Fintech Ecosystem Orchestration: Unifying Fragmented Support Channels into a Governed Intercom Stack
A scaling Mexican digital SOFIPO struggled with a fragmented tech stack that erased customer context between voice and digital support. We consolidated Zendesk and Infobip into a unified Intercom environment via Apache NiFi, and integrating Vozy voice bots and Incode KYC kept context clean across handoffs, accelerating loan disbursements.
Executive Summary
Context
A mid-sized Mexican digital SOFIPO was burdened by a fragmented support stack (Zendesk, Infobip, SendGrid) that prevented real-time context sharing and inflated operational costs during a period of rapid customer growth.
What We Built
We engineered a unified fintech support ecosystem centered on Intercom, integrating third-party voice (Vozy), identity (Incode), and messaging gateways (Airmovil) into a single governed workspace.
Tech Stack
- Intercom (Chat/Bots/Ticketing)
- Vozy (Voice Bots)
- Incode (KYC Video)
- Apache NiFi (Middleware)
Not a fit for organizations that rely purely on voice-only support or those without the technical capacity to manage API integrations between specialized fintech microservices.
The Challenge
A customer might start an interaction with a Vozy voice bot to inquire about a loan. Switching to digital chat or video verification through Incode lost all of that previous context, forcing the customer to restart authentication and the inquiry, sometimes more than once. That context fragmentation drove high drop-off rates in the lending funnel. The bank was also paying for redundant seat licenses across Zendesk and Infobip, with no single source of truth for customer interaction history. The technical team also had to comply with Mexico's federal data protection law (LFPDPPP) while syncing sensitive financial data from a legacy core into this multi-vendor support stack.
Our Approach
Apache NiFi bridged the legacy core and the new Intercom-centric ecosystem in a stack-consolidation strategy where every tool referenced the same customer ID and compliance status. We had to choose between a complex, long-term sync for Zendesk and a hard cutover, and we picked the hard cutover: decommissioning Zendesk and moving straight to Intercom's unified helpdesk forced a single workflow for all agents. To solve the voice-to-chat gap, we engineered a custom handoff protocol between Vozy and Intercom, so the voice bot's intent data gets pushed into Intercom as a pre-filled conversation before the user even types a first message. We also embedded Incode's video verification directly into the Intercom messenger interface.
This engagement is part of our systems integration work for finance & insurance organisations.
Impact
30% Reduction in Agent Redirection
Unified Customer Context
Voice, chat, and video interactions consolidated into Intercom, so customers no longer had to repeat information during channel switches, and that zero-friction handoff directly improved the completion rate for complex loan applications.
Decommissioning of Legacy Tech Debt
Replacing Zendesk and Infobip removed redundant licensing costs. Consolidating the stack into a unified environment reduced the technical management burden on the bank’s internal DevOps team.
Secure KYC Video Integration
We built an API bridge that captures intent and entity data from Vozy voice interactions. When a user is escalated to digital chat, this data is pushed into Intercom as custom attributes, so the agent or bot immediately knows the previous voice-based context.
To facilitate secure identity verification, we embedded Incode's KYC video workflows into the Intercom Messenger, enabling one-click KYC within a support thread where the verification results are automatically logged against the Intercom user record.
Apache NiFi serves as the fintech router, pulling real-time customer data from the legacy banking core and distributing it to Intercom and Vozy. So a Do Not Contact flag or a balance change reflects across all communication channels within seconds.
We shifted the support metrics from "Ticket Response Time" to "Automated Resolution Rate." By using Intercom’s Resolution Bots, we established a cost-per-resolution model that aligns vendor expenses with successful customer outcomes.
A customer who started with a Vozy voice bot lost all context after switching to chat or Incode video verification. They had to restart authentication and the inquiry, sometimes more than once. Now Apache NiFi gives every tool the same customer ID and compliance status. The voice bot's intent data lands in Intercom as a pre-filled conversation before the customer types a first message.
FAQ
We implement attribute tokenization at the Apache NiFi level. Instead of passing raw PII (personally identifiable information), the middleware passes unique tokens that let each system reference the customer's core record securely. Sensitive data only decrypts within the specific environment where it's needed for a verified interaction.
For high-growth fintechs, ticketing systems often create a waiting-room experience that kills conversion. Moving to an Intercom-led stack allows for real-time resolution and proactive marketing triggers, turning support into a growth channel where a simple inquiry about a card can convert directly into a loan disbursement opportunity.
