with the strongest learning loops can position themselves for competitive advantage.
The third layer is front-office operations—the teams spanning marketing, sales and commerce, service and product. In an agentic front office, these teams still do the core work, but no longer as isolated silos. Instead, they work across functions to support a coordinated response across the customer journey. Human judgment, expertise and creativity remain essential as teams apply context and collaborate with AI to turn shared signals into action.
From connected layers to coordinated action
As a connected system, this model can help compound value across the enterprise by elevating customer experience, driving growth and improving efficiency. Among customers who have used AI for online shopping, 85% say it improved their experience.³ And, a recent analysis found that AI-assisted holiday shopping sessions at a global online retailer resulted in purchases at 3.5 times the rate of unassisted sessions.⁴
The goal is not to automate every interaction. It is to apply AI purposefully, adapting to a customer’s comfort, context and expectations, so that the experience becomes intent-driven and delivers meaningful results.
The stakes are now commercial as much as experiential. Revenue is increasingly won or lost in the moments when customers make decisions. Enterprise AI can help organizations recognize what customers need and respond faster—whether that means improving conversion, reducing drop-off, or tailoring an offer or service to the moment. But speed alone isn’t enough. Customers also value brands that make things easier, remember the context and follow through in ways they can trust. That combination of responsiveness, continuity and credibility is becoming a real differentiator.
The sections that follow unpack the role of each layer: the experience customers encounter, the mesh that coordinates enterprise response and the operations that make the response real.
KEY CONSIDERATIONS:
● Are AI agents already shaping who gets considered in your category?
● Can AI systems find and credibly represent your organization?
● Where are customer decisions being won or lost—and can you respond in time?
Make the experience as individual as the customer
The experience layer is where customers encounter the enterprise: across web and mobile, chat, in-person interactions, connected devices, AI search and embedded agents. A unified experience is not simply a consistent interface across channels. It recognizes the customer’s context, responds to the moment and preserves the customer’s ability to decide how they want to engage.
The appropriate balance of automation and human connection depends on customer preferences and how much autonomy they want to keep. Some customers may be comfortable letting agents search, decide and act on their behalf. Others may want AI to explain its reasoning while keeping them in control. Still others may prefer familiar human channels with AI working quietly in the background or remaining optional.
Using more than 25 AI-specific signals, Deloitte analyzed data representing about 80% of US customers and discovered tangible differences.⁵ Drawing on a deep analysis, we developed four AI customer archetypes—Frontrunners, Orchestrators, Entrusters and Traditionalists—reflecting different levels of AI fluency, trust, comfort and willingness to delegate. The journeys below illustrate how brands can turn these data-driven insights into experiences tailored to customers’ distinct needs and expectations.
AI use and AI comfort do not always move together. The four AI customer archetypes below reflect different combinations of behavior and preference, not demographics. The same person may move between them depending on the context. Together, they show how brands can create agent-enabled, people-powered experiences that preserve context, build confidence and make human interactions more meaningful.
KEY CONSIDERATIONS:
● What does each AI customer archetype need from your organization?
● How should your organization tailor experiences to different levels of AI comfort and confidence?
● How effectively does your experience and AI architecture allow customers to move between AI and human support while preserving context and control?
From fragmented silos to enterprise orchestration
Many organizations’ front offices are not yet equipped to deliver real-time, connected and hyper-personalized experiences across the full customer journey. Workflows still rely on static handoffs, fragmented data and delayed insight. Legacy systems that handle repeatable transactions well are less adept at interpreting nuanced behavior and acting on live context. As touchpoints multiply, experiences can feel disjointed from one channel to the next.
The agentic mesh turns fragmented signals into shared context and coordinated action. It preserves customer intent as work moves across teams, routes each signal to the right capability, and gives agents access to the tools, knowledge and guardrails needed to act.
