Another day. Another product question teams aren’t sure how to answer.
A customer service engineer knows the solution exists somewhere in a 2,400+ article knowledge base, but they’re knee-deep in documentation and Slack conversations. By the time insights arrive, valuable time’s been lost — and your customer is still waiting for an answer.
Box has been there, and we’d like to share with you how we cut service questions in half and saved 1,000+ hours better spent on valued client interactions.
Key takeaways
- AI-powered knowledge management is deflecting an estimated 50% of questions, freeing teams to focus complex problem-solving for customers
- Along with dramatically improving response times and customer satisfaction, this Box use case demonstrates how cost savings can be achieved
- By preserving existing workflows and focusing on access rather than process change, Box achieved rapid adoption without disruption
Time savings hidden in plain sight
Box had an extensive support knowledge base resulting from years of customer interactions.
Thousands of meticulously crafted articles covered every conceivable customer scenario. But critical answers lived in different locations — from Zendesk and Slack threads — and Box needed to deliver fast insights.
“Customers expect support at the speed of their problem,” says Scott Juster, Director & Chief of Staff at Box Customer Success.
Meeting that expectation meant reimagining how support teams gather information. Every internal question, every “quick check” in Slack, every cross-team consultation represented a chance to streamline.
You’d have to manually go through a Zendesk article to find out what your customer needs. It’s cumbersome. And you’re on the clock. This allows teams to self-service, and it’s fast.
“Before, when looking for answers from support, the majority of my time was either spent perusing the support website, asking around in different Slack channels, or pinging other contacts I knew may have the answer,” Aaron Spriggs, Senior IT Program Manager at Box, explains.
Stephanie Slate, Box Senior Value Engineer, adds context: “You’d have to manually go through a Zendesk article to find out what your customer needs. It’s cumbersome. And you’re on the clock. This allows teams to self-service, and it’s fast.”
Numbers that drove the solution
When Box teams quantified the opportunity, the numbers spoke volumes:
- Customer Service managed 20 Slack channels
- Teams answered 500 questions per channel annually
- That amounted to 10,000 questions per year
- Answering each request 7.5 minutes on average (two employees per exchange)
- Ultimately, the process ate up 1,250 hours annually
And while their much-needed knowledge was right at their fingertips in Box, they needed fast ways to harness it. With Box AI, this wealth of information could be instantly and securely routed to the people who needed it most. So engineering and customer service experts got to work on an elegant simplification: an AI-powered support portal.
The solution leveraged three core Box technologies:
1. Automated knowledge integration
The team created API calls to Zendesk, automatically pulling in all support articles. No manual uploads, no version control challenges, no lag between documentation updates and availability. The comprehensive information repository that previously required manual searching was now instantly accessible.
2. AI-powered intelligence
Box AI, trained specifically on the support content corpus, became the intelligent layer that could parse natural language queries and deliver precise answers. Unlike generic search that returned dozens of potentially relevant articles, Box AI understood context and intent, serving up exactly what was needed.
3. Source attribution and trust
Support teams could verify answers, dive deeper as needed, and leverage the system to accurately source answers. No process overhauls, no retraining documentation teams, no disruption to existing workflows.
Immediate and profound results
With a solution in place, teams were empowered to deliver instant replies and more of the exceptional service that customers have come to know at Box. Now, more time is spent on high-level challenges, not hunting for information that already exists.
“Nearly all of that searching has been cut out by being able to ask our Hub what I’m looking for and getting a concise answer that references our entire pool of documentation,” Spriggs reports.
Estimated impact:
- 50%+ of questions resolved through self-service
- 5,000 questions deflected from Slack channels to self-service annually
- 1,000+ hours returned to high-value work
Nearly all of that searching has been cut out by being able to ask our Hub what I’m looking for and getting a concise answer that references our entire pool of documentation.
The multiplier effect:
Slate explains that when customer service engineers can instantly access information, several benefits emerge:
- Customer response times improve dramatically when engineers have immediate access to answers
- First-contact resolution rates increase as teams arrive prepared with comprehensive information
- Knowledge gaps become visible when the AI can’t find answers, highlighting documentation opportunities
- Onboarding accelerates as new team members can self-serve from day one
What these results mean for your business
Years of documentation efforts can result in information repositories that teams struggle to navigate efficiently, but as Box shows, any organization can overcome this hurdle with the right tools in place. Box Chief Customer Officer Jon Herstein explained that organizations can solve these challenges by simplifying access with an agent, as Box did.
“With Box AI, it’s very applicable to anyone who sells a product and has a need to support it,” he said. “Whether you’ve got policy knowledge bases, HR policies, or SOPs, you can do the exact same thing.”
He added that when critical information lives across multiple systems and team members, businesses miss opportunities for consistency and scale. Centralizing access connects information to people who need it so they can focus on strategic initiatives rather than repetitive questions.
Herstein offers a blueprint for any organization looking to optimize support efficiency:
- Determine whether customer-facing employees struggle to find accurate, up-to-date information
- If the answer is yes, you’ve confirmed you can benefit from a customer service agent
- In creating your agent, ensure you’re confident with content you’re referencing, and (if possible) start with content you’ve already published
- Solve any content sourcing issues before creating your agent
- Implement without disruption: Your agent should enhance, not stifle, current tools and workflows
Learn more about customizing AI-powered support portals in Box Hubs.
