Case study
How global fintech Jeeves 3x'd their
support capacity with AI agents
With support demand rising almost 20% each month, Jeeves’ customer experience team faced a critical challenge: how to triple capacity without adding headcount.
Backed by $380M in funding from Andreessen Horowitz, CRV, Tencent, and Y Combinator, Jeeves is a global fintech infrastructure company supporting 3,000+ businesses across 25+ countries.
Fintech
Behind Jeeves’ global reach was a small bilingual team managing thousands of monthly tickets, most through live chat.
Ticket volume was projected to triple within a year, while hiring plans remained focused on commercial and product growth. For Daniela Martínez, Global Head of CX, this created an urgent balancing act.
Daniela Martínez, Global Head of CX
Many customer requests were repetitive, from card deliveries and payment confirmations to declined transactions. Agents were spending valuable time switching between Zendesk, Metabase, Datadog, and partner portals.
Daniela’s team needed to protect response times and CSAT, maintain the customer experience Jeeves is known for, and meet company-wide AI adoption goals – all without expanding the team.
Jeeves partnered with Outverse to launch a focused pilot designed to prove value quickly. The first AI agent went live in just two weeks, integrating directly with Zendesk to keep disruption minimal.
Following our embedded agent design model — where Outverse specialists work directly inside the customer’s environment to design, test, and iterate AI policies — we collaborated closely with Daniela’s team to map high-impact workflows and ensure a smooth rollout. This hands-on approach meant Jeeves’ agent was configured with real operational context from day one.
Together, we:
Identified the most valuable automation opportunities through
pre-launch data analysisConfigured routing based on Jeeves’ existing queues, languages,
and out-of-hours rulesTuned escalation sensitivity to minimise unnecessary human hand-offs while keeping safeguards for negative sentiment or complex edge cases
Securely connected to Jeeves’ internal tools to verify customer identities and retrieve live data for contextual responses — without exposing sensitive systems or requiring engineering involvement
Jeeves’ AI agent, Leonora, now resolves the majority of routine tickets across chat and email, in Spanish, Portuguese, and English.
Leonora was designed to reflect Jeeves’ brand and team identity. “Our logo is a lion,” explains Daniela. “So we called her Leonora: a lioness who represents both strength and care.”
Policies are written in natural language, so Daniela’s team can adjust them as business rules evolve.
Every automated action leaves a clear, auditable record in both Zendesk and the Outverse platform. Daniela’s team can see how Leonora reasons, test policy updates safely before deployment, and monitor every workflow in real time, giving full observability across all AI-led operations. That transparency has been key to maintaining trust and compliance as automation scaled.
Outverse gave Daniela’s team the capacity they needed to scale without hiring, freeing them to focus on strategic work and new initiatives.
Increased capacity: Leonora now autonomously handles routine requests, reducing workload
Scale without hiring: Jeeves is growing month on month, and has been able to maintain existing team size due to Outverse absorbing workload.
Redeployment of resource: talent is now able to work on additional internal projects and spend more time on career development.
Multilingual coverage: instant, consistent responses in Spanish, Portuguese, and English.
Audit-ready visibility: every automated interaction is traceable, supporting risk and compliance standards.