A comprehensive overview of the Ada Web SDK

Let’s be honest, a generic, one-size-fits-all chatbot often does more harm than good. To actually create a great customer experience, you need an AI assistant that feels like it belongs on your website, knows who the user is, and acts the way you want it to. For teams with developers on hand, a Web Software Development Kit (SDK) is usually the tool for this kind of deep customization.

This article will walk you through the Ada Web SDK, explaining what it is, what it can do, and what it asks of your team. We’ll look at how it gives you fine-grained control, but also dig into the real-world trade-offs of an SDK-first approach, especially when it comes to speed and developer time. It’s powerful, sure, but it can also become a real bottleneck. We’ll also show you a more modern, integration-focused alternative designed to get you results in minutes, not months.

What is the Ada Web SDK?

Think of a Web SDK as a developer’s toolkit. It’s a package of pre-written code that lets a developer plug a third-party app, like a chatbot, directly into their own website and control it with their own code. So, instead of just pasting a simple widget onto your site, an SDK gives you the building blocks to weave that widget into the very fabric of your user experience.

The Ada Web SDK is a JavaScript library that lets your developers embed and manage Ada’s AI agent. According to Ada’s documentation, it’s built for "flexibility and customization." In plain English, this means a developer adds a script to your website’s code and then writes more JavaScript to tell the chatbot how to behave. This code-first setup is what gives you that deep level of control over how everything looks and feels.

Key features of the Ada Web SDK

The Ada Web SDK gives developers a set of levers to pull, letting them wire the chatbot directly into how a user interacts with the website. This control is all managed through code, which is great for customization if you have the engineering hours to burn.

Personalizing the user experience

Right away, developers can use a global settings object ("adaSettings", for the curious) to control the bot’s appearance and how it first greets a user. This is a big step up from a basic plug-and-play widget.

Based on Ada’s API reference, here are a few key things a developer can set up:

• "greeting": You can create custom welcome messages for different pages. Someone on your pricing page might see a totally different greeting than someone browsing a specific product.

• "language": The bot’s language can be set automatically based on the user’s browser settings or profile, which is a nice touch for international users.

• "parentElement": Instead of the default chat bubble floating in the corner, you can embed the chat window directly into a specific part of your webpage, like a sidebar.

• "metaFields": This one is super useful. You can pass customer data you already have, like their name, email, or account ID, straight into the conversation. This lets the bot offer a personalized "Hi, Sarah!" experience from the very first message.

Controlling the chatbot with programmatic actions

Beyond the initial setup, the SDK lets you control the bot in real time based on what a user does on your site. This is where the real power of a code-based approach shows up, but it’s also where things start to get complicated. A good example is how a tool like SendSafely handles file uploads; it needs custom JavaScript to listen for an upload and then call specific Ada SDK functions to drop the file link into the chat.

Here are a few of the actions developers can use:

• "adaEmbed.toggle()": Open or close the chat window from any button or link on your site, not just the default chat icon.

• "adaEmbed.setLanguage()": Change the chat language on the fly if a user requests it, without losing the conversation history.

• "adaEmbed.triggerAnswer()": Force the bot to kick off a specific conversation. This is perfect for proactively starting a workflow when a user clicks a certain button, like "I need help with my refund."

Listening for events

The SDK also has an event subscription feature ("subscribeEvent") that lets your website "listen" for things happening inside the chat window. This is for more advanced stuff, like sending chat data to your analytics platform or even changing your website’s interface based on what’s happening in the conversation.

You can subscribe to events like:

• "ada:agent:joined": Get a heads-up the moment a human agent joins the chat, which could trigger a notification on your site.

• "ada:conversation:message": Fire an event every time a new message comes in, which is handy for logging or analytics.

• "ada:csat_submitted": Grab a customer’s satisfaction score the second they submit it and send it over to your data warehouse or CRM.

The drawbacks of the developer-first approach

While the Ada Web SDK offers a ton of control, it doesn’t come free. This developer-heavy model can create a lot of friction, slow down your support team, and ultimately delay the benefits you hope to get from AI.

Why everything depends on developers

Every single customization we just talked about, from a simple greeting to a complex conversation flow, requires a developer to write, test, and ship JavaScript. This creates a massive bottleneck. Your support team might have a brilliant idea for a new proactive message, but they can’t just launch it. They have to file a ticket with engineering, wait for it to get prioritized, and then wait for the next deployment cycle.

