How does Embedded iPaaS Help You Build Integrations?

First, What is an Embedded iPaaS?

“Embedded integration platform as a service” (or Embedded iPaaS) is quite a mouthful. Normally, product categories can be more concise, but Embedded iPaaS actually needs all of these terms to properly describe it.

1. Embedded describes a pattern where all of the integrations and functionality will be encapsulated in your app. No need for your users to navigate to Zapier or n8n to integrate their CRM or messaging with your app.

2. An Integration Platform helps you build product integrations embedded in your app that your users can connect to.

3. Lastly, as a service means that the platform will come with support, monitoring, and new features that you can expect from SaaS products.

SaaS builders use an Embedded iPaaS to offer their users product integrations between their app and integrations, like Slack and Shopify, to connect their users’ 3rd-party data and leverage 3rd-party APIs.

What are Examples of Product Integrations?

Products like Zapier and n8n are internal integration platforms for internal automations. In Zapier, you can trigger workflows whenever a record gets created in your CRM and send a Slack message from your Slack account/app.

Product integrations are user-facing integrations that connect your application with each of your users' unique integrations. If your sales application has a product integration with HubSpot that syncs Contact data, your users will authenticate their HubSpot in your app, and your app can call the HubSpot API to pull CRM data on their behalf.

Here’s an example of product integrations in AppSmith’s product where their users can add agent’s with Jira, Asana, Slack, and other integration tools.

How Does Embedded iPaaS Work?

Embedded iPaaS enables SaaS teams to quickly develop integrations for their users and build use cases with connected data. But how do they embed integration features and work with 3rd-party APIs?

Embedded iPaaS accelerates integration development through two core features:

1. Managing 3rd-party integration authentication with embedded UI

2. Workflows for building integration-level automations that leverage your users’ credentials

Throughout this section, we’ll show you how Embedded iPaaS works with examples from Paragon’s Embedded iPaaS features. Paragon is an integration infrastructure platform that started as an embedded iPaaS.

After years of working with customers on their integration needs, we noticed that workflows are not the perfect solution for some common integration use cases - from high-volume syncs to low-latency actions. This observation prompted Paragon to expand our platform, suited for any integration use case across any product integration.

We’ll start with the two core Embedded iPaaS features and work our way to expanded integration use cases.

Embedded UI for Authentication

Embedded iPaaS starts on your frontend with SDKs that take your users through the OAuth process for their integration. Paragon does this with a frontend SDK. With just a few lines of code, you can implement an embedded popup (Connect Portal) for users to log in to an integration.

await paragon.authenticate(
  // You can find your project ID in the Overview tab of any Integration
  "38b1f170-0c43-4eae-9a04-ab85325d99f7",
  // A signed token with your user's unique ID
  "eyJhbGciOiJSUzI1NiIsInR5cCI6IkpXVCJ9.ey..."
);

paragon.connect("salesforce");

An Embedded iPaaS will then manage authentication, everything from storing your users' access tokens to refreshing tokens.

Your users' integration credentials are stored by the Embedded iPaaS so that workflows and actions that use the integration provider's API will automatically use your users' valid credentials.

Workflows for Integration-Level Automations

Embedded iPaaS has extended workflow builders like Zapier for user-facing product integrations. When you build workflows in an embedded iPaaS like Paragon, you can perform different actions that call the integration providers' APIs (i.e. send a Slack message, query Salesforce records, and pull a Google Drive file).

These integration actions can be used to pull data (pulling Google Drive files for indexing to your vector database) or push data from your app to an integration (updating a Salesforce record with enriched data from your app).

These workflows can be triggered in a few ways:

1. From an event in the integration - a Slack message, a Jira ticket created, a GitHub commit

2. From an event in your app - your app can call one of your embedded iPaaS APIs with a payload

3. CRON triggered - configure a cadence in the embedded iPaaS workflow builder to trigger workflows on a schedule

Workflows are an artifact of embedded iPaaS's roots as traditional iPaaS platforms. While theoretically, workflows should be flexible enough to build any integration feature, we've seen that workflows are not an ideal solution for every integration use case modern SaaS and AI companies need.

