DSW’s CX Strategy? Let Customers Choose

Table of Contents

• Built From Stores to Contact Center: A CX Leader’s Unconventional Path
• Building a CX Function From Scratch
• Rapid Growth Forced a New Operating Model
• Automation and Partnership Helped DSW Scale Smarter
• Automation Paid Off When the World Shut Down
• AI Has Limits — and Brands Need to Acknowledge Them
• A Global Team With a Growing Focus on Virtual Agent Oversight
• Why DSW Chose to Buy — and Not Build — Its AI Capabilities
• Customer Sentiment Drives How Virtual Agents Improve
• The Human-in-the-Loop Model Extends Beyond QA
• From Information to Action: The Maturing Role of Virtual Agents
• Patterns in the Data Are Driving Smarter CX Decisions
• What’s Next: Virtual Agents That Can Complete Purchases
• Eliminating Friction Means Letting Customers Choose the Channel
• Designing CX Around Customer Preference, Not Constraints
• Voice Still Leads — and That’s by Design
• Expanding Automation Beyond Core Use Cases
• Agent Experience Starts With Trust and Autonomy
• Why Empowerment Reduces Agent Stress — and Improves CX
• AI Is Now Improving the Agent Experience, Too
• Measuring Agent Performance Beyond Speed Alone
• Why Metrics Must Work Together — Not in Isolation
• Why Slowing Down Can Lead to Better Outcomes
• Closing the Loop Between Metrics and Experience
• Renewing the Brand Experience Beyond the Contact Center

The Gist

• DSW builds customer operations around choice, not channel forcing. Tim Harpe explains that customers should be able to interact with the brand however they want, whether that means voice, virtual agents, chat, email or a live human agent.
• Automation works best when humans stay in the loop. DSW uses sentiment analysis, QA reviews and close partner collaboration to continuously improve virtual agent performance instead of treating AI as a set-it-and-forget-it solution.
• Agent empowerment remains the heart of the experience. Harpe says DSW gives human agents autonomy, removes unnecessary guardrails and uses AI to reduce friction, helping agents solve problems instead of getting trapped by scripts and speed metrics.

Editor's note: This interview was edited for clarity.

Built From Stores to Contact Center: A CX Leader’s Unconventional Path

Tim Harpe doesn’t come from a traditional contact center background — and that’s exactly the point.

After nearly two decades at DSW and a career that started in retail stores, Harpe brings a frontline-first perspective to customer operations. He’s run stores, districts and regions, and spent time in marketing operations before stepping into what was, at the time, a brand-new role leading customer service.

“Never spent any time in a contact center and not even on my radar," Harpe said,

That didn’t matter to DSW leadership. What they needed wasn’t contact center tenure. They needed someone who understood the brand, the culture and what it would take to build something from the ground up.

Fifteen years later, that decision still shapes how DSW approaches customer operations and customer experience.

Building a CX Function From Scratch

When DSW first launched its ecommerce business, customer service barely existed in real time.

Harpe says the company initially had only a small group of associates answering voicemails. But once DSW.com launched, that setup quickly stopped making sense. Customers shopping online needed support in the moment, not delayed callbacks, and the company moved to create a formal contact center operation.

About a year later, DSW took the idea further and rebranded the contact center as Shuoephoria, a name meant to capture the excitement, energy and emotion customers attach to shoes. For Harpe, the goal was to preserve the brand feel of the in-store experience even when the interaction was happening over the phone or through email.

“We needed to build a place where it felt like you were still interacting with the DSW that you know and love from being in a store,” Harpe says. “You’re just doing it in a different avenue.”

Rapid Growth Forced a New Operating Model

That early model scaled quickly. Harpe says the team started with about 40 associates, then eventually grew to more than 650.

But that growth created a new constraint. DSW was still operating in the same physical space it had used at the beginning, even as the team expanded dramatically. Harpe expected the next step would be a larger contact center footprint.

Leadership had a different idea.

When Harpe brought forward a plan for a new contact center, he says his president pushed back immediately. The business was not going to solve the problem by growing physically. If anything, the operation needed to get smaller and more efficient.

That became a turning point. Instead of tying scale to more people and more space, DSW had to rethink how the work got done.

Automation and Partnership Helped DSW Scale Smarter

By 2018, DSW identified two ways forward: automation and external support.

Harpe notes that automation was not yet as dominant a topic as it is today, but the opportunity was already clear enough for DSW to move. The company began using automation as one lever to reduce pressure on the team and handle growing demand more efficiently.

