How Firm360 turned its support stack into a self-improving system with AI

Gonzalo Cayo

Chief Customer Officer

How Firm360 turned its support stack into a self-improving system with AI

Gonzalo Cayo

Chief Customer Officer

How Firm360 turned its support stack into a self-improving system with AI

Gonzalo Cayo

Chief Customer Officer

Company

Firm360

Industry

Complex SaaS

Company size

Scale-up

The headcount question

Firm360 started how the best businesses often do: someone fed up with the tools they had, with an idea for a better solution, and a small group of like-minded professionals willing to try it first. What began as a tool one accountant needed for his own firm has grown into a platform used by accounting firms across the United States.

Today, the company faces a familiar challenge: increases in customer volume require increases in team capacity. When Gonzalo Cayo joined a year ago to structure operations, the choice was a Customer Success Manager to drive renewals and upsell, or dedicated support agents to take tickets off the people doing onboarding and customer success.

A self-improving knowledge base

The most distinctive thing about how Firm360 runs Lorikeet is what happens to the knowledge base. The team treats Lorikeet as a contextual data layer, not a frontend. Every two weeks, Firm360 runs analysis against three sources at once: Lorikeet's MCP server, Slack, and recent customer conversation transcripts. The output is a ranked list of knowledge base gaps with draft answers attached.

"I run analysis using Claude and the Lorikeet MCP server. The MCP server works really, really well. Right off the bat it works. I aggregate all these data sources and that gives me a succinct list of: these are the knowledge base articles to update and here are the answers. That's how I'm feeding the machine, and the gaps get less and less every time. This results in a better experience not only for our team, but most importantly, for our customers." — Gonzalo Cayo, Chief Customer Officer

Once new articles are published, Lorikeet's Auto QA reviews them continually. The next time a similar conversation comes in, Lorikeet handles it end-to-end. With each cycle the gaps decrease.

This is the loop that compounds. Most AI support deployments stall when a knowledge base isn't updated, because nobody owns the work of finding gaps. At Firm360, the work of finding gaps is automated, and the work of closing them is heading the same way.

From support to operations

At Firm360, the teams Lorikeet assists are not support agents. They are onboarding specialists, CSMs, and ops leads, all answering tickets between other priorities. By helping reduce the load, Lorikeet acts as both customer operations and customer experience. It makes sense that the next use case is internal. Lorikeet also lives in Slack, where the Firm360 sales team can ask product questions during demos. Months of historical Q&A in that channel become the seed context for an internal-facing knowledge base that the sales team can talk to in real time to improve and scale capability without scaling the team.

"Imagine a customer has a complex issue that would require the Firm360 team to go away and determine a solution, wasting the customer's valuable time ... now the team can just ask Lorikeet, get those responses in the moment, and immediately act on customer needs. I went from eight people to fifty-four people at a previous company. Hiring is a full-time job, onboarding is a full-time job. Lorikeet reduces that load, and over time it'll make everyone more efficient." — Gonzalo Cayo, Chief Customer Officer

Read other stories

Connect, configure and preview

Read other stories

Connect, configure and preview

Read other stories

Connect, configure and preview

Read other stories

Connect, configure and preview