What should be a quick, five-minute change turns into a multi-day (or multi-week) project. It completely kills the support team’s ability to be agile and respond to customer needs quickly.

The slow start and hidden costs

The time it takes for a developer to implement and test all these SDK customizations adds up. This is a huge difference from platforms designed to be used right out of the box. Modern AI tools like eesel AI are built to be radically simple. You can connect your help desk, point the AI to your knowledge sources, and have it working in minutes. The entire setup is handled in a dashboard that makes sense, putting the support team in the driver’s seat without them ever having to ask a developer for help.

No easy way to test the AI

Testing code is one challenge, but testing the quality of an AI’s answers is something else entirely. With a setup that lives in the codebase, a support manager has no simple way to simulate how the AI will handle thousands of real-world questions before it’s live. You’re left hoping the logic works when faced with messy, unpredictable customer queries.

This is a problem that platforms like eesel AI solve with built-in simulation tools. A support manager can test their AI against thousands of historical tickets in a completely safe environment. You get a solid forecast of your resolution rate, see exactly how the AI would have answered, and spot gaps in your knowledge base, all from a simple dashboard, with zero risk.

A better alternative: An integration-first platform

The biggest limitation of an SDK is that it only handles the chat window. A truly helpful AI support system needs to be connected to all your company knowledge and business tools. And it needs to be managed by the people who know your customers best: your support team.

Unify knowledge without writing code

An SDK doesn’t do anything to help you with the most important part of AI, the knowledge that fuels it. That’s a whole separate, often manual, process. In contrast, eesel AI is built around seamless, one-click integrations. You can instantly pull in knowledge from all the places your team already works, like Confluence, Google Docs, Notion, and even past ticket history from help desks like Zendesk or Freshdesk. This lets you build a powerful, context-aware AI from day one, without a complicated setup.

A showcase of eesel AI's one-click integrations with platforms like Zendesk, Freshdesk, and Notion, which avoids the limitations of the Ada Web SDK.

Build custom workflows with a visual tool

Instead of asking developers to code every custom action, eesel AI gives you a no-code workflow builder. A support manager can easily define the AI’s personality and tone of voice with a simple prompt. They can then set up custom actions, like looking up an order in Shopify or escalating a ticket in Jira Service Management, through a visual, click-and-build interface. This puts the power to build, test, and improve right back where it belongs: with the support team.

eesel AI's no-code workflow builder, an easier alternative to the code-based actions of the Ada Web SDK.

Transparent pricing you can actually understand

Enterprise tools that lean on developer-heavy SDKs, like Ada, often hide their pricing behind a "Book a Demo" wall. This usually means complicated contracts, long sales cycles, and costs that can be hard to predict.

eesel AI keeps things simple. Our pricing is public, with flexible monthly plans you can start on your own. We don’t charge you per resolution, so you won’t get a surprise bill after a busy month. You can start small, prove the value, and scale up as your team grows.

Feature	Ada Web SDK Approach	eesel AI Approach
Setup Time	Days or weeks (and you need a developer)	Minutes (genuinely self-serve)
Custom Actions	A developer has to code it in JavaScript	Click-and-build workflows, no code needed
Knowledge Sources	Handled separately inside Ada’s platform	100+ one-click integrations (docs, tickets, etc.)
Pre-launch Testing	Manual testing by developers	Automated simulation on your historical tickets
Pricing Model	Enterprise (you have to talk to sales)	Transparent, predictable monthly/annual plans

Is the Ada Web SDK the right tool for your team?

The Ada Web SDK is a solid tool for companies with dedicated developers who need to control every pixel and every interaction of their chat widget. If you have an engineering team ready to build and maintain a custom solution, it gives you all the flexibility you could ask for.

But for most support teams, the goal is to provide better, faster support without creating more work for other departments. They need tools that are fast, agile, and empowering. eesel AI was designed for that reality. It’s the platform that gives you the benefits of advanced AI and deep integrations, without trapping your great ideas in an engineering backlog. You get the power of a custom solution with the speed and simplicity of a tool you can manage yourself.

See what a self-serve AI platform can do

Ready to automate support without waiting on developers? Connect your sources and see how eesel AI can start resolving tickets in minutes. Try it free today.