A New Era for Shopping: The Rise of AI-powered Conversational Commerce

Get ready for the ride: online shopping is about to be completely upended by AI.

Over the past few years, the scale and rate of AI development has been astounding. Also, the release of ChatGPT opened the door for millions of people to experience the power of AI first-hand. This has led to a rapid rise in companies cashing in on the hype – everyone wants to be a leader in AI innovation and dreams of magically co-creating content and dynamic conversations with their users. So the current reality where every company slaps on a bland chatbot devoid of customer context and character is really disappointing.

At Algolia, we know that our customers sweat the details for the home screens of their apps – after all, they’re the digital front doors for their businesses. They’re carefully curated with findings after customer research, refined and polished through numerous design iterations, and built using end-user profile information to keep content relevant and interesting.

Replacing this digital front door with a blank chatbot without any context strips away all the carefully curated relevance and delight. We’re at risk of turning a familiar, simple process of browsing an app into something less human, just for the sake of technology.

But what if this new technology was more attuned to the habits of consumers and didn’t require them to learn new ways to do familiar tasks? What if we empowered them with the power of conversational and generative AI in the behaviors they’re already experts at? To do just that, we’re excited to announce a new framework that brings the power of Conversational AI and Generative AI to your Search and Discovery user experiences. Now, you will be able to strategically deploy AI-powered features that enhance your customers’ journey, inspiring and delighting them, guiding them to find exactly what they need, and, without having to teach them new metaphors or behaviors to use an app.

As we started developing our AI framework, we spoke to many customers and technology experts alike to truly understand what they need. Here’s what we learned:

• No Innovation for the Sake of Innovation: The hype and prospects of AI are exciting, and the speed and rate of change of development are pretty astounding. However, innovation for innovation’s sake is insufficient for Algolia customers – these features should enhance UX and connect business goals to their consumers’ needs. A blank chatbot that lives independently of the carefully curated user journey doesn’t fit these goals and often conflicts directly with them.

• Trust and Safety are Paramount: In spite of the breathtaking rate of AI development, it’s still early days. Many of us have seen AI products released to large audiences with poor vetting with the effect of generating misleading, incorrect, or even harmful information. This is a direct function of releasing products too quickly to market without a thorough understanding of the end-user and what they’re trying to achieve.

“Organizations that don’t manage AI risk are much more likely to experience negative AI outcomes and breaches. Models won’t perform as intended, and there will be security and privacy failures, financial and reputational loss, and harm to individuals. AI that is carried out wrongly can also cause organizations to make poor business decisions.” ~ Avivah Litan, Distinguished VP Analyst,  Gartner Research

Algolia takes trust and safety very seriously, and our customers expect nothing less. Our Conversational and Generative AI features are designed with stringent guardrails that ensure trust and safety for our customers and their end-users in such a way that it enhances the user experience further.

• AI needs to benefit from inclusive design: Cutting-edge technologies give us glimpses of what the future may look like. In order to make them valuable to everyday life, they need to be approachable and usable by a wide variety of people. By using a familiar chat/messaging interface, OpenAI made ChatGPT and its Large Language Models (LLMs) accessible and meaningful to anyone – as opposed to forcing developers who are able to comb through GitHub to install and test an LLM. Today, anyone can login and create queries. In a similar way, Algolia’s AI framework will build on familiar heuristics to enhance the end-user journey, which means less time spent learning how the AI features work and more time spent on tasks that matter to customers.

This approach to inclusivity is important for Algolia customers too – they’ve spent time, energy, and attention in building and optimizing the best Search and Discovery experience for their users. Our framework is designed to be installed easily and with minimal changes to the customer journey – resulting in less time spent setting up a framework.

So what’s in the Algolia conversational AI framework?

Our customers’ needs are wide and varied, which means a one-size-fits-all approach to AI is insufficient. The Algolia AI framework will allow Algolia customers to compose and deliver the power of AI based on specific problems and challenges they see in their Search & Discovery journey. To do this, the framework will include two important building blocks – AI Actions and AI Assists:

AI Actions: AI Actions are UI controls that allow end-users to invoke the magic of AI in their experience. For example, a button next to a carousel allows the end user to generate a list of related items. Or, a link next to a shirt in a product detail page can bring up lookbooks from a curated set of blog posts – all summarized by AI.

Actions are approachable (for example, a button, a widget, or a link) and aren’t very different from the kinds of controls that users typically expect on a website and app. They are humanistic by design and seek to add more context, give more support, or round the edges of the task a user is trying to complete. Using an AI Action starts the AI-powered conversation or generates relevant, contextually valuable content.

AI Assists: AI Assists are event and context-driven prompts that allow you to start an AI-powered conversation with your consumers. For example, starting a conversation with a customer if their search query is vague by asking follow-up questions in order to refine the query to garner the best, most relevant result.

AI Assists enter a conversation thoughtfully and are opt-in. This means you are not forced to interact with a blank chatbot without context – assists usually come with an understanding of what the user is trying to do.

These controls are only possible because Algolia is in a unique position to understand an end-users history with an app, while also understanding their intent in real-time. Together, these can lead to some powerful, magical AI-powered experiences.

Some of the ways customers are using our AI framework

Rescuing consumers from a failing “search and discovery” session: A user who is unable to find the perfect gift for their partner at a popular apparel site may be greeted by an AI Assist that guides them to the perfect gift. The AI Assist recognizes the context (x+ browse sessions, y+ search queries, 0 clicks) and generates a recap message for the user with prompts and suggestions for what to buy.

Building dynamic, contextual shopping lists: When a consumer is looking to make a set of purchases (say for example, a new home theater that has a number of individual items that make the theater come together) and adds the first item to their cart, AI Action can provide a contextually intelligent input to them. In this case, an item such as “Build out the ideal home theater” can be shown on the UI. These actions can show related items that complete the purchase, generate AI-powered content, share links to curated blogs and sources, etc.

Why was this result recommended to me? Helping end-users understand why a search result or recommendation is important to build trust in the ability to surface the best suggestions for them. It also gives them an understanding of how to make better search queries. An AI Action popover next to a recommendation carousel gives an AI-generated summary of the contextual reasons that were responsible for this recommendation. For example, “We chose this result of a Kale Salad because of your query ‘lunch foods’ and your historical preference of ‘organic only’”.

We’re excited to share more!

While these explorations are incredibly promising, they are just the tip of the iceberg for the AI revolution. We’re excited to expose more use cases with this Conversational AI framework, either as programmable primitives or as ready-to-use patterns that customers can simply adopt. We expect this framework to interact with every part of Algolia’s portfolio. Just like InstantSearch and Autocomplete, this will be a valuable addition to your user experience toolkit. To help shape development and get early access, join us by signing up for our waitlist.

Sign up for our waitlist to join our first cohort of customers.