At the same time, DSW brought in an outside partner to support work that could no longer live entirely inside its own four walls. Harpe says that partner relationship remains in place today.

On the automation side, DSW partnered with SmartAction in 2018. Through acquisition, that company later became part of Capacity, but Harpe says the relationship has continued to produce strong customer experiences.

Together, those two decisions — investing in automation and bringing in a trusted partner — helped DSW solve its scale challenge in 2018 and 2019. Then 2020 arrived, and the environment changed again.

Related Article: The CX Reckoning of 2025: Why Agent Experience Decided What Worked

Automation Paid Off When the World Shut Down

When the pandemic hit in 2020, DSW’s earlier investment in automation quickly proved its value.

Harpe says the company was able to rely heavily on its virtual agent during a critical transition period, when live agents were temporarily unavailable while teams shifted from centralized contact centers to remote work environments.

“We have never been more happier to have, at that time, SmartAction as our virtual agent provider,” Harpe says. “So much of our volume was able to be handled through that virtual agent when our live agents weren’t available.”

That early foundation in automation helped DSW navigate one of the most disruptive moments in modern customer operations. While the transition to remote work still required major logistical effort, the company avoided a complete breakdown in customer support.

Harpe says that experience reinforced a broader lesson: automation works best as part of a long-term strategy, not a reactive fix.

AI Has Limits — and Brands Need to Acknowledge Them

Even as automation and AI capabilities have evolved, Harpe pushes back on the idea that technology will replace the need for human-led customer service.

He says the narrative that AI will solve every problem in the contact center is overly simplistic and often disconnected from operational reality.

“It’s a great theory,” Harpe says. “But there are limitations to it.”

Those limitations show up in multiple ways. Brand expectations play a role in determining how far automation can go. Backend systems and data availability create additional constraints. And perhaps most importantly, customer preferences vary widely.

“Not every customer is created equal and not every customer sees the value in working with a virtual agent versus a live agent,” Harpe says.

For Harpe, the key is balancing efficiency with experience. Cost savings alone cannot drive decision-making in a customer-centric organization.

“At the end of the day, those are the two people you have to satisfy,” he says, referring to the brand and the customer. “It’s not what’s going to save me the most money.”

A Global Team With a Growing Focus on Virtual Agent Oversight

Today, DSW’s customer operations team is distributed globally, with a core group based in Columbus, Ohio, focused on training, engagement and real-time operations management.

But one of the most notable shifts has been how the organization supports its virtual agents.

As automation capabilities have expanded, DSW has also expanded the human roles responsible for overseeing those systems. Harpe says the company has transitioned much of its quality assurance effort away from live agents and toward evaluating virtual agent performance.

Rather than treating automation as a separate channel, DSW evaluates virtual agents using the same standards applied to human agents.

“We’re going to train these agents. We’re going to QA them. We’re going to monitor feedback — everything the same way that we do with live agents,” Harpe says.

The goal is consistency. Customers should not feel a disconnect between interacting with a virtual agent and speaking with a human representative.

“What’s important from a brand perspective is that you don’t have these different experiences,” Harpe says.

Why DSW Chose to Buy — and Not Build — Its AI Capabilities

When it comes to managing virtual agent technology, DSW has made a deliberate choice to rely on external expertise.

Harpe says companies typically face a familiar decision in the contact center space: build technology internally, or buy it and either manage it themselves or rely on a partner.

For DSW, the decision was to buy and have a partner manage the platform. That approach allows internal IT resources to stay focused on other business priorities, while specialists handle the complexity of maintaining and improving the virtual agent.

That’s where the company’s partnership with Capacity continues to play a central role, he said.

Harpe says DSW’s team provides direct feedback based on real customer interactions, highlighting where the experience can be improved or made more intuitive. Capacity then takes that input back to its developers and CX teams to refine the system.

The result is a feedback loop that continuously improves the virtual agent without requiring DSW to build and maintain the technology itself.

“They are so receptive to what we see and hear through these interactions,” Harpe says. “We can provide them, here’s what we’re hearing, here’s where we think it could be a little better.”

This model, Harpe says, has proven to be the most effective way for DSW to scale its automation capabilities while maintaining control over the customer experience.

Related Article: Why Agent Experience Just Became the Center of CX

Customer Sentiment Drives How Virtual Agents Improve

At DSW, evaluating virtual agent performance starts with one core signal: customer sentiment.

Harpe says every interaction handled by a virtual agent is scored for sentiment, giving the team a consistent way to understand how customers feel about the experience. But the process doesn’t stop with automated scoring.

