Scale it

You’ve proven AI can deliver results. Now comes the real opportunity: scaling its impact.

At scale, AI Agents reshape how your business delivers value. But they require an evolution of your customer experience, the role of your team, and the economic model of your support function. If you don’t evolve the system around AI, you’ll hit a ceiling.

So the questions you need to ask now are:

That’s what this chapter covers. How to scale beyond quick wins and unlock the compounding value across three transformational pillars:

The outcome is a new model for support: AI as infrastructure, humans as system designers, and a transformed customer experience – where every interaction starts with an AI Agent, human experts collaborate when needed, and customers get real value.

If you were building support from scratch today, you’d design around AI from day one. That’s the mindset you need to adopt.

But many teams treat AI like a feature instead of infrastructure. They tack it onto existing processes and tools, limit its scope to tier-one issues, and fail to evolve the systems around it.

These teams are thinking too small. They chase incremental efficiency gains, underinvest in the system change needed to make AI successful, and get stuck as a result. The customer experience remains fragmented. The support team stays reactive. The business leaves value on the table.

AI Agents don’t just answer questions, they act. They pull in real-time data, trigger backend processes, and complete multi-step tasks. They’re not limited to “simple stuff,” and they’re not just a triage layer. They’re fully capable, end-to-end resolution engines, and are central to transforming your customer experience.

Before we dive into the mechanics, it’s important to look at the end state. What does it look like when AI is deeply embedded, not just deployed?

This chapter will show you how to avoid stalling, and scale with confidence.

It’s easy to think of AI Agents as a support team initiative, but the initiative won’t scale without the C-suite.

AI Agents reshape how support works, how teams are structured, how performance is measured, and how cost and value flow through the organization. That kind of change touches everything, and it can’t succeed without top-down commitment: budget, air cover, and a willingness to rethink the old cost models. That means aligning your CFO around ROI, your CCO around journey design, and your CEO around customer experience as a strategic advantage.

Even a high-performing AI Agent will stall without this. You’ll see early wins, but the bigger opportunity – the one that shifts your economics, elevates your team, and transforms your customer experience – won’t materialize without someone at the top who understands that AI is infrastructure.

If you’re a support, CX, or AI transformation leader: use early wins to show your exec team what’s possible.

If you’re a C-suite leader: your support team can’t do this alone. But with the right mandate, they’ll help lead your company through the next era of customer experience.

For many companies, transforming the customer experience is AI’s biggest promise.

Efficiency might justify the investment, but the most ambitious support teams don’t just deploy AI to drive efficiency or automate Tier 1 volume. They use it to create improved and entirely new customer experiences; ones that are faster and smarter, that remove friction, anticipate needs, and deliver value in every conversation.

These teams aren’t layering AI on top of the old model. They’re designing around it, building new models with AI at the core.

This section outlines how to make that transition from reactive service to AI-powered experience design. It’s built around five principles that help teams scale trust, consistency, and delight across the customer journey.

What do customers actually want?

Those needs haven’t changed. What has changed is your ability to meet them – more consistently, more intelligently, and at greater scale.

In the traditional support model, if you had no constraints (infinite time, people, and budget), you might have delivered on those principles by:

But you didn’t have infinite resources, so you made trade‑offs. You might have buried the “Contact Us” button behind help articles, deflected when you could, and triaged the rest. You relied on queues, rules, and SLAs to manage complexity and control cost.

But those constraints are no longer fixed.

Modern AI Agents don’t just answer questions immediately. They understand context. They complete tasks. They get better over time. And when you design your systems with that in mind, something powerful happens: support stops being reactive.

So you can stop working around limitations, and start designing around what customers actually want.

How you design your customer experience shifts:

Each of these principles is engineered to deliver on the first principles of great support: fast, accurate resolutions, empathy, and effortlessness.

Treating support as a product means designing, building, and managing your support experience with the same rigor and accountability you would apply to your core product.

It’s a shift in mindset from seeing support as a reactive function or operational cost to treating it as a customer-facing experience that shapes perception of your brand, drives loyalty, contributes to growth, and accelerates the value customers get from your product.

Just like product teams:

AI isn’t replacing the human touch. It’s redefining when, where, and how it’s most valuable.

In a scaled model, AI becomes the first responder: the default entry point for every conversation (and the end point for most of those conversations too). But it doesn’t work in isolation. The experience should be hybrid by design:

Every message, whether from AI or a human, should feel like part of one coherent experience. Behind the scenes, you define who handles what, and why.