In the future state, front-office operations shift from executing isolated tasks to providing standardized capabilities that an agentic mesh can coordinate across the enterprise. Front-office teams, processes and capabilities supply the agentic mesh with domain knowledge, data, skills, workflows and human expertise. The agentic mesh then translates external signals into coordinated action across functions through six core components. This system helps the enterprise act as one.
A real-world example of an agentic front office in action
Consider a flight delay. In a traditional model, the disruption can trigger rebooking, hotel changes, loyalty recovery offers, service outreach and mobile notifications. Each team only sees part of the situation. In an agent-enabled model, that same event becomes a shared signal that coordinates the next best action. The experience layer captures the customer’s needs and preferences, the agentic mesh coordinates the relevant decisions and actions, and front-office operations execute the response across channels and functions in real time.
This distinction matters. The agentic mesh creates a coordinated, governed way to interpret signals and decide what should happen next. Front-office operations provide the reusable functional capabilities, people and processes required to carry out those decisions.
KEY CONSIDERATIONS:
● Are your organization’s front-office functions built for handoffs—or reusable capabilities?
● Which customer signals should move across experience, technology and operations in real time?
● Who owns outcomes when humans and agents act—and how does the system learn?
From process lanes to ecosystem motion
Front-office operations is the functional layer supporting the mesh: the teams spanning marketing and advertising, sales and commerce, service, and product and experience. These functions remain accountable for the work that creates customer and business value. In the future state, however, they do not operate as isolated process lanes. They act as a dynamic service fabric that feeds, learns from and collaborates within the unified agentic mesh.
The operational layer supplies the domain knowledge, human expertise, workflows and services required to act on coordinated decisions. Its success depends on how well humans, data and agents are connected through shared standards, continuous learning and responsible autonomy.
Rather than relying on each function to optimize siloed interactions, the enterprise can sense signals, make decisions and respond as one. This creates a more contextual, unified customer experience while improving efficiency through real-time coordination that continuously learns and improves.
Up to 90% cost savings
Reported by an organization after rebuilding its AI orchestration⁶
Six core dynamics of the operational layer
The operational layer is where the agentic front office becomes executable. Functions shift from managing siloed process lanes to providing reusable capabilities, shared signals, governance and human expertise that the agentic mesh can coordinate across the enterprise.
Six core dynamics define this shift, enabling the front office to move from isolated workflows to ecosystem motion—and from fragmented activity to sustained customer and business value.
Modular, exposed capabilities. Functions shift their people and processes from running siloed workflows to providing clean, reusable building blocks that the agentic mesh can orchestrate safely and consistently across the enterprise.
Model and token governance. Functions operate within defined guardrails for model selection, routing and token usage, helping ensure that the right model is used for the right task, costs remain predictable and quality does not erode as usage scales.
“Always on” signal exchange. Functions continuously publish and consume relevant insights and state changes through the mesh, creating shared customer and operational understanding that supports faster signal-to-action response.
Continuous improvement. The mesh provides enterprise agent telemetry and performance data, while humans retain oversight and accountability. Functions close feedback loops to improve workflows, agent behavior and cross-functional operations.
Human-agent operating model. Organizations redesign roles around the division of labor between humans and agents, with clear accountability and standards for how they collaborate and what they deliver.
Learning and upskilling. Organizations deliberately develop how people learn to work alongside AI, investing in upskilling, training and certification paths that help their people succeed.
Why the front office must be rewired
Together, these dynamics enable the enterprise to sense signals, make decisions and respond as one. They translate front-office operations from a set of siloed functions into a coordinated system of reusable capabilities, shared context, accountable human–agent collaboration and continuous improvement.
The urgency is real—and so is the opportunity. Deloitte’s 2026 State of AI in the Enterprise research shows that 37% of organizations report little or no change to existing processes, while another 30% are redesigning key processes around AI but keeping dated business models.⁷ At the workforce level, 84% have not yet redesigned jobs around AI capabilities.⁸
In short, many companies are experimenting, but few are rewiring how work gets done.
