Overview
If you're working at a startup, speed and efficiency aren't nice-to-haves; they’re essential for survival. In the FinTech space, companies also face the added hurdle of compliance, where even one minor misstep can completely derail momentum.
So how does a young, ambitious company navigate the delicate balance between rapid growth and rigorous regulatory demands?
FrankieOne made a bet: that by incorporating agentic AI into different facets of the business, every team could move faster, operate more independently, and scale without sacrificing quality.
Besides, you don’t become Fintech Company of the Year without finding smart, scalable ways to stay ahead. Here’s how Cassidy became a key part of FrankieOne’s success.
Removing the Friction to Move Faster
As noted, FrankieOne operates in one of the most demanding sectors: FinTech.
Though the finance world is traditionally not known for its speed, startups in this space are playing catch-up, working to innovate within a framework shaped by decades-old rules and regulations.
As FrankieOne began building new systems and processes, they had to navigate the pressure of adhering to those laws while also being held to a high standard: delivering polished products quickly, offering precise and informed support, and keeping operational friction low.
“We always want to be technology-forward,” said COO Warren Oakes. “But there’s also good pressure from our VCs to make sure we’re using the latest tools and being the most efficient.”
This isn’t just a startup issue. Scaling efficiently and making the most of your team’s time is the expectation across different industries.
And as FrankieOne evolved, that expectation became harder to meet without better systems in place. Several key friction points became increasingly hard to ignore:
- Support tickets were piling up, even the easy ones. Many incoming tickets were low-effort to resolve but sat untouched because they weren’t flagged as urgent.
- Customer onboarding diagnostics were eating up time. When a customer failed onboarding, someone had to dig through audit trails and logs to figure out why. It could take up to 30 minutes per case, dragging down operational speed and tying up valuable time.
- Release cycles kept getting tripped up by small errors. Manual configuration issues were creating friction at key moments in the development cycle. Even minor mistakes were enough to delay timelines and create pressure for engineering teams.
- Support answers weren’t always customer-ready. Engineers understood the product deeply, but their responses weren’t always clear or empathetic when responding to customers. This led to extra QA and sometimes created confusion or inconsistent tone.
FrankieOne wasn’t looking to tear things down and rebuild. They needed practical tools that could reduce friction, support faster execution, and give teams the clarity and confidence to keep moving forward.
That search led them to Cassidy.
When Cassidy Came In
Like many forward-thinking startups, they started by asking: Where can we be faster? Smarter? More consistent? That curiosity led them to Cassidy.
One of their early champions, Vanessa (Data Project Manager), discovered the platform while searching for tools that could consolidate internal knowledge and deliver smart, actionable answers across the company.
Cassidy was brought in as a low-risk experiment. The first use case? A smart assistant embedded in Slack, designed to pull answers from internal sources like Confluence and past Slack threads.
Starting with Slack: Frankie Help becomes an instant hit
Cassidy started as a quiet experiment. The team embedded it into Slack to see if it could help answer repeat questions that typically clogged up channels like “Frankie Help.”
It didn’t take long to prove itself. By connecting Cassidy to their Confluence documentation and past internal discussions, FrankieOne had a live knowledge assistant inside Slack that could respond in seconds.
"We relied on the goodwill of team members to respond. Now questions are answered right there in slack" - Warren Oakes
It became one of the company’s most used tools helping people move faster without having to interrupt each other or wait on a reply.
Onboarding diagnostics in seconds
One of the earliest internal wins came from a frustrating bottleneck: onboarding failures. Previously, if a customer failed verification, someone had to dig through audit logs and system responses to figure out why. Each case could take up to 30 minutes.
Using Cassidy, they built a diagnostic assistant that let the team input an entity ID and instantly get a breakdown of what went wrong and what to check next.
“It would take 30 minutes… but now we get a near-instant answer.” - Warren Oakes
Though this workflow was more custom than off-the-shelf, it highlights what’s possible when teams combine imagination with real operational needs. It’s a great example of how Cassidy can go beyond the pre-built use cases to support highly specific, high-impact solutions tailored to your organization.
Automating ticket triage and status page updates
Support operations also leveled up quickly.
FrankieOne gets a steady stream of support tickets. While many are simple, they still take time to triage. To speed things up, the team built workflows in Cassidy that could classify and respond to repeat tickets, and process vendor outage emails by automatically updating their public status page.
“The delay is actually time to get eyes on it. If someone looked at [the ticket], the answer would be quick.” -Warren Oakes
By automating the front end of this process, FrankieOne made a noticeable dent in response times and support overhead:
- 11% of all support tickets are now fully automated
- Median ticket resolution time dropped from four days to just over two
This helped the team stay on top of volume without relying on constant manual triage, and freed up time for higher-impact work.
Bringing real-time support into the customer portal
Once Cassidy had proven itself internally, the next step was to bring that same level of responsiveness to FrankieOne’s customers.
The team embedded Cassidy directly into their customer portal, giving users the ability to ask questions and get real-time answers without submitting a form or waiting for a support agent.
For a company operating in a high-trust, high-speed environment like fintech, this wasn’t just a UX improvement. It was a way to deliver answers with the same urgency and accuracy customers expect from the product itself.
“We now have literally the Cassidy UI in the bottom right-hand corner… so customers can directly ask questions. And obviously, if they can’t get an answer, they can submit a request.” -Warren Oakes
The bot is powered by the same knowledge base used internally, but tuned for an external audience. When Cassidy can’t provide a response, it escalates cleanly and passes the relevant context to the support team so they can jump in without backtracking.