Ask:

It’s also important that you don’t retrofit AI onto old processes or team structures.

A big mistake teams make when adding AI to the customer experience is tacking it onto something that wasn’t designed for it. That’s when experiences break. Handoffs feel clunky. Customers lose trust.

You need to design around AI from the beginning. For example, write knowledge content for AI, not humans. Design workflows where AI is the primary resolver. Define what happens after success: what feedback loops get triggered, what metrics matter, and who owns what.

When AI is at the core of the system, not tacked onto it, it performs better. And the customer experience feels better.

Use AI to anticipate customer needs and offer help, guidance, or nudges before they become problems.

It’s not about deflection, it’s about momentum: identifying drop-offs, surfacing friction, and stepping in at just the right time.

At Intercom, our team is working towards Fin acting as a “digital customer experience agent” for every customer; not only answering support queries but also offering tailored onboarding and proactive help. The aim is for AI to feel integrated into the full lifecycle experience, delivering relevant support without even needing to be asked.

Some customers still assume AI won’t help them. You’re dealing with the legacy of bad chatbots that gave vague answers, clunky menus, and left people in endless loops.

You build trust in AI by showing that it works. At scale, every interaction becomes a test. And every successful resolution is proof.

To build trust at scale:

At Intercom, we surveyed over 1,000 US consumers to see if their sentiment toward AI-first support changed once they saw what an AI Agent was capable of. The change was significant: by simply showing people that modern AI Agents can conversationally engage with customers, perform actions, and seamlessly escalate queries to a human, positive sentiment increased by 20 percentage points and trust levels jumped by 18 points.

Your AI Agent represents your brand. The way it speaks, follows policies, and responds matters. With AI Agents, you can use capabilities like tone control, fallback logic, and language preferences to align the experience to your brand’s standards.

For Fin, we call this “Guidance.” It’s a way to control how the AI Agent sounds and behaves using simple, natural language prompts.

Consistency builds trust, but personality builds connection and loyalty.

In an AI-native customer experience:

To realize the true value of AI and the promise of a transformed customer experience, the way your team works needs to change.

That doesn’t mean starting from scratch. But it does mean rethinking how support is structured, how performance is owned, and how your systems evolve.

The teams that scale AI successfully treat it like infrastructure and design around it. It’s an intentional shift that helps the technology keep performing as its scope grows.

In this section, we’ll show you how to make a strong case for change, evolve your team structure, and create systems and ownership models that sustain AI performance over time.

Scaling AI is as much about people as it is about technology. If people don’t understand what’s changing, doubt creeps in. They worry about their roles. They question whether the new system will work. They aren’t open to trying new things and embracing a “new way.” This lack of alignment puts the AI initiative at risk of stalling out, so it’s important to address before it takes hold.

Here’s how to approach this.

Scaling AI successfully requires leadership backing to move at speed. Executive sponsorship provides the air cover to unlock resources, align cross-functional teams, and make AI a visible priority across the company.

You can have the right team and tooling in place, but without strong leadership support, momentum is harder to sustain. Scaling AI reshapes how people work, roles evolve, and systems are structured. That kind of shift works best when leaders are behind it.

To secure exec air cover, make the proof tangible by focusing on one or two clear, measurable wins that show what’s possible. For example, choose a high-volume use case where impact is easy to track. Let the AI Agent handle it, observe what changes, and then share the story broadly, consistently, and in places where people will notice.

That’s what Dotdigital did:

Executive backing will give you momentum, but you need to build trust on the ground to sustain it. People need to see how AI will improve their work and how it fits into the bigger picture. That starts with transparency: what’s changing, why, and what success looks like.

Then, give them a role in shaping the system. They need to know exactly what the AI Agent is responsible for, when it hands off, and who owns its performance. That kind of clarity is what enables scale. It makes the change feel grounded and safe, and gives the organization something to align around.

Ask for feedback on AI performance. Show how their insights improve it. Turn the AI Agent into something the team is involved in building, not something that’s happening to them.

When AI Agents resolve the majority of customer conversations, the shape of your support team has to change.

The old tiered model built around queue management, handoffs, and volume-based productivity no longer fits. AI now handles the bulk of customer interactions, including many that used to require human involvement.

That shifts the role of your human team entirely. Responsibilities evolve, and success is measured differently.