Integration Infrastructure Platform, The Evolution of Embedded iPaaS

Workflows are low-code interfaces that allow technical and non-technical audiences alike to build integration logic. Where this low-code interface starts to break down is for use cases with:

• Quick low-latency actions

• High-volume data pipelines

Quick Low-Latency Actions

There are many use cases for these low-latency actions.

1. Tool calling for AI agents

Agents need to call 3rd-party APIs on your users' behalf in real-time.

2. Embedded workflow actions

If you're building a workflow builder, 3rd-party workflow actions like sending a Slack message or creating a Notion page need to be synchronous when you execute workflows for your users.

3. UI components that call an integration API

Your UI components may call an integration API, such as buttons to update a CRM contact or retrieve Notion documents for a file picker.

Embedded iPaaS workflows are great for multi-step durable jobs that don't have low-latency requirements, but for these one-step jobs, calling an API is a more appropriate interface that sends a direct API call to the integration provider, rather than using a workflow engine.

Paragon's ActionKit is the solution to these low-latency integration actions, with a simple API that has 700+ integration actions. These actions come with LLM and human-readable descriptions for agentic tools and workflow steps. These actions aren't just API wrappers, they’re abstracted actions for common use cases for integrations.

{
  "name": "NOTION_GET_PAGE_AS_MARKDOWN",
  "description": "Get a Notion page with contents in Markdown format",
  "parameters": {
    "type": "object",
    "properties": {
      "pageId": {
        "type": "string",
        "description": "Page ID of the page to query"
      },
    "required": ["pageId"],
    "additionalProperties": false
  }
}

High-Volume Data Pipelines

The flexibility of workflows is one of its defining features. But because workflows weren't purposefully designed for any one type of job, there are API products that are better suited for specialized jobs.

One specialized job type that we were seeing over and over from customers were high-volume data ingestion and syncs from their users' Google Drive, SharePoint, Salesforce, etc. Purpose-built APIs like Managed Sync's Sync API and Permissions API were designed to sync hundreds of thousands of files and records in minutes.

Rather than forcing our customers to build standard workflows for each integration with this sync use case, the Managed Sync APIs normalize data across an integration category - syncing Salesforce, HubSpot, Pipedrive, and Dynamics records all use the same API and return a normalized data schema. The Permissions API also syncs permissions data on files and documents in a managed permissions graph that users can query. This is essential for RAG use cases where data not only needs to be ingested and synced, but permissions need to be enforced on retrieval.

Paragon as Your Integration Infrastructure Platform

Paragon was and still is the top choice for Embedded iPaaS. We noticed that the workflows that traditional embedded iPaaS is associated with weren't the perfect solution for some of the hardest integration problems that modern SaaS and AI companies are trying to build for.

ActionKit and Managed Sync are our response to the challenge of solving for every integration use case our customers are building.

1. Quick, low-latency integration actions with ActionKit

2. High-volume data pipelines for ingestion and syncs

3. Workflows for flexible asynchronous jobs

All of our products come with first-class developer experience and core Embedded iPaaS features.

• Custom integrations to cover every integration your users are asking for

• Embedded authentication and management so your developers can build integrations with the highest velocity

• Monitoring and observability so you have visibility on how your customers are using integrations

Wrapping Up

If you've made it this far, you should have a solid understanding of:

• What an embedded iPaaS is

• What product integrations are

• How embedded iPaaS works

• Limitations of embedded iPaaS

• Why integration infrastructure is the evolution of embedded iPaaS

What we've seen is that embedded iPaaS is a great starting point for building product integrations for your users. AI and modern SaaS products have raised the bar for product integrations. Users expect deep integrations and AI features that traditional Embedded iPaaS is no longer optimal for.

Try Paragon to see the difference a full integration infrastructure makes. Book a demo to talk with our team and build deep product integrations for your users.