Driving Multi-API Adoption

Overview

  • Project Name: Multi-Unit Enterprise Agent Deployment Use Case Guide (Boomi Agent Garden APIs)

  • Role: Technical Writer and API Documentation Architect

  • Objective: I designed and authored a narrative-driven use case guide to help developers and enterprise architects understand the distinct types of APIs available in Boomi Agent Garden (Design vs. Execute) and choose the right integration pattern for multi-unit rollouts.

  • Key Outcome: Increased feature awareness and API adoption by framing abstract technical capabilities within a relatable, multi-persona enterprise scenario (financial services rollout across regional branches and HQ).

Problem

  • The Feature Education Gap: Boomi’s Agent Garden offers powerful APIs spanning two distinct paradigms (Design APIs for management/CI-CD and Execute APIs for runtime interaction), alongside embeddable widgets (EmbedKit). However, developers often struggled to distinguish when to use which API.

  • The “One-Size-Fits-All” Fallacy: Users frequently tried to force a single integration method (like custom building an entire chat UI) for use cases that were better suited for rapid widgets or automated CI/CD pipelines.

  • The Documentation Hurdle: Technical references list endpoints and payloads, but they don’t explain architectural intent or how different company personas (DevOps, regional staff, enterprise developers) can leverage different APIs simultaneously.

Target audience

  • Primary: Enterprise architects, engineering managers, and technical leads planning large-scale AI agent rollouts across multiple business units or branches.

  • Secondary: Frontend and integration developers looking to choose the correct SDK, API endpoint, or CDN embedding strategy for their applications.

My solution and technical process

To educate users on the full breadth of the API ecosystem, I structured the guide around a concrete, multi-constraint enterprise narrative:

  • The Financial Services Scenario: I introduced a fictional global firm rolling out a support-triage AI agent across 12 regional branches and a central HQ—each with wildly different technical constraints (zero engineering staff vs. deep CI/CD and custom CRM needs).

  • The “Choose Your Path” Information Architecture: I categorized the platform’s capabilities into three clear, distinct technical paths:

    1. Path 1 (Execute API): For deep integration, custom chat UIs (React/Vue), session lifecycle control, and CRM/data warehouse logging.

    2. Path 2 (Design API + CI/CD): For DevOps teams treating agent configuration as code, automating multi-environment promotions via Git and version-controlled snapshots.

    3. Path 3 (EmbedKit CDN): For resource-constrained regional branches needing rapid, one-line widget deployment without build tools.

  • The Hybrid Approach Strategy: I demonstrated how mature enterprises don’t have to choose just one—showing how organizations can combine EmbedKit for public-facing sites with Execute APIs for internal tooling, all managed under the hood by Design APIs.

Tools and technologies used

  • APIs and Protocols: Boomi Agent Garden Design APIs, Execute APIs, Server-Sent Events (SSE) / JSON Polling

  • Ecosystem and Tooling: Agentstudio, EmbedKit CDN, Git / CI/CD pipelines, custom JavaScript/React UI frameworks

  • Documentation Formats: Markdown, interactive documentation hierarchies, architectural use-case guides

Results and impact

  • Drove Cross-Product API Adoption: Successfully encouraged developers to explore beyond basic chat UIs and adopt advanced automation tools (Design APIs and CI/CD integration).

  • Clarified Product Boundaries: Eliminated confusion between runtime execution and lifecycle management by mapping specific technical needs directly to the correct API endpoint.

  • Enhanced Strategic Planning: Provided enterprise architects with a blueprint for scaling AI agents across complex organizational structures without friction.