News & Events Timeline

We’re excited to share that in October 2023, Allison became significantly smarter by adding conversational context and memory. Instead of treating each message as a fresh input, Allison now remembers what was said earlier in a conversation — making interactions much more natural and coherent.

Why Conversational Context Matters

Traditional chatbots often start each response “from scratch,” with no memory of what came before. That means if you ask a follow‑up question — even one clearly connected to the last — the system might not understand the reference. To overcome this, modern AI research shows that storing and utilising long‑term dialogue memory dramatically improves consistency in extended conversations. Read more in this paper from Cornell University: Recursively Summarizing Enables Long-Term Dialogue Memory in Large Language Models, which describes how summarising and retaining past context help models generate cohesive responses across many turns.

How It Was Built

To give Allison conversational context, we store key parts of the chat history as the conversation unfolds and feed that accumulated memory back into the AI’s understanding of the next input. The steps look something like this:

1. Capture and store ongoing conversation turns
2. Summarise or index important context from those turns
3. Use stored memory to inform future replies

This isn’t just logging every word — it’s about identifying what matters in a conversation and letting Allison reference that history when needed (e.g., for follow‑ups or callbacks to earlier points).

What This Enables

With conversational context and memory:

• Allison can follow multi‑turn exchanges without losing track of the topic.
• She can recall details from earlier in the dialogue, such as preferences or project info.
• Conversations feel more natural and coherent, reducing the need for users to repeat information.

This upgrade moves Allison closer to human‑like conversation, where each message builds on what came before — making her responses smarter, more relevant, and easier to trust.

We’ve upgraded Allison, our AI chatbot, with vector embeddings and Retrieval-Augmented Generation (RAG) — a major improvement in how she understands questions and generates accurate, grounded responses.

What Are Vector Embeddings?

Vector embeddings are numerical representations of text that capture semantic meaning rather than relying on exact keyword matches. Similar ideas are placed close together in a mathematical space, allowing systems like Allison to understand intent even when phrasing changes. You can read more about Word Embeddings on Wikipedia, or dive into the original research like Mikolov et al.’s paper on Efficient Estimation of Word Representations in Vector Space. The Transformer architecture paper “Attention Is All You Need” also explains how embeddings are used in modern AI models.

This approach enables:

• Semantic search instead of keyword search
• Better understanding of natural language
• More relevant and context-aware responses

What Is Retrieval-Augmented Generation (RAG)?

Retrieval-Augmented Generation (RAG) is an AI technique that retrieves relevant information from external sources before generating a response. Instead of relying solely on a model’s internal knowledge, RAG grounds responses in real data using vector embeddings for retrieval. You can learn more on Wikipedia: Retrieval-Augmented Generation or check the seminal paper RAG for Knowledge-Intensive NLP Tasks (Lewis et al.).

This helps ensure responses are:

• More accurate
• More up-to-date
• Less prone to hallucination

Why We Added This to Allison

As Allison takes on more complex conversations, accuracy and context become critical. By combining vector embeddings with RAG, Allison can now:

• Understand meaning instead of relying on keywords
• Retrieve the most relevant information from trusted sources
• Generate responses grounded in real data
• Maintain context across longer, multi-turn conversations

This makes Allison more reliable, more scalable, and far more useful in real-world business scenarios.

What This Unlocks

With vector embeddings and RAG in place, Allison gains a strong foundation for future improvements — including deeper personalization, better memory, and more advanced reasoning across large knowledge bases.

This upgrade moves Allison closer to being a truly helpful assistant — one that understands intent, respects context, and answers with confidence.

We’re excited to share that our Digital Displays feature is now available in a limited release. This marks an early step toward making real-world screens as easy to manage as any website.

Testing Our CMS for Digital Displays

We’ve begun testing our CMS for digital displays, giving businesses a simple way to send content directly to screens in physical locations—no specialized hardware required.

Our CMS can deliver static HTML to off-the-shelf TVs, allowing teams to control what appears on their displays using the same tools they already use to manage web content.

What You Can Do Today

With this early release, businesses can:

• Send static HTML content to standard TVs
• Display menus, promotions, or announcements
• Keep content in sync across multiple screens
• Update displays remotely from a single CMS

What’s Next

This limited release helps us validate performance, reliability, and real-world use cases. As testing continues, we’ll expand access and add more flexibility to support larger, more dynamic display networks.

Digital displays should be simple, scalable, and easy to manage—and this is the first step in getting there.

We’ve started testing our CMS for digital displays, giving businesses a simple way to send content directly to screens in real-world environments.

CMS Testing for Digital Displays

Our CMS can now deliver static HTML to any off-the-shelf TV, making it easy to power digital displays without specialized hardware.

What This Enables

• Send static HTML content to standard TVs
• Display menus, promotions, or announcements
• Sync content across multiple screens
• Update displays remotely from one CMS

This testing phase lays the groundwork for flexible, reliable digital displays that are easy to manage and ready to scale.

Allison now supports multiple languages, allowing businesses to communicate with customers around the world directly on their website.

Multilingual Website Conversations

Allison can understand and respond in different languages, creating a more natural and accessible experience for every visitor.

Key Benefits

• Communicate with global audiences
• Automatic language detection
• 24/7 support in multiple languages
• Consistent responses across all languages

With multilingual support, Allison helps turn your website into a truly global customer experience.

We’re excited to showcase Allison, our conversational chatbot designed to live directly on a customer’s website. This release focuses on testing Allison in a real-world environment to ensure smooth, helpful, and natural interactions with website visitors.

What Is Allison?

Allison is a natural language processing (NLP) chatbot built to handle conversations and customer experiences across multiple platforms and languages. When embedded on a website, Allison provides instant, intelligent responses that help visitors get the information they need—any time of day.

Website Chatbot Capabilities

Real-Time Website Conversations

Allison engages directly with website visitors:

• Embedded seamlessly into customer websites
• Responds to visitor questions in real time
• Works across desktop and mobile devices

24/7 Customer Support

Allison acts as an always-on support assistant:

• Handles frequently asked questions
• Provides instant responses outside business hours
• Reduces the need for live support staff

Appointments & Automated Actions

Allison can take action, not just chat:

• Book or reschedule appointments
• Guide users through common tasks
• Trigger custom workflows based on user intent

Voice Assistant Compatibility

Allison can also power voice-based experiences:

• Build actions for Google Assistant devices
• Maintain consistent conversations across web and voice platforms

Why Test Allison on a Customer Website?

Testing Allison in a live website environment helps ensure:

• Accurate and helpful responses for real visitors
• Smooth conversation flows and user experience
• Readiness for full deployment as a customer service solution

Allison turns your website into an interactive, always-available customer experience.