How to boost your website conversion rate with a predictive UX based on explainable AI?



Tous les articles


As many web-based marketplaces, banks rely on their website and mobile app to propose products like loans and insurances to the users. The conversion rate, that is to say the percentage of visitors that purchase a product, is often pretty low. The reasons are multiples: a great diversity of offers, non-targeted propositions, doubtful users… This article presents how Craft AI boost a website conversion rate by proposing adapted products to the users using predictive UX and hyper-personalized content.

À télécharger

Understanding user behavior...

Before proposing hyper-personalized content, the system needs to understand the user behavior on the platform. If a new user lands on the website, what path will they take? What element will they interact with? Obviously, this is not a simple task, there are hundreds of different pages on the website and a user can take thousands of different paths. As a human operator, creating an adaptive UX using rules that take into account all of the possible actions the user can take would require a tremendous amount of work. Moreover, adding new products and new pages on the website means that the operator needs to regularly update those rules.

Feeding analytics data to an AI enables the understanding on how a user can evolve in this environment. Craft AI is able to predict the most probable paths that any visitor will take. It is able to find the product which is the most likely to trigger the conversion and which path leads to it.

… To create a goal oriented AI…

In order to meet the company’s objectives, the AI needs to have a goal related to them. It is possible to fix the goal to be all pages related to a specific subject (e.g. real estate) on a specific period (e.g. September), a specific product type (e.g. insurances) or simply the user's most adapted product. With this information given to the AI, it can hyper-personalized the website's UX.

… That will hyper-personalized the UX

An user arrives on the website! The AI predicts possible paths that they can take based on their profile. By using a tree search algorithm, similar to what is used in Game AI, and in particular AlphaGo, it finds the best path to one of the goal pages. This personalized path will lead the user to their conversion.

We can suggest the user to take this path in multiple ways depending on the platform they are on: we can show a banner to the next page, push a notification, display information related to the topic…

The UX adapts itself to the user and evolves with their exploration of the website to show hyper-personalized content.

Results: increasing the user engagement

Choosing Craft AI to realize a predictive UX will make you able to push hyper-personalized content to user related to your goal. It will increase the user engagement as the proposed product is chosen according to their profile. Moreover the AI will evolve automatically with the website and with the new proposed products and new pages that would be added. This AI is directly integrated in your existing website with the help of an expert data scientist from craft ai who drives you from the expression of your need to the production launch of the solution.

This AI can be adapted for any e-commerce website to find in which products the user can be interested in. In an e-commerce website for example, it would display a link to the next products based on the already seen ones.

Of course, predictive UX can be used in a lot of sectors and on every platform that use UX. Craft AI is notably applied on the pilot interface for a project with Dassault Aviation to minimize the informations displayed to the pilot depending on the current situation.

Une plateforme compatible avec tout l’écosystème

Google Cloud
OVH Cloud
Tensor Flow
mongo DB

Vous pourriez également apprécier


How MLOps will streamline your AI projects?

When speaking of Artificial Intelligence, the efficiency and profitability of projects depend on the ability of companies to deploy reliable applications quickly and at low cost. To succeed, you need to organize and improve the processes for creating, implementing, and maintaining AI models with a diverse and sizable team.

Lire l'article


Don’t just build models, deploy them too!

You don’t know what “model deployment” means? Even when you try to understand what it means, you end up searching for the meaning of too many baffling tech words like “CI/CD”, “REST HTTPS API”, “Kubernetes clusters”, “WSGI servers”… and you feel overwhelmed or discouraged by this pile of new concepts?

Lire l'article

IA de confiance

Un-risk Model Deployment with Differential Privacy

As a general rule, all data ought to be treated as confidential by default. Machine learning models, if not properly designed, can inadvertently expose elements of the training set, which can have significant privacy implications. Differential privacy, a mathematical framework, enables data scientists to measure the privacy leakage of an algorithm. However, it is important to note that differential privacy necessitates a tradeoff between a model's privacy and its utility. In the context of deep learning there are available algorithms which achieve differential privacy. Various libraries exist, making it possible to attain differential privacy with minimal modifications to a model.

Lire l'article