Even in early stages, the AI-powered portal assistant has already helped reduce pressure on support, shorten response times, and set a new standard for how FrankieOne delivers real-time service at scale.
Helping UX and Product Teams move even faster
Cassidy also found its way into product and UX.
Teams started using it to generate discovery questions for user interviews and to help structure early product thinking. As usage evolved, they began asking Cassidy to review product requirement documents, rate their completeness, and suggest improvements.
It gave teams a way to self-audit their work, refine ideas quickly, and move into build cycles with more clarity. Over time, this helped improve consistency across documentation and led to faster alignment between teams.
Improving the quality of technical communication
Even written communication got an upgrade.
Engineers often handle support tickets directly, but they aren’t always trained in customer-facing language. While their answers were technically correct, the tone could vary or require editing from team leads before being sent.
Cassidy helped streamline that process. Engineers could use it to rewrite messages in a clearer, more polished tone, helping reduce back-and-forth and improving consistency across support communication.
Drafting RFPs Without Burning Hours
On the commercial side, Cassidy helped reduce the time spent on one of the most repetitive tasks: RFPs and security questionnaires.
These forms were time-consuming, even when most answers had already been written in past submissions. Instead of starting from scratch each time, Cassidy generated high-quality first drafts based on previous responses and internal documentation.
“It’s taken out hours of time to give an initial pass at RFP responses. That’s something the commercial teams appreciated.” - Warren Oakes
This gave the team a strong starting point, saved hours per submission, and freed up time to focus on strategy and customer engagement.
Cassidy helped FrankieOne establish an AI-first culture
What started as a single assistant in Slack evolved into a company-wide transformation.
Cassidy didn’t arrive with a mandate or a major launch plan. It showed up as a helpful tool, solved a few real problems quickly, and earned its way into more workflows. Over time, it became a platform for teams across the company to build their own solutions, on their own terms.
What made it work was not just flexibility. It was usability.
“We wanted a tool any person could pick up and start creating workflows with, without needing to involve an engineer.” - Warren Oakes
Adoption spread naturally. As Cassidy became embedded in day-to-day operations, at least one or two people in each department began experimenting with it.
They created custom workflows, shared what they built, and helped their teams get more value from the tool. It did not require formal onboarding programs. It took root through visible results and peer-led momentum.
As adoption deepened, FrankieOne began hosting AI hack days. Teams used the time to explore new automations, test ideas, and build on what had already become part of their workflow.
Though there are many improvements, here are some of the key results so far ...
- 11% of all support tickets fully automated
- Median ticket resolution time dropped from 4.0 days to 2.1 days
- Onboarding diagnostics reduced from 30 minutes to under a minute
The results so far have given the team room to think bigger. With key processes now automated and everyday pressure removed, FrankieOne has set new goals. These include reaching sub-one-day resolution times and expanding automation across more complex workflows.
Cassidy didn’t just help FrankieOne meet their goals. It helped them raise the bar.
“I would recommend Cassidy. We looked at other tools, but none made it this accessible. It’s proven itself across multiple departments, and it’s done it in a way that’s effective and easy to use.” - Warren Oakes
If you are looking to scale your efforts with ease, like FrankieOne, it might be time to try a demo of Cassidy.
Read more success stories
Curious about Cassidy?
Reach out to our success team to see how Cassidy can help your business automate manual tasks and unlock new growth
Trusted by global executives at leading organizations
"Cassidy allows us to be much more proactive in addressing critical customer complaints — automatically escalating issues to a manager who can solve customer problems quickly."
“I never thought I’d be able to automate responding to our RFPs until I found Cassidy. It was incredibly easy to connect my tools, and helps us respond 5x faster to customers so we can close more deals.”
“Cassidy’s real-time context allowed us to finally implement AI automations across our marketing, HR, and sales teams."
"Across the board, our team is obsessed with using Cassidy as a copilot to refine solutions, research industry insights, streamline knowledge transfer across teams and find creative ways to automate the day-to-day monotonous tasks"
"Cassidy makes it easy for us to get our employees exactly the information they need — right when they need it."
"I would recommend Cassidy. We looked at other tools, but none made it this accessible. It’s proven itself across multiple departments, and it’s done it in a way that’s effective and easy to use."
FAQ
Cassidy is enterprise-grade secure. All data is encrypted in transit and at rest, and we’re SOC 2 Type II compliant. Most importantly, your data is never used to train AI models. If your team is using AI, Cassidy is the safest way to do it—trusted by security-conscious teams in finance, healthcare, and government.
Cassidy supports all major large language models, including OpenAI’s GPT-4, Anthropic’s Claude, Google’s Gemini, and more. Each model has unique strengths—Cassidy lets you choose the right one for your use case, or route tasks dynamically with a model-agnostic approach.
Cassidy lets your team build AI-powered automations for almost any task—without writing code or waiting on IT. From AI assistants to complex end-to-end workflows, Cassidy gives you the flexibility to automate whatever your business needs.
Cassidy integrates with hundreds of tools across your stack. That includes CRMs like Salesforce and HubSpot, support platforms like Intercom and Front, data warehouses, internal APIs, and more.
Cassidy is used by startups, mid-market leaders, and Fortune 100 enterprises alike. Whether you’re just starting out or running global operations, Cassidy scales with your business and security requirements.
Most tools give you a chatbot. Cassidy gives you a platform. With deep business context, powerful automations, model flexibility, and enterprise-grade security, Cassidy helps teams actually get work done with AI—not just chat with it.
