27. January 2020From AI to product: The TEAL prototype

Yes, we have big plans: TEAL wants to become the new Smart Blockchain Ecosystem from Europe. With its combination of proprietary, proven AI and decentralized blockchain infrastructure, TEAL wants to create a wide range of new application possibilities. Over the past few months we have been working on many milestones on the way to this goal: Preparations for our IEO are in full swing (more information coming soon). Our white paper explains on more than 50 pages TEAL’s USP and vision as well as the development of the e-commerce and recommendation market in the decentralized web. We are also working on first products.
Our first release is the TEAL Marketplace. A marketplace that connects users with each other and provides each individual with the best recommendation or answer based on the information actively provided – not with classic tracking, but based on sigma milieu and similar interests, dislikes and buying histories. The TEAL communication wizard takes care of this. This personal assistant will be the user’s visual and voice interface to the marketplace and will serve as input and output template for our proprietary TEAL AI.
This may sound like any other personalization solution from the WWW, but it is completely different. You want to see for yourself? Using our AI and completely anonymized, publicly available data we have built a prototype. It is still in the alpha phase and of course does not have all the features the final product offers. But it will give you a first impression of what is possible with associative intelligence and freely available data without infringing the privacy of individuals.

The way to the prototype
In theory, it’s simple: Take your own AI programme and a data set. You use that data set to derive patterns. Merge the programme with the data set, so that these results are displayed on a large scale and individual correlations are revealed. Write additional programmes that visualise these relationships and make them accessible to the user.
In practice, we already had the programme, but not the data set. On the TEAL Marketplace, users will be given the opportunity to train the communication wizard to their needs with publicly available data such as their social media posts. We have not yet had any users for the prototype, but we have worked with freely available data sets that are provided for training such algorithms. We have included postings from Facebook and Instagram for the target group calculation and considered the factors time, interest, place, locations and geodata.
With this training data set, a product catalogue, like the one an online marketplace would later contain, and a map of the island of Mallorca (one of the most diverse test locations in the world), we were able to train our prototype and output it in an app interface.

The prototype at a glance
The resulting app shows four functions that the communication assistant on the TEAL Marketplace will later have.

  1. Start area: Every (fictitious) user sees his personal user profile with all stored information and can alter it at any time. In the Marketplace he will later secure this information with his private and public key – this is where the blockchain comes in. The AI accesses this information, compares it with other users and recognizes patterns and connections that make each user unique.

    Here you can see how our prototype calculates the user types according to the scheme of sigma milieus and weights the information so that an individual user with his or her likes and dislikes can be identified.
  2. Discover: a user makes a search query and, based on this, receives recommendations for his or her purchase or trip. Unlike traditional recommendation engines, these are also based on the tone of his postings, the stores he likes to visit, and other preferences. We only include the information that the user wants to be considered. He can also find advertising or further information here.
  3. Radar: With this function, TEAL Marketplace Online Shopping interlinks with the store. Which restaurants that suit my needs are nearby? I was just looking for Pandora jewellery for my wife? Then a Swarowski store will also be shown to me, based on the AI’s experience with people who are similar to me. But not Thomas Sabo – the AI recognizes that buyers of Pandora bracelets don’t want another collector’s bracelet from the competitor.

    In our video example, you can see where other people interesting for our test user are located.
  4. People: The TEAL Marketplace does not only bring users and products together, it also allows us to share common experiences such as travelling, spontaneous concert visits or simply to keep in touch with people who are important to us. These can also be new friends who have things in common: For example dog owners, jazz fans or free climbers. Social networks like Facebook need groups with moderators. Our prototype can bring people together with a central common ground, who will also get along well in other ways, at the push of a button and make it easier for them to build up a friendship with chats. For example, if you’re looking for a hiking companion on Mallorca with whom you also have something to talk about.

    This video demonstrates the search for like-minded people – called digital twins. This function creates the basis for the subsequent bringing together of people via the TEAL Marketplace and also serves as a learning environment for the discovery function of products and services explained above.

What happens next?
If you want to try it yourself: We are currently working on an interface that will allow you to dock your own accounts and just browse around. The data will of course remain in the test data set and will not be transferred to the later marketplace.