Human reviewers step in to validate whether the sentiment aligns with what the business would consider a successful or unsuccessful interaction. From there, the team digs deeper into the drivers behind each outcome.

“We start with the customer sentiment from each of those interactions,” Harpe says. “We then go in and verify that that sentiment matches what we would deem as a successful call or a non-successful call.”

That validation process helps identify gaps — including scenarios where the virtual agent could not complete a task. Those gaps then become inputs for future development.

The result is a continuous improvement loop driven by real interactions, not assumptions.

10 Lessons From DSW’s Customer Operations Playbook

Editor’s note: Key takeaways from Tim Harpe’s Beyond the Call appearance on how DSW blends automation, agent empowerment and customer-first design.

Lesson	What It Means for CX Leaders
1. Let customers choose the channel	DSW avoids channel forcing and designs experiences so customers can move seamlessly between voice, virtual agents and human support based on preference.
2. Build automation early — not reactively	Investing in virtual agents before COVID gave DSW resilience when disruption hit, proving automation works best as a long-term strategy.
3. Keep humans in the loop	Sentiment analysis, QA validation and partner collaboration ensure AI is continuously improving instead of operating in isolation.
4. Treat virtual agents like real agents	DSW applies the same QA, training and performance expectations to virtual agents to ensure consistent brand experiences.
5. Don’t chase AI hype blindly	AI has limits tied to data, systems and customer expectations, so decisions must balance efficiency with brand experience.
6. Empower agents with autonomy	Removing rigid scripts and decision trees allows agents to solve problems faster and deliver more human, personalized service.
7. Measure performance holistically	Customer satisfaction, handle time and QA must work together — no single metric tells the full story of agent performance.
8. Speed alone doesn’t equal success	Fast interactions can miss key details; sometimes better CX comes from spending more time understanding and resolving the issue fully.
9. Use AI to reduce agent friction	Tools like auto summaries and real-time analytics free agents from administrative work and improve interaction quality.
10. Let data reveal systemic issues	Patterns in virtual agent interactions — like shipping-related inquiries — can uncover operational problems beyond the contact center.

The Human-in-the-Loop Model Extends Beyond QA

While automation plays a growing role, Harpe emphasizes that humans remain deeply embedded in the system.

That “human in the loop” concept applies not only to quality assurance, but also to how DSW works with its technology partners.

Rather than building and managing AI internally, DSW collaborates closely with partners who specialize in virtual agent development. Harpe says that approach reflects a practical reality for many CX organizations.

“I don’t have an army of data scientists sitting on my team,” he says. “I don’t have an army of IT professionals who can’t wait to build the next generation of an AI agent.”

Instead, DSW focuses on identifying opportunities, sharing insights and shaping direction, while its partners handle the technical execution.

“We partner with them on ideas and then it comes to fruition,” Harpe says.

That same dynamic works in reverse, with partners bringing new ideas back to DSW based on what they’re seeing across clients and use cases.

The relationship becomes less about vendor management and more about shared ownership of outcomes.

From Information to Action: The Maturing Role of Virtual Agents

As DSW’s virtual agent capabilities have evolved, so has the type of work those systems can handle.

Harpe describes a clear progression, starting with basic informational use cases.

Early on, the virtual agent focused on answering common questions — especially high-volume requests like order status. In a retail environment, “Where’s my order?” is one of the most frequent customer inquiries, making it a natural starting point for automation.

From there, the focus shifted from providing information to taking action.

DSW began enabling its virtual agent to complete tasks on behalf of customers, such as updating an address or canceling an order. That shift marked a move toward more advanced, agentic capabilities.

“Now it’s starting to do things for you and not just give you information,” Harpe says.

Patterns in the Data Are Driving Smarter CX Decisions

As virtual agent usage has expanded, DSW has also gained clearer visibility into customer behavior and operational friction points.

Harpe says one key insight has come from analyzing patterns tied to shipping and logistics partners. When certain carriers consistently generate more customer inquiries, it signals a breakdown in the experience.

Those patterns create opportunities for targeted improvements, whether through more proactive communication, better tracking visibility or closer coordination with partners.

“There always seems to be more contacts from one provider than another,” Harpe says, pointing to how those insights help guide operational fixes.

Instead of treating each interaction as a one-off issue, DSW uses aggregated data to identify systemic problems and address them at the source.

What’s Next: Virtual Agents That Can Complete Purchases

Looking ahead, DSW is exploring even more advanced capabilities — including virtual agents that can place orders on behalf of customers.

