Evolving Jaime: From Chatbot to AI-Powered Digital Assistant

Challenge

When First Line Software launched Jaime 1.0 in 2024, the goal was clear: create an AI-powered website assistant that could answer visitor questions and guide them through content. The first version quickly gained traction, supporting both English and several other languages. Users engaged, explored services, and discovered answers on their own. 

But real usage brought deeper insight. Visitors weren’t just looking for fast answers. They were evaluating services, preparing for calls, and comparing vendors. They wanted continuity, personalization, and context. Conversations needed to evolve into guided experiences, with Jaime not just reacting but anticipating. 

That’s when our challenge became clear: how do we move from a chatbot to a true digital companion—one that adapts to a user’s goals, speaks their language, understands where they are in their journey, and helps them take the next step? 

Approach

Building on insights from Jaime 1.0, we assembled a cross-functional product team—product owner, business analyst, developers, and UX leads—focused on making the assistant more intelligent, more helpful, and more aligned with real user behavior.

We started by observing and interviewing users. What were they trying to do? What stalled them? Which interactions moved them forward? The goal wasn’t just to expand features, but to improve the experience, making Jaime feel more like a partner than a bot.

From this, several core design directions emerged. We introduced intent recognition, enabling Jaime to detect user goals based on current page content and conversation history, and dynamically offer context-aware suggestions. We added the ability to favorite important responses for later reference. A clear, persistent chat history structure was introduced to support multi-step journeys. Jaime’s multilingual capabilities were retained and improved, and we began laying the foundation for future voice support.

We also expanded interaction options to manual typing, clicking smart prompts, and preparing for voice input. 

We began building and in May 2025 released Jaime 2.0.

Solution

From Conversations to Companion

Jaime 2.0 reimagines the assistant’s role. It no longer waits passively for questions but instead adapts in real-time to the visitor’s goals.

When a user lands on a service page, Jaime responds by offering targeted action buttons, such as “Summarize this page,” “Show case studies,” or “Talk to an Expert.” It suggests relevant prompts tied to the page’s content and adjusts its behavior based on the user’s recent conversation.

These suggestions aren’t generic. They’re informed by both what the user is interested in now and what they’ve asked before. That shift, powered by intent recognition and contextual awareness, transforms Jaime from a simple Q&A assistant into a meaningful guide.

Multilingual and Voice-Ready

Visitors can chat with Jaime in their own language, and it will respond accordingly. This functionality was already present in Jaime 1.0 but has been expanded for broader coverage and smoother transitions. Additionally, users can speak naturally, dictate texts, interrupt responses, or switch between voice and text mid-conversation, making Jaime more accessible, especially on mobile.

Favorites & Memory

With the introduction of Favorites, visitors can now mark specific responses to revisit later. Combined with a searchable and date-organized chat history, this helps users compare services, gather information over time, and come back prepared with follow-up questions.

In the near future, Jaime can also be enabled to identify and remember user interests and instructions and consider them during further conversations, making them more targeted.

Jaime becomes not just a live help tool, but a persistent research assistant, which is especially valuable in high-consideration journeys.

Seamless UX

We’ve redesigned the experience for flexibility and speed. Jaime is fully embedded within the website, working in light or dark mode, and allowing users to switch between compact and expanded layouts.

Visitors can edit or resend previous messages, rate responses for feedback, and copy answers with a click, all without navigating away or opening a separate window.

Built for Growth

Jaime’s roadmap includes several upcoming features designed to enhance personalization, accessibility, and integration while remaining easy to scale across client environments.

Site-wide smart search is underway, allowing Jaime to more intelligently surface and respond to content across the website.

To support long-term user journeys, sign-up and sign-in options will enable visitors to retain preferences, revisit previous chats, and opt in to follow-up communication, directly from the chat window.

Additional enhancements include scenario dialogs for multi-step guidance (e.g., lead qualification) and multimodal outputs, such as link previews, tables, or visual templates, to make responses more actionable.

Multi-source capabilities are planned to make Jaime a universal tool for querying information from modern tools like Amazon Q Business, Google Vertex AI, and similar solutions, offering dialog-based access to data. 

These multi-source capabilities will also be extended to allow dialog and agent-based multi-step interaction, providing transparent access to dialog/scenario-based services while maintaining all other benefits of Jaime 2.0.

It is worth mentioning that Jaime 2.0 also follows the market demand for MCP integrations and plans to offer this feature in the next releases.

All features are being built within Jaime’s modular framework, making them reusable, configurable, and deployable across client implementations.

Try Jaime Live

Click the chat icon in the lower-right corner of firstlinesoftware.com to see Jaime in action.

To explore how AI-powered assistants like Jaime can elevate your customer experience, connect with us or follow First Line Software on LinkedIn.