ROLE
PRODUCT DESIGNER
AT FRONT.COM

TEAM
DEVELOPER PLATFORM

I joined Front as the Product Designer for the Developer Platform squad, as the sole designer on the team, responsible for the integrations ecosystem, specifically Connectors, a key component for enabling advanced automations.

Front is a customer operations platform that combines AI with human support.

My role and key responsibilities

→ Led end-to-end design as the only designer on the squad
→ Designed user flows and added new components to the design system
→ Collaborated closely with developers during implementation and QA
→ Conducted customer interviews alongside the product manager
→ Planned and executed usability tests in Maze

CONTEXT ABOUT Connectors

Connectors allow users to integrate external systems into Front and trigger actions directly from a conversation. For example canceling an order in a CRM or creating a Jira ticket using information detected in a message. Their core value lies in automation, based on specific inputs, the system can query external data and generate personalized automated responses.

Example: A customer asks about order status → the system checks logistics → responds automatically with up-to-date information.

Without active Connectors, the AI has no access to external data, significantly limiting the quality and accuracy of automated resolutions.

Research
understandING Connectors usability

Before the project was defined, I conducted exploratory customer interviews to understand real Connector usage and adoption gaps, which surfaced key issues around discovery, hidden complexity, and administrative friction.

Interview insights summary from initial customer calls about Connectors

The problem

At a company level, Front’s north-star goal was to improve AI-powered resolutions, increasing the quality and relevance of automated answers delivered to customers.

Within that context, our squad’s goal was clear: increase the adoption of Connectors, which are a critical dependency for powering AI resolutions with real, external data. Only 5.2% of eligible customers used connectors weekly.

While we had already gathered multiple insights from customer interviews explaining low awareness and adoption of Connectors, we began the quarter with a focused initiative: improving the discoverability, understanding, and activation of pre-built Connectors (ready-to-use integrations).

Intended user journey, which in reality was highly fragmented and full of friction.

Connectors Home in App Store

The project started with a clear hypothesis:

Users aren’t adopting Connectors because they are hard to discover, hard to understand, and difficult to configure.

The goal was to turn the App Store into the home for Connectors, a clear, guided place where users could discover, understand, and implement integrations with minimal cognitive and technical friction.

RESEARCH

A study conducted in Maze helped identify critical issues.

Most users did not understand what a Connector was. Additionally, when entering the App Store, users didn’t know whether they needed to install an app as a plugin or as a connector. The problem wasn’t only functional, it was conceptual and semantic. As a result, even highly motivated users dropped off before completing an integration.

The study also revealed navigation preferences that shaped the new App Store structure, its search behavior, and its AI-driven discovery.

Maze study results showed users need for better education about Connectors and improved installation and configuration around Connectors.

architecture analysis

I started with a benchmark to understand how platforms like Slack, Zapier, and Notion structure integration marketplaces, specifically how they balance exploration, task-driven discovery, and setup complexity.

I then mapped Front’s existing App Store architecture and identified systemic issues:

→ Competing parallel navigation paths that fragmented discovery
→ An unclear hierarchy that obscured Connector value
→ A lack of guidance for users arriving with a specific task or integration goal

Mapping the existing information architecture helped surface structural issues and inform a more coherent content architecture.

Benchmark helped understand how other platform balance exploration and discovery.

Validating navigation and mental model assumptions

Before defining the final navigation model, me and the PM formulated a set of hypotheses based on early customer interviews and product observations.

These hypotheses focused on how users approach integrations, how they reason about system concepts, and where conceptual friction occurred in the App Store experience.

Reframing our goal

How might we transform Front’s App Store into an intent-driven integration experience that enables teams to discover, understand, and activate Connectors with confidence, reducing conceptual friction, aligning with how users think about outcomes rather than system concepts, and ultimately enabling AI to access external data required for high-quality automated resolutions?

The new search experience surfaces integrations that power automation and AI resolutions earlier in discovery.

Reframing the App Store as a workflow entry point

The previous experience positioned integrations as items to browse. Research showed that users rarely explored the App Store this way, they arrived with a specific goal in mind.

The redesigned entry experience introduces:

→ A larger, primary search entry point focused on apps, workflows, or problems
→ Outcome-oriented categories instead of technical groupings
→ Contextual sections such as AI Essentials to guide discovery 

removing ambiguity at a critical decision moment

When an integration offered multiple implementation models (e.g., Jira Plugin vs Jira Connector), users previously had no clear way to understand the difference.

The redesign introduces:

→ Side-by-side comparison flows directly from search
→ Clear descriptions of when to use each integration type
→ Contextual tags such as automation or AI resolution support

This approach reduces incorrect installations and removes ambiguity at a critical decision moment.