Harpe says that idea would have felt unrealistic just a few years ago. But improvements in technology and ongoing collaboration with partners have made it a viable next step.

“If someone said to me four or five years ago that you’re going to have a virtual agent who will be able to place orders, I’d be like, I don’t know,” Harpe says. “But with the technology where it is now, that’s not a far-fetched idea at all.”

As these capabilities expand, DSW’s guiding principle remains consistent: give customers control over how they interact with the brand.

Some customers will prefer self-service through a virtual agent. Others will want to speak with a human. Harpe says both options need to be equally accessible.

Eliminating Friction Means Letting Customers Choose the Channel

For Harpe, one of the most important elements of modern customer experience design is minimizing friction between channels.

That includes making it easy for customers to move between virtual and human support without barriers.

“When they want to use the virtual agent, they can use it seamlessly,” Harpe says. “But when they want a live agent, they can get to that live agent seamlessly.”

While that approach may sound straightforward, Harpe notes that not all providers share the same philosophy. Some prioritize cost control over flexibility, discouraging escalation to human agents.

DSW takes the opposite view. If a customer wants to speak with a person, the experience should support that choice — even if it comes at a higher cost.

“That’s what the customer wants,” Harpe says.

In a world where many brands still force customers into specific channels, Harpe’s approach reflects a broader shift: designing customer experience around preference, not constraints.

Designing CX Around Customer Preference, Not Constraints

For Harpe, one principle continues to guide how DSW builds its customer experience: let the customer choose how they engage.

“We’re not here to dictate how you interact with us,” Harpe says. “We’re here to make sure that when you interact with us, you do it in a way that you’re comfortable with and that you get the support that you want.”

That philosophy shows up in how DSW approaches channel strategy. While many brands continue to push customers toward lower-cost digital channels, DSW is seeing a different reality in its own data.

Voice Still Leads — and That’s by Design

Despite the rise of chat and automation, voice remains DSW’s dominant channel.

Harpe says the company’s customer base still prefers the phone, reinforcing the idea that channel strategy must align with customer demographics, not industry trends.

“Voice is still our number one channel of choice,” he says.

That doesn’t mean digital channels are lagging. Chat and email are nearly tied as the second and third most-used channels, with performance influenced in part by how DSW manages response times and service levels.

Harpe notes that email, often a pain point for customers at other organizations, is handled with speed and consistency, avoiding the multi-day delays that frustrate many consumers.

Expanding Automation Beyond Core Use Cases

As DSW continues to refine its automation strategy, it is moving beyond tightly defined use cases and into more flexible, conversational interactions.

Harpe says the addition of frequently asked questions within the voice experience has opened up new possibilities. Customers can now ask broader, more spontaneous questions and receive immediate answers without needing to navigate rigid workflows.

That success is shaping how DSW thinks about its next phase of channel expansion.

The company is now looking to bring similar capabilities into chat, using lessons learned from voice interactions to create more natural, responsive digital experiences.

“Some random questions — if they can answer it, great,” Harpe says. “That’s all they wanted. They just wanted to know.”

Agent Experience Starts With Trust and Autonomy

While automation continues to evolve, Harpe is clear that human agents remain central to the customer experience — and their success depends heavily on how they are empowered.

At DSW, that starts with autonomy.

Harpe says the company has deliberately avoided building rigid scripts, decision trees or excessive approval layers. Instead, agents are trained to use what he describes as a “service compass,” guiding them to respond based on the customer’s needs in the moment.

“We make sure that every one of our agents has the autonomy to do whatever they need to do in the moment to take care of the customer,” Harpe says.

That approach removes the friction that often slows down resolution, such as transferring calls between departments or requiring multiple approvals for simple fixes.

It also reflects a broader cultural belief in self-expression — not just for customers, but for employees.

“We come from the heart,” Harpe says. “We treat people the way they are needing to be treated in the moment.”

Why Empowerment Reduces Agent Stress — and Improves CX

Giving agents more control doesn’t just improve customer outcomes. It also changes how agents experience their work.

Harpe says one of the biggest stressors in contact center roles is the need to default to “no” — or to navigate complex rules before saying “yes.”

By removing those constraints, DSW allows agents to approach each interaction with the mindset that they can solve the problem.

“When that’s your starting point, those interactions are way easier to handle,” Harpe says.

Instead of managing limitations, agents focus on resolution. That shift not only improves customer satisfaction but also helps reduce burnout in what is often a high-pressure role.

AI Is Now Improving the Agent Experience, Too

After initially focusing AI efforts on customer-facing use cases, DSW has begun applying the same technology to improve agent efficiency.

