Yotpo: Seamless Support Automation via Enjo

— CHALLENGE

Replacing a Failing Tool - Without Missing a Beat

Yotpo faced a critical situation when Halp, the tool their DevOps teams relied on daily, needed to be replaced as quickly as possible. With over 40 Slack channels serving different teams and functions, the stakes were high - any gap in tooling would ripple across internal workflows and customer-facing support.

The team needed a solution that could integrate effortlessly with Slack while preserving their preferred emoji-based actions for ticket creation and management. It also had to support DevOps second-level support, ensuring complex issues were handled efficiently without compromising response times.

But Yotpo's ambitions went further than a like-for-like replacement. Their employees were spending significant time navigating internal tools and systems to find answers to routine questions - deployment statuses, system configurations, access requests, and operational procedures. These repetitive queries drained engineering and DevOps bandwidth, and the answers were often scattered across disconnected internal services. Yotpo needed a way to bring those answers directly to employees in Slack, without requiring them to context-switch across multiple tools or wait for a human to respond.

The objective was clear:

• Replace Halp immediately - Halp could no longer meet Yotpo's needs for customization, reliability, and scale. A fast, zero-disruption transition was critical.
• Establish robust Jira–Slack integration - DevOps teams handling second-level support needed seamless workflows across both tools without channel switching.
• Deploy across 40+ channels in two weeks - Any downtime would directly impact internal operations and customer satisfaction.
• Enable AI-driven resolution for employee queries - Beyond ticket management, Yotpo wanted to use AI Agents to autonomously resolve common employee questions by connecting directly to internal systems.

— SOLUTION

A Two-Phase Approach: Stabilize, Then Automate

To help Yotpo navigate this transition, Enjo implemented a focused, two-phase approach - first stabilizing their ticket operations, then unlocking AI-powered resolution.

Phase 1: Rapid Deployment of Enjo Ticketing

• Enjo Ticketing for Jira - Deployed the ticketing system with native Jira integration, ensuring all support requests were tracked and managed within Yotpo's existing workflow.
• Slack–Jira automation - Connected Slack with Jira to enable automated ticket creation through emoji actions, preserving the workflows teams already knew.
• 40+ channels live in two weeks - Enjo AI Agents were deployed across all channels simultaneously with zero downtime, backed by hands-on training for every team.

Phase 2: AI-Powered Employee Query Resolution with Agent Studio

Once ticketing was stable, Yotpo expanded their use of Enjo to tackle a broader challenge: resolving employee queries autonomously using AI.

Using Enjo Agent Studio, Yotpo connected their AI Agent to their internal MCP (Model Context Protocol) server — giving the agent secure, real-time access to internal systems, APIs, and operational data. This allowed the AI Agent to go beyond answering questions from static knowledge bases. Instead, it could query live systems, retrieve contextual information, and resolve employee requests end-to-end, all within Slack.

With this setup, Yotpo's employees can now ask questions like:

• "What's the status of the latest production deployment?"
• "Who has access to the staging environment?"
• "Can you check if the payment service is healthy?"

The AI Agent resolves these queries instantly by pulling live data from Yotpo's connected systems through the MCP server - no human intervention needed. When the agent's confidence is low or the request requires human judgment, it escalates cleanly into the Enjo Inbox with full context and a recommended next step.

Agent Studio made this possible without requiring Yotpo to write custom code. The visual builder allowed their team to design, test, and deploy AI behaviors and multi-step flows that connect to internal tools through the MCP server - with guardrails to ensure every action is safe, auditable, and policy-compliant.

— BUSINESS IMPACT

Results That Compound

Results

• 100% adoption - Full team adoption within the first month of deployment, demonstrating strong user engagement and system effectiveness across all 40+ channels.
• 30% faster ticket creation - Achieved a 30% reduction in time spent creating and managing Jira tickets compared to manual workflows with Halp.
• Instant employee query resolution - Common employee questions about deployments, system status, and access are now resolved autonomously by the AI Agent via MCP server integration - freeing up DevOps bandwidth for higher-value work.
• Zero-downtime migration - Transitioned from Halp to Enjo across 40+ Slack channels within a strict two-week timeline with no disruption to daily operations.
• Scalable automation foundation - With Agent Studio and MCP server connectivity in place, Yotpo can continuously expand automation coverage - connecting new internal systems and resolving more query types without additional engineering effort.

Why Enjo

Yotpo's experience demonstrates the value of Enjo's approach to support automation: start with immediate, high-impact wins - like replacing a failing tool overnight - and then expand into autonomous resolution powered by AI Agents.

• Resolve issues, not just route them - Enjo AI Agents don't just create tickets. With Agent Studio and MCP server integration, they query live systems and complete requests end-to-end.
• Fast time-to-value - From zero to 40+ live channels in two weeks. From ticket management to autonomous query resolution in the same platform.
• Safe escalation when it matters - When the AI Agent can't resolve a request, it hands off to the human Inbox with full context and suggested next steps - no lost information, no rework.
• Built to expand - Agent Studio's visual builder and MCP server connectivity mean Yotpo can keep adding automation without writing custom code or waiting on engineering cycles.