It goes beyond just adding automation to existing ways of working. You’re building an operating model that’s entirely new.

Most teams don’t hire a dedicated AI function from day one. They start by distributing a few critical responsibilities across existing team members, and formalize those responsibilities as AI becomes central to how support works.

Once you have executive support and a clear strategy in place, these are the four foundational roles we believe are key to getting AI off the ground in a meaningful way:

You’ll also need a new kind of leadership needed to make this model work. The traditional support leader doesn’t map cleanly to an AI-first organization. You need a new layer: leaders who are part strategist, part operator. They roll up their sleeves to analyze the AI Agent’s performance, retrain content, and debug handoffs, but they also coach the team through a new way of working.

This is the “player-coach model” – leaders who actively shape both the system and the people inside it.

These leaders see the AI Agent as a teammate to manage, not just a tool to monitor. They can’t be purely people leaders or purely systems thinkers. They need to be both, and they’re emerging as a critical hire in support right now.

Some teams are restructuring their organizations around the AI Agent as a core product, not just a support tool.

At Intercom, our Support team is now structured around three pillars: Human Support, AI Support, and Support Operations and Optimization.

Each function includes evolving roles and responsibilities, but all of them work together as a system, with clear ownership and shared accountability for AI performance.

As your ways of working evolve, so will the traits you value in your team.

Traditional strengths like good communication skills, empathy, and the ability to work under pressure are still important, but they’re not enough.

In an AI-first model, you’ll want to work with people who demonstrate:

Hire people who notice what’s broken and suggest how to fix it. Who bring customer insights back to the product team. Who are quick to automate repetitive work. Who are curious, proactive, and genuinely excited about what AI makes possible.

And that mindset doesn’t just apply to agents. It extends to the managers leading them. You need the whole team to care about how the system performs and treat the AI Agent like a product: something to tune, improve, and evolve over time.

You need people who wake up thinking about how to make the whole thing better.

Once the right people are in place, you need an operating model that makes progress part of how things work and keeps the AI Agent improving over time.

At Intercom, we use a simple mantra to guide how we think about this: “The first time you answer a question should be the last.”

We’re not chasing perfection. We’re trying to build an operating model where every resolution improves the system, so that fewer issues repeat, quality compounds, and support becomes more scalable over time.

Getting this right takes intentional design. It takes clear ownership. Guardrails that let you move quickly without risk. A way to feed insights back in. And a culture that celebrates the work, not just the outcomes.

Let’s break that down.

One of the most common reasons AI performance plateaus is ambiguity.

When no one owns how the AI Agent performs, feedback gets lost, issues linger, and improvements stall. High-performing teams assign a single owner who’s responsible for making the AI Agent better by:

As the AI Agent handles more volume and complexity, change might start to feel risky. And when teams hesitate to make changes, performance stalls.

That’s where lightweight governance comes in: a clear way to keep iterating without bureaucracy or endless approvals.

The teams that have developed a good rhythm with this put a few principles in place:

AI performance isn’t static, but most teams treat it like a one-time implementation. The most successful organizations design systems that learn: they analyze where the AI Agent struggles, then feed that insight directly into structured improvement.

That might look like:

Even the best system won’t keep improving if people stop believing in it.

Belief will fade quietly if you don’t reinforce it. Keep it strong by:

When done right, AI goes far beyond improving support efficiency. It rewires the financial model: breaking the link between support costs and revenue growth, and turning support into a contributor to customer activation, retention, and lifetime value.

Stop measuring AI like software and thinking about ROI in isolation. Start treating your AI Agent as a new workforce capability that changes how your support function creates and captures value.

Traditional support economics are built around linear growth: more tickets means more headcount, more outsourcing, more software. Value was measured by containment: deflections, handle times, tickets closed.

This is what it means to decouple cost from growth (and what every CFO wants to see).

But that curve doesn’t bend on day one. The early stages often look like additive spend because you’re funding a new workforce layer while your old one is still intact. The real gains come gradually through attrition (not backfilling open roles), pausing BPO expansion, slowing down hiring plans, and the compounding effect of automation.

(See Chapter 2: Launch it for more detail on how automation rate influences time-to-ROI.)

Metrics like average handle time (AHT), cases handled, and first contact resolution (FCR) defined what “good” performance looked like when human agents managed the majority of support queries and success meant moving quickly through high volumes. But they no longer reflect the value of your team’s work.

