If you’ve ever spent hours online searching for the perfect pair of shoes to go with that new dress or a shirt that says ‘hip but not too hip’, then Berlin-based startup Wunder.ai is for you.

Founded in 2016, the company is using AI and deep learning to read customers’ minds and filter the millions of products on the web to those which are more likely to appeal to them. Being exposed to a narrowed-down choice of items helps alleviate the information overload which slows down the shopping and decision-making process.

The real-time cognitive AI senses customers’ buying intent and the unconscious drivers of decision-making by analysing their first party clickstream data. Sounds techy, and it is, but Wunder.ai understands that one of the most important propellors of decisions is emotion.

The company spoke to entrepreneurial website Startup Night  about how their system works, Their ‘deep shopping bots’ initially work to understand the customer and then use an AI-powered Sense-Infer-Act paradigm to function in real time.

Sensing is carried out by interactive personality quizzes, preference games and semantic searches, designed to be enjoyable for the customer but with deep mind sensors lurking beneath. The inference is implemented using ‘mind machines’ which state a customer’s preferences, activities, styles and emotions. This leads to a virtual ‘customer personality.’ Acting matches this online personality with product data, showing you what you want without you even consciously knowing it was what you wanted.

An important priority of the company, especially given the current data privacy climate, is that all the data and information these AI bots hold on the customer is easily accessible. In their product description on founder platform F6S they appreciate that ‘reading the mind of customers’ appears to be a complicated and futuristic power and so they’ve used a ‘whitebox’ approach. This provides complete transparency for users and businesses about the information held and allows alterations.

“We believe that this is a key element to foster peoples’ trust and seize the great new opportunities which AI will bring to society,” the company wrote.

Time magazine wrote an article explaining the subtle ways in which AI is altering the way we shop online with subtlety being the keyword. These algorithms and technologies have the ability to push people to buy a product when they’re on the fence or automatically show a customer a matching belt to go with those fabulous new trousers.

However, it is a standard human response to fear what we don’t understand, and if the technology becomes too overt or too pushy then customers will withdraw from the shopping experience, uneasy in the knowledge that something is intruding on their late-night Asos binge. It’s a fine line and many companies are using trial and error to find the sweet spot of non-intrusive snooping.

It may seem like online shopping is experiencing a new AI takeover but, in reality, companies have been sneaking this technology into various aspects of e-commerce for a while.

Adobe Target uses machine learning to provide real-time, data-driven information to retailers allowing them to provide perfectly timed and matched products for customers. They have also created a technology that allows the user to sketch their dream item and the tool finds similar products.

Pinterest has created a tool called Pinterest lens, effectively a Shazam for things instead of songs. You point the camera at something and it using the visual clue to help you find the product, or similar products, online without having to trawl through the slew of websites that potentially stock that sofa you saw in a cafe once.

New companies are taking AI and tiptoeing gently into the online shopping world to create a seamless and personalised retail experience for consumers, while keeping it firmly behind the scenes.