How Vanquish Trader automated 95% of support tickets and reclaimed 60% of its founders' time

"In options trading, a delayed payout or an unclear rule can cost a trader real money," says Trevor Feldmann, co-founder of Vanquish Trader. "Support isn't a cost center for us. It's part of the product."

For most of Vanquish's growth, co-founders Trevor Feldmann and Max Mastbaum were in the support queue themselves, alongside a 12-person team handling thousands of tickets a month through Intercom. The hands-on approach had built real trust with their trader base, but it was costing each of the founders roughly 60% of their week.

"We were the bottleneck in our own business," Trevor says. "Every hour in tickets was an hour we weren't spending on product, partnerships, or the next phase of the company."

They had already tested Intercom's native AI, Fin, and hit a wall. Fin couldn't handle the complexity that Vanquish needed or take actions on the users' behalf. Trevor explains, "We attempted it with Intercom with Fin and Fin was giving very frequently inaccurate responses. It was offering users services we didn't have. Even if we would train on the help center and try to give it guardrails, it was very very frequently wrong. It was slow and the wording of Fin's responses sounded very AI." Duckie on the other hand is able to handle sensitive information extremely securely, never hallucinates, and is only noticed as AI twice out of a thousand times.

What Vanquish needed was an AI that behaved less like a chatbot and more like a trained support rep, one that could pull up a trader's account, check their activity, tell them exactly what went wrong, and handle account cancellations or upgrades.

"Duckie was the first thing we saw that actually investigates," says Max Mastbaum, co-founder of Vanquish Trader. "It doesn't guess from a help doc. It opens the account, reads the activity, and works the ticket the same way one of our reps would."

Just as important was the onboarding. Vanquish was live inside their existing Intercom workflow within a week, with no integration project, no runbook-writing marathon, and no in-house AI team to staff up. "The Duckie team got us connected, the agent trained on our ticket history, and we were reviewing its work by the end of week one," Max adds. Trevor also explains that they use the playground daily to quickly improve and train their agents, turning what used to take hours with other AI agents, or months with human agents, into a matter of minutes.

Duckie is now handling 95% of Vanquish's support tickets with extremely high accuracy, transforming how the founders operate day-to-day. "Me and my partner would spend, I don't know if Joel ever told you, we would probably spend 80% of our day in support because it was like uncontainable," Trevor says.

"I could confidently walk away from support for probably five days right now and not need to really worry about looking at it. Which is something that was never possible for us." When a ticket does get escalated, the team hops into the playground, identifies what was answered incorrectly, and has it corrected within a minute or two for the next customer. It's essentially talking to customer support to fix your customer support in the fastest way possible with a lot of accuracy, and as the owner of a business, that goes an extremely long way.

That confidence compounds in two directions. Duckie flags its own gaps and proposes improvements the team approves, which means the system gets sharper week over week without engineering lift on Vanquish's side. And every escalated ticket arrives pre-investigated, so the human rep isn't starting from zero.

Vanquish's agent now handles the full range of trader and admin workflows. Traders asking about payout status get a specific answer tied to their account, often explaining that an open position is holding things up per Vanquish's rules, and what to do next, rather than a link to a FAQ. And if the human wants Duckie to address the issue for them, it actually does that. Questions about account eligibility, rule interpretations, and trade reviews get worked the same way: account opened, activity reviewed, answer delivered.

Because Vanquish handles sensitive financial data, balances, payment history, subscription details, the team stress-tests Duckie daily, including attempts to extract information about high-profile influencer accounts whose account numbers are publicly known in the trading community. Just as critical are the guardrails Vanquish has built around negotiation: in an industry where traders routinely try to bargain their way to refunds, free upgrades, or rule exceptions, Duckie holds the line and explains policy clearly without ever offering concessions it shouldn't. "Other AI we tested would cave and start offering free things the moment a user pushed back," Trevor says. "Duckie doesn't negotiate, it educates, and that alone has changed how our customers interact with support."

For the Vanquish team, that investigative depth is the difference between deflection and resolution. "A bot that can't look at the account is just a smarter search bar," Trevor says. "Our customers need the answer, not the article."

Within the first week of going live, Duckie was autonomously resolving 67% of incoming tickets, handling over 1,000 tickets without human touch. By week three, that resolution rate had climbed to 95%. Vanquish has since reduced its support team by 50%, reallocating those resources into growth functions, and the founders have reclaimed roughly 60% of their time. Trevor and Max now review a handful of escalated tickets a day and spend the rest of their week on product and partnerships.

"We're growing insanely fast and support is absolutely crucial to our business," Max says. "In prop trading, when customers go on tilts trading options, they get furious. Support needs to be actually helpful and fast. The fact that we're able to offer very, very fast high quality support has gone an extremely long way, we're even getting Trustpilot reviews specifically calling out how fast and beneficial our customer support is, which in our industry is dire. We're starting to build a reputation for having such good customer support because of using Duckie."

Most AI support tools plateau because they can't actually do anything. They search, they summarize, they deflect. If your tickets require looking at a customer's account, and for most real businesses, they do, you need an AI that can investigate, not just respond.

Beyond handling 95% of Vanquish's support volume, Duckie makes implementing and updating support agents remarkably simple.

The transformation has been complete. After testing the big names in AI support, Vanquish finally found what actually works. "I have a competitive advantage right now having you guys," Trevor admits when asked if he'd recommend Duckie to other prop firms. "But yeah, I would one hundred percent recommend it to them. No exaggeration, we have people leaving reviews applauding how beneficial our support is. Which we did not get beforehand."

Vanquish is already exploring where else Duckie's investigative approach can apply across the business. Ticket volume continues to grow every month, and for the first time in the company's history, growth on the support side doesn't mean growth in headcount.

"We're just getting started with Duckie," Trevor says. "As we keep growing and ticket volume goes up, we're not worried about scaling support anymore. Duckie grows with us. We're already thinking about where else we can apply it across the business."