Transform Slack Confusion into AI-Driven Organization: Insights from Our Latest Podcast

When it comes to customer support, opening a ticket often feels like tossing a message into the abyss. You might receive a response eventually, if you’re lucky, but then what? The record of your issue often vanishes into an ocean of data, barely glanced at again.

For Tom Bachant, founder and CEO of Unthread, this seemed like an enormous waste of potential. Each conversation hides valuable insights about how people truly interact with a product, waiting to be discovered. Today, Tom is transforming this chaotic scenario into something far more productive.

His innovative company helps organizations manage IT and HR requests directly within Slack. This turns a flood of messages into organized, searchable support. But his latest creation, the Support Ticket Analyzer Agent, takes this concept a step further, leveraging AI to extract and highlight what truly matters in those conversations.

Navigating the Journey from Ride-Sharing to AI Solutions

Tom’s entrepreneurial path began long before AI became a household term. While studying at the University of Connecticut, he developed a ride-sharing app named Dashride, well ahead of Uber and Lyft. The goal was straightforward: provide students with a convenient and safe way to travel around campus.

This endeavor grew into a successful company, later acquired by Cruise, where Tom served as an engineering manager. Here, he gained invaluable experience in delivering products that impacted thousands while recognizing the transformative power of automation.

By the time he established Unthread, one theme was crystal clear: solving complex, human-centric problems through clever technology was his passion.

“I’ve always considered myself a hacker,” he shared. “It’s exhilarating to see how a simple line of code can revolutionize the way people work.”

Quelling the Chaos of Slack

Anyone familiar with Slack understands the overwhelming noise it can create. Important messages can get buried, requests disappear into lengthy threads, and the same question might be asked repeatedly.

Unthread aims to tame this chaos. It functions as an AI-powered help desk seamlessly integrated within Slack. For example, when someone requests a new laptop or a password reset, Unthread captures the request, monitors its progress, and automates responses.

Yet Tom recognized a potentially deeper layer within this communication. Each message and ticket presented pieces of operational truth, illuminating how teams actually work. What if this information could be analyzed to uncover larger patterns?

Crafting the Support Ticket Analyzer Agent

This curiosity led to the birth of the Support Ticket Analyzer Agent, now available on Agent.ai (and yes, it’s free!). The tool accepts CSV exports of support data—whether from Zendesk, Jira, HubSpot, or directly from Unthread—and utilizes advanced language models to identify recurring themes, sentiments, and levels of urgency.

“Summarizing data with AI is the easy part,” Tom explains. “The challenge lies in making that summary actionable.”

The agent doesn’t merely count occurrences of issues; it ranks them based on their impact, identifies urgent problems, points out documentation opportunities, and flags repetitive tasks that could be automated. For those responsible for customer experience or product management, it’s like having a data-savvy teammate already in the loop.

Initially created for Unthread’s internal use, Tom pushed the boundaries by running numerous support tickets through the tool. To his surprise, the AI consistently provided insights that aligned with—and occasionally surpassed—his own analysis.

“I realized I could step back; the system recognized the same patterns I did and communicated them even more clearly,” he shared.

Transforming Data into Actionable Insights

Every creator dreams of feedback mechanisms that enhance their processes. For Tom, the Support Ticket Analyzer became just that. His team now reviews its insights during sprint planning, determining which issues to prioritize or document next. Tasks that once consumed hours can now be completed in seconds, allowing the team to concentrate on impactful actions.

More importantly, Tom made a significant choice: he decided to share this tool with everyone. The underlying logic of the agent, which instructs the AI on how to interpret ticket data, is freely accessible through Agent.ai. Even those unfamiliar with Unthread can upload their own CSV files to unveil insights.

This philosophy of accessibility aligns with Tom’s vision. He believes that AI tools shouldn’t confine users to a specific ecosystem; they should integrate seamlessly into existing workflows.

“There shouldn’t be a need to overhaul systems to benefit from automation,” he emphasized. “If AI can grasp your data, it should interface with whatever tools you’re already using.”

Insights from a Builder’s Mindset

Tom’s journey offers an inviting blueprint for those venturing into AI. He didn’t start with a grand ambition of overhauling workflows. Instead, he focused on addressing a single, pressing challenge: navigating excessive support data and lacking time to analyze it.

This targeted approach kept the project grounded. Each evolution of the agent tackled a distinct aspect of the challenge, progressing from summarizing data to prioritizing issues and ultimately recommending next steps.

Additionally, Tom learned a vital lesson: AI thrives when combined with human insight.

“The agent provides a solid starting point,” he noted. “But we don’t rely on it entirely. It enables teams to align more swiftly, so they can invest their energy into making decisions, rather than sifting through hundreds of tickets.”

In practice, this human-AI collaboration is what makes the Support Ticket Analyzer truly valuable. It doesn’t seek to replace individuals but enhances their judgment and efficiency.

Broadening the Vision: Documentation that Self-Updates

Next on Tom’s agenda is tackling a problem that every organization faces: outdated documentation.

“We all have those Confluence pages or Google Docs gathering dust,” he chuckled. “Someone addresses the same question in Slack, but the documentation remains untouched.”

Unthread’s forthcoming Documentation Generation Agent aims to resolve this. It will analyze existing documents, compare them against real conversations, and automatically suggest updates or create new articles when gaps arise. The goal extends beyond saving time; it’s about preserving crucial knowledge within the organization.

As Tom described, the agent will handle the often-dull task of identifying and drafting updates, with humans simply reviewing and approving the changes. This philosophy mirrors Unthread’s support tools: let AI manage the heavy lifting, allowing individuals to focus on what truly matters.

Why This Matters for Everyday Users

For many, the notion of building an agent might seem daunting. Yet, Tom’s story illustrates that it doesn’t have to be complicated. He created the Support Ticket Analyzer as a helpful experiment, solving a real issue for his own team, which ultimately led to developing something beneficial for others.

That’s the beauty of AI agents. They often begin small—perhaps as an automation, prompt, or script—and can evolve into tools that genuinely enhance productivity.

Whether you oversee customer support, manage operations, or simply wish to decode your organization’s communication noise, Tom’s agent exemplifies the potential when AI transforms listening into learning.

If you’ve ever pondered the hidden messages within your support tickets, now is the perfect moment to explore! You can try Tom Bachant’s Support Ticket Analyzer Agent directly on Agent.ai. Simply upload a CSV file from your system, let the agent do its magic, and discover the insights waiting to surface.

You may stumble upon your next groundbreaking product idea, a crucial process enhancement, or an overdue documentation update, all nestled within the everyday dialogues you already have.