Key takeaways
Faire's risk operations team is integral to their business growth: every new product needs new risk logic, and existing logic has to keep pace with fraud. Until recently, both waited on engineering and data science capacity. Since their operations team took ownership of risk logic on Taktile, Faire has seen:
- KYB/KYC flows live in 3 weeks, about 3x faster than expected.
- Same-day turnaround for reactive logic fixes from diagnosis to deployment.
- 1–2 week cycles for fully backtested broader logic changes, shipped on demand instead of in a bundle.
"Taktile has enabled the operations team to take on the workflows that would otherwise involve a lot of heavy engineering and data science resourcing. And the AI Copilot has really supercharged our ability to own that more end-to-end."
Meet Faire: The global wholesale platform powering independent retail
Faire is where local shops buy what they sell. On the platform, over 100,000 brands supply wholesale to hundreds of thousands of independent retailers worldwide. Faire has introduced 10 million connections between brands and shops so far, all in service of its mission to strengthen the unique character of local communities.
"Faire is the tool that will help a small business grow," Therese says, "and we can't exist as a marketplace if it's an unsafe place to transact." As she puts it: "Risk in general is dynamic. We have to constantly be iterating on our logic."
The opportunity: Evolving without leaning on engineering
Until recently, iterating often meant waiting on other teams, and that time really added up in two places.
The first was the fraud logic. Fraud patterns change constantly, and naturally, the risk team is the first to see those changes in the cases it reviews every day. As Faire rolled out new products for brands and retailers, even rules that backtested cleanly a few months ago needed recalibrating. Updating them, however, depended on other teams:
"The operations team can notice pretty quickly that this logic isn't holding up. The harder part is what happens next: fixing it means looping in data science and engineering, having them scope the change on their end, and then waiting for it to land a slot on the roadmap," says Therese.
The second place was the product roadmap itself. Faire was preparing to launch a new product which lets retailers pay for their orders on one consolidated monthly balance, backed by a partner bank. Faire needed a robust KYB/KYC process that met the partner bank's requirements and ensured the right customers were approved quickly – so their business and their marketplace stayed safe. Based on the engineering team’s estimate, building the end-to-end KYB/KYC process internally would have delayed the launch by another 2 months.
Instead of asking for more engineering capacity, the team pitched a different approach:
"Engineering resourcing is limited, data science resourcing is limited. Operations is so deeply ingrained in all of these processes: why can't we take on more ownership? We just need a tool that's a little more plug and play. And Taktile was that solution for us."
The innovation: An onboarding journey built and run by the experts
Today, Faire's KYC/KYB decision flows run on Taktile, built, back-tested, and maintained by the risk operations team, with minimal engineering support. These flows also became the team's first use of Case Manager, Taktile's built-in queue for manual reviews, sparing them from extending their legacy case system. In Therese's words: "The KYB/KYC flows would not have launched as quickly as they did without Taktile, for sure."
AI also supports both sides of the team's work: building flows and reviewing cases. On the building side, the team uses AI Copilot for error handling and for writing the code inside flow steps, so they can spend their time iterating on the logic rather than adjusting the syntax. On the review side, Case Insights gives the team AI-powered analytics on their case review operation: the team builds its own dashboards and asks operational questions in plain language, such as which queues are growing and what is driving review volume.
Prior to Taktile, these questions were relayed in a data request to another team. For both layers, AI speeds up the experts' work and still leaves the decisions to them.
"I wanted to see where customers drop off after we reach out to them, so I asked Copilot in Case Insights for a chart of manual review cases per week, split by whether we'd asked for more information and whether they got back to us. I've built that same dashboard in our internal tools before. This was much faster."
Ownership doesn’t end when a rule ships either. A dedicated group of spot checkers reviews live decisions on a continual cycle because, as Therese explains, "trends change over time, fraud vectors change over time." When a rule needs to be recalibrated, the spot checkers flag it and the team makes the change in Taktile.
The outcome: Speed to market, safety at scale, and capacity to grow
By putting risk logic in the hands of its risk operations team, Faire gained three things its growth depends on: new products that launch on business timelines, a marketplace that stays safe as fraud changes, and more capacity from the same team.
- Faire’s KYB/KYC flows went live in 3 weeks, about 3x faster than expected. A new revenue product reached the market about 2 months sooner, while engineering stayed focused on the core product and operations handled the build.
- The team has same-day turnaround for reactive logic fixes, from spotting the change to deployment. When a rule misfires in production, the team can diagnose, patch, and ship necessary adjustments right away. "Now, I go in and make the change," Therese says, "and it takes me seconds."
- Fully backtested broader logic changes can be shipped on demand in just 1–2 weeks. Backtesting against historical data now takes a few days to a week, followed by review and sign-off. Testing used to be slow and manual, so changes were batched; now each one ships when it's ready, and the hours once spent on manual tests belong to operations again.
That freed-up capacity now goes toward improving the platform.
"The team now has time to do more project work that contributes to the overall product and customer experience. Those projects are really meaningful, but they need time. Now we can focus on that higher-ROI work."
The vision: A playbook for the whole company
Company-wide understanding was a condition from the start. The setup had to work for every team, from risk operations and customer experience (CX) to engineering and data science, as Therese explains: "This has to be digestible and usable by everyone at Faire: risk operations, CX, engineering, data science. Everyone needs to be able to understand it, hop in and make changes. And that's what Taktile showed us."
What Therese is most excited about is the playbook: an internal guide, with Taktile at its core, for rolling operations-led risk logic out to the rest of Faire. The team's first experiments focus on underwriting, extending the approach to a brand-new workflow that needs "little to no input from engineering."
Her advice to other risk teams weighing the same move: "Be ready to learn, and don't be afraid to ask a million questions. Just get in, start building. It might be messy, but you're really going to learn a lot really fast."
Messy or not, the result at Faire is a team that can fix the problems it sees, and a marketplace that remains a safe place for local retail to thrive.