In many cases, they create the wrong incentives: pushing speed over quality, or signaling failure when agents are actually solving more difficult problems. Here’s why they don’t work:

You don’t need to throw out these traditional metrics entirely, but they become less relevant once you’ve deployed an AI Agent. Instead, you need to focus on:

Instead of asking, “How many cases did we close?”

Ask:

In an AI-first model, your AI Agent is your front line. It needs to be evaluated with the same rigor, ownership, and outcome-focus you apply to human teams.

These are the core metrics that matter when measuring AI Agent performance at scale:

These metrics work together. Resolution tells you if it’s working. Involvement shows where it’s working. Automation reveals how much work it’s doing.

Now, you need to measure how customers feel about it.

CSAT only covers a small fraction of conversations (often less than 10%) and tends to capture extremes. Deflection rate assumes that if a query didn’t reach a human, it must have been resolved. But that’s a risky assumption. These metrics miss context and nuance. They make it hard for teams to know what to fix next, or whether anything’s actually improving.

At its core, assessing the customer experience comes down to two things:

These questions aren’t new, but until recently, support teams didn’t have reliable ways to answer them.

Here’s how to shift your measurement model to get a true read on the customer experience.

Traditionally, deflection rate focused purely on whether or not queries reached a human agent. It was used as a proxy for success, assuming that if a customer didn’t reach a human, their issue was resolved. But that’s not always true.

In the AI era, deflection alone doesn’t tell you enough. While helpful as an early signal, it’s a limited, and potentially misleading, measure of success. What matters is whether the issue was resolved, regardless of who handled it.

With AI now capable of resolving the majority of queries, we need to shift our focus from deflection to resolution.

Resolution rate tells you what’s working. But what’s unresolved tells you where to improve. Look at what AI isn’t resolving, and why. Is the content unclear? Is it a configuration issue? Use this to continuously improve the system.

Fin surfaces recommended fixes right alongside these breakdowns – editing content, refining a prompt, or adding a missing step. This creates a feedback loop built directly into the system:

Deflection tells you what didn’t reach your team. Resolution tells you what’s actually working, and what to fix next.

If you want to scale AI successfully, you need to understand what AI is (and isn’t) resolving, and how the customer feels about it.

In the next section, we’ll unpack how to measure customer experience at scale.

In the early phases of AI adoption, success is measured through hours saved, headcount avoided, CSAT maintained. But at scale, leadership needs to understand the broader value story: how support performance connects to retention, conversion, and revenue growth.

This means modeling not just cost reduction, but business impact over time:

The most advanced teams go a step further: they directly attribute business value to support data. They track product usage lift after successful resolutions. They measure feature adoption linked to consultative support.

In this model, support becomes a real lever for growth. Not a cost center. A value driver.

In a traditional model, support spend is siloed. You fund people, tools, and services separately, and scale them linearly with support volume. When volume goes up, you hire. You expand BPO contracts. You add software to help the team move faster.

AI Agents break that model. They decouple headcount from volume. And that shift should be reflected in how you model cost and capture value.

AI Agents introduce flexibility. They allow you to break down the budget categories that have historically constrained how you invest in customer experience. When a single AI Agent can take on work previously spread across in-house agents, BPOs, and point solutions, your job isn’t just to count the savings. It’s to reallocate them strategically.

If you treat your AI Agent as “just software,” these decisions don’t happen. But if you recognize it as a new workforce capability, it becomes clear: your budget should follow your operating model.

Blending budgets doesn’t mean cutting. It means realigning spend to match your operating model. You need to shift your thinking from “cost savings” to “budget reallocation.”

Every hour your AI Agent saves, every role you no longer need to backfill, every BPO contract you downsize: these are dividends that can be reinvested for long-term growth.

You can use these dividends to reinvest in:

You’ve proven that AI works. You’ve seen the early wins. But at scale, something bigger happens. AI stops being a tool and starts becoming infrastructure. Your customer experience gets faster, smarter, and more personalized. Your team evolves into system designers and strategic advisors. Support becomes a true lever for growth, not a cost center.

This is a chance to lead the way for your company, not just your team.

Not every team will get there. But you can. With the right systems, roles, and mindset, your AI Agent doesn’t just support customers. It becomes the foundation for a smarter, faster, more resilient business.

For support leaders, this is a career-defining opportunity. The chance to pioneer a new model. To show what’s possible. To build something the rest of the business follows.