That gap limits the value they can capture. When processes remain organized around siloed handoffs and roles remain unchanged, agents are layered onto existing work rather than integrated into a coordinated system of people, data and capabilities. The result is isolated productivity gains, duplicated effort, unclear accountability and inconsistent customer experiences—rather than compounding improvements in growth, efficiency and responsiveness.
Closing the gap requires moving from process lanes to ecosystem motion: organizing work around shared outcomes, enabling capabilities and signals to move across functions and establishing clear human–agent roles, guardrails and learning mechanisms so the front office can improve over time.
KEY CONSIDERATIONS:
● What must each of your organization’s functions expose as a reusable capability?
● Where do guardrails, shared signals and feedback loops enable responsible autonomy?
● How must roles, decision rights and skills evolve as humans and agents share execution?
Designed around trust. Delivering on intent.
Customers are moving fast. And AI is raising their expectations. The challenge is no longer personalization at scale—it’s making every interaction feel genuinely useful, timely and tailored to the moment. The archetype scenarios in this article illustrate what that can look like: agentic experiences designed around different customer needs, levels of complexity and moments that matter. There is no single starting point. Organizations should begin where their understanding, constraints and opportunity to build momentum intersect. The opportunity is to make customer experiences easier, smarter—and more human—while creating a more dynamic and connected front office.
Done well, the agentic front office brings all three layers together: experience strategy and customer understanding, technology and data readiness, and reimagined processes and workflows. Agents can help coordinate capabilities across a de-siloed front office, while people focus on judgment, empathy, creativity and accountability.
The result is a more unified experience—one that connects customer insight, integrated foundations and coordinated operations to create lasting value.
KEY CONSIDERATIONS:
● Where should your organization start to unlock the greatest customer and enterprise value?
● Where are the most consequential gaps—in experience, technology and data, or operations?
● Where must leadership intervene to scale an enterprise-level, agent-enabled, people-powered front office?
About the authors
Briana Mangialardi, senior manager of design and experience at Deloitte Digital, helps clients define and create human-centered experiences, guiding teams towards human-led, AI-driven delivery that balances efficiency, efficacy and creativity.
Christine Kang, principal at Deloitte Digital, helps complex organizations drive growth and innovation, translating customer data and insights into digital products and experiences that deliver business value.
Julius Tapper, senior manager of design and innovation at Deloitte Digital and co-lead of Deloitte Digital's UX research community, helps organizations connect AI, business strategy and human experience to shape their growth choices.
Chris Sommerfeld, principal at Deloitte Digital, helps large organizations unlock the value of a human-centered, design-led, AI-powered approach to designing and delivering better experiences, services and products.
SOURCES
1. Bobby Stevens and Simon Chafetz, Emerging Retail and Consumer Trends, Deloitte, Q1 2026, page 5.
2. Deloitte Digital, Your customers have already moved on. Has your personalization strategy?, September 29, 2026. (Note: Cited data was not included in the published research report but reflects results from the underlying survey. For full survey methodology, see the published report.)
3. John Valenti and Simon Chafetz, Commerce strategy: Agentic shopping, Deloitte, 2026, page 3.
4. John Valenti and Simon Chafetz, Commerce strategy: Agentic shopping, Deloitte, 2026, page 4.
5. Converge™ by Deloitte consumer data collected February–April 2026. Analysis drew on more than 750,000 predictive and descriptive data points, including transactional, behavioral, lifestyle, financial, demographic and other signals, from more than 235 million US consumers.
6. Tim Smith, Amanda Nelson, Nicholas Merizzi, Diane Ma and Diana Kearns-Manolatos, AI tokenomics: A CFO’s guide to governing the AI P&L, Deloitte, August 14, 2026.
7. Jim Rowan, Beena Ammanath, Nitin Mittal and Costi Perricos, State of AI in the Enterprise: The untapped edge, January 2026, page 11.
8. Jim Rowan, Beena Ammanath, Nitin Mittal and Costi Perricos, State of AI in the Enterprise: The untapped edge, January 2026, page 13.