Artificial intelligence is rapidly transforming how companies interact with customers. For businesses managing thousands of support conversations every week, scaling customer service without increasing headcount can be a major challenge.
In a recent customer panel hosted by HubSpot, Ryan Orner spoke with Andrea Martin from fintech company Wallet about how they implemented AI agents inside their CRM and the measurable results they achieved.
This real-world case study shows how AI-powered customer support—especially within a unified CRM—can dramatically improve efficiency, scale conversations, and deliver better customer experiences.
Wallet serves gig-economy workers—such as rideshare drivers and delivery couriers—who rely on mobile access to financial tools. Their users often need fast answers about cash advances, account status, or payment timelines.
Because their customers are primarily mobile-first, support interactions happen mostly through in-app chat.
However, the company faced several operational challenges:
At peak times, Wallet was receiving over 3,000 chats per week—a volume that simply wasn’t sustainable for a small team.
The result?
The company knew it needed an AI-powered solution that could handle routine conversations and operate 24/7.
While evaluating AI chat vendors, Wallet initially explored third-party AI chat platforms. The team was close to signing with an external provider when something unexpected happened.
Several team members attended INBOUND and discovered that HubSpot’s new AI agents could support mobile chat directly inside the CRM.
That discovery changed everything.
Instead of adding another tool to their tech stack, they realized they could deploy AI directly within the platform they were already using.
This mattered because:
Using AI inside the CRM meant the system could understand the full context of each customer interaction.
One of the biggest advantages of using AI directly within HubSpot CRM is access to real-time customer data.
This allowed Wallet’s AI agent to:
Compared to the third-party tools they evaluated, the response quality was significantly better because the AI had access to the company’s actual CRM data.
Many companies hesitate to adopt AI because they assume implementation will be complex.
In Wallet’s case, deployment was surprisingly quick.
Within less than one week, the team had:
Interestingly, the biggest effort wasn’t technical.
Instead, the most time-consuming part was training the knowledge base so the AI could provide accurate answers.
This reinforces an important lesson for AI adoption:
AI success depends heavily on the quality of your data and documentation.
Whenever companies introduce AI into customer support workflows, one concern often arises:
Will AI replace human jobs?
At Wallet, the opposite happened.
With thousands of weekly chats coming in, the support team welcomed the help.
Instead of replacing agents, AI allowed them to:
The AI agent handled repetitive tasks such as:
This freed human agents to focus on meaningful problem-solving.
Since implementing HubSpot’s customer AI agent, Wallet has seen dramatic results.
Previously, customer support only operated during business hours.
Now, customers can get answers any time of day, without waiting for human agents to come online.
The benefits go far beyond support efficiency.
Customers receive immediate answers instead of waiting in queues.
Human agents can focus on higher-value support cases.
AI provides around-the-clock assistance without increasing staffing costs.
By analyzing AI conversation data, the company can optimize staffing hours based on demand.
Wallet isn’t stopping with customer support.
They’re now exploring AI across additional departments, including:
Using tools like Breeze, the company plans to automate repetitive tasks across the organization.
The goal is simple:
Free employees to focus on the work that matters most—helping customers.
The Wallet case study highlights several important lessons for companies exploring AI adoption.
AI works best when solving specific operational bottlenecks.
Deploying AI inside your CRM reduces complexity and improves context.
Better documentation leads to better AI responses.
AI should remove repetitive tasks, not eliminate human expertise.
Track conversations handled, response times, and operational efficiency.
As AI capabilities continue to improve, companies will increasingly rely on CRM-embedded AI agents to manage conversations at scale.
Platforms like HubSpot are leading this shift by embedding AI directly into customer workflows rather than offering disconnected tools.
The result?
And as Wallet’s experience shows, companies that adopt AI strategically can unlock massive operational improvements without expanding their tech stack.
A customer AI agent in HubSpot is an AI-powered assistant that automatically responds to customer conversations using CRM data and knowledge base content.
HubSpot AI can automate responses, verify customer data, summarize conversations, and route complex issues to human agents.
No. AI is best used to automate repetitive tasks while human agents handle complex or sensitive customer issues.
Many companies can launch a proof of concept within days, depending on how well their knowledge base and CRM data are structured.
Industries with high conversation volumes benefit most, including fintech, SaaS, e-commerce, and customer support operations.