Harpe points to tools like automated call summaries, which eliminate the need for agents to manually document every interaction after it ends. That alone reduces administrative burden and allows agents to move more quickly between conversations.

Advanced analytics also play a role, helping identify where agents may need additional support and where they are performing well.

In real time, agents can see signals about how customers are responding during an interaction, giving them greater awareness and the ability to adjust on the fly.

“There’s total transparency and clarity,” Harpe says, describing how those tools help agents understand where they stand in each conversation.

Together, these capabilities reinforce DSW’s broader approach: combine autonomy with insight, so agents are both empowered and informed.

Measuring Agent Performance Beyond Speed Alone

When it comes to evaluating human agents, Harpe returns to a familiar starting point: customer feedback.

Customer satisfaction scores remain the primary signal for understanding how well an agent is performing, even as some in the industry question whether those metrics come too late to influence outcomes.

“It all starts with the customer feedback,” Harpe says.

From there, DSW layers in additional metrics — but always in context. Handle time, for example, is not ignored, but it is not treated as a standalone indicator of success.

Instead, Harpe says performance is evaluated by combining efficiency, effectiveness and customer perception into a single view.

“If it takes you 20 minutes to solve a problem and your customer satisfaction is no better than the person who it took them 10 minutes, what value did you add with that additional time?” he says.

At the same time, speed without substance creates its own problems. Agents who rush through interactions may miss key details, fail to resolve the issue fully or create repeat contacts that ultimately degrade the experience.

Why Metrics Must Work Together — Not in Isolation

Harpe cautions against over-indexing on any single metric.

Handle time alone can be manipulated. Customer satisfaction alone may not reflect long-term outcomes. And efficiency without quality can create downstream issues that are harder to detect in the moment.

To address that, DSW combines multiple data points — including customer satisfaction, handle time and quality assurance — to create a more complete picture of agent performance.

“As long as there’s a component to how efficient you are, how well you’re using the tools you have available, and how well your customer is perceiving the support you’re providing, you walk away with a pretty good understanding,” Harpe says.

Quality assurance plays a critical role in that mix, helping ensure that what agents promise during interactions is actually delivered.

Without that layer, even strong performance metrics can mask underlying issues.

“I can have the best C-Stat in the world,” Harpe says. “But everything I promised a customer never ever happened.”

Why Slowing Down Can Lead to Better Outcomes

In an industry often driven by speed, Harpe says there are times when agents need to do the opposite.

DSW frequently coaches new hires — especially those coming from more rigid environments — to spend more time with customers, not less.

“We need this person on the phone longer,” Harpe says, describing conversations with team leaders.

That additional time allows agents to ask better questions, provide more complete answers and anticipate follow-up needs — all of which contribute to stronger outcomes.

When agents are overly focused on getting through interactions quickly, those opportunities disappear, and the quality of the experience suffers.

The same principle applies to virtual agents. Faster responses are not always better if they lack clarity or completeness.

“The one that’s super fast wasn’t the best,” Harpe says, recalling comparisons between different solutions.

Closing the Loop Between Metrics and Experience

Ultimately, Harpe’s approach reflects a broader shift in how customer operations leaders think about performance.

Instead of optimizing for individual metrics, DSW focuses on outcomes — ensuring that customers leave interactions with their issues fully resolved and their expectations met.

That requires balancing speed, accuracy and empathy across both human and virtual channels.

It also reinforces a consistent theme throughout DSW’s customer experience strategy: the goal is not to move faster, but to serve better.

Renewing the Brand Experience Beyond the Contact Center

While much of Harpe’s focus is on customer operations, he says the broader DSW brand is also evolving in ways that will shape the customer experience in 2026 and beyond.

One of the biggest shifts is a renewed focus on product assortment.

Coming out of the pandemic, DSW — like many retailers — leaned heavily into categories that reflected changing customer behavior, particularly athletic and casual footwear. But that shift also created gaps in other areas that had historically been part of the brand’s strength.

Harpe says the company is now working to rebalance its assortment, bringing back a stronger mix of dress and casual comfort options while maintaining its gains in athletic.

“We’re really going back and finding pockets of the assortment that we maybe walked away from,” he says.

Alongside assortment changes, DSW is also preparing to roll out an updated loyalty program designed to re-engage its customer base.

While Harpe stops short of sharing details, he says the goal is to build on an already highly engaged and vocal customer community.

“Our customer base is so loyal and they’re extremely vocal,” he says. “That’s a fantastic thing because they care about what’s happening.”