Hi UTU Guardians,

Our partners Kephi Gallery invited us to participate in an AMA on their Telegram group on September 3 2021. Our CEO Jason Eisen and CTO Bastian Blankenburg talked about our partnership and the latest UTU roadmap developments.

In case you missed it, don’t worry. Below is the transcript. Learn the latest in this UTU AMA!

UTU AMA Kephi

Q1: Please tell us a bit about yourselves, your position within UTU Trust and a bit about your background.

A1 [Bastian]:I studied and did a PhD in computer science/distributed AI. This involved multiagent systems using game theory, risk models, trust models etc. It also involved building payment protocols that incentivised agents to adhere to the protocol, before blockchain came along. So nowadays one would use smart contracts for some of this, and I find it exciting which possibilities for distributed systems are now available.

Later I worked in industry, then moved to Kenya because of private reasons. There I met Jason who convinced me to join his startup, the taxi app MARAMOJA, because he had this great idea for a trust mechanism. That’s because the taxi sector had traditionally a lot of crime here, and people prefer known drivers. Then from that we spun off UTU, to make its own product of the trust mechanism.

A1 [Jason]: Yea for sure….I’m Jason, Founder and CEO of UTU. I’m a serial entrepreneur. I grew up between Boston and Nashville, spent 10 years in Washington DC, first at university at The American University (studied International Relations) then 7 years as a consultant for USAID, World Bank, and others. I started spending time in East Africa through that work in 2010 and moved to Kenya in 2013 to start (at the time) the first Taxi App anywhere in Africa, MARAMOJA. We realized in the process of building that taxi app that we were solving the wrong problem. we realized the problem was actually about Trust and began to focus our attention there, building better models of digital trust. This was the origin of UTU.

R1 [Nico]: Excellent!! I cannot imagine how hard it was to do all those things before blockchain @databu!

[Bastian]: Haha yeah, there were some long nights coming up with a payment sequence for autonomous agents working together on joint problems such that there would be no opportunity for any of them to just take the money and run at any time. Easy with smart contracts…

Q2: You both mentioned MARAMOJA as the kick-off point for UTU. Could you tell us a bit about how you decided to create UTU and provide us with an overview of what UTU Trust is and why it’s a very useful product?

A2 [Jason]: absolutely, one of my favorite stories 🙂

A2 [Bastian]:I’ll leave the business case to Jason, but on the tech side, the trust mechanism we built is just very general and lends itself to all sorts of problems. Basically anything where users need to choose between different options in any online marketplace or platform.

Actually, users and/or agents/algorithms.

[Jason]: I tell the story in full here http://utu.io/blog/utu-origins/ but I can summarise it just for you guys…

UTU’s vision is to become the trust infrastructure of the entire internet. Our mission is to bridge the gap between how people trust in real life and how they are asked to trust online.

We believe in a more human-friendly internet, data as a human right, and the need to avoid digital trust dystopias as we’ve seen portrayed and played out in various contexts around the world.

At our core I can say we are building the trust infrastructure for the entire internet, based on first principles of how digital trust should have always looked.

R2 [Nico]: That’s quite a vision. And a very necessary service, considering how things are lately on the internet and particularly in crypto space.

[Jason]: It really grew out of this specific use case of how do I find a trusted taxi driver that we needed to solve for in MARAMOJA but the amazing results of our initial solution seemed to get the attention of people all over the world that started coming to us, asking to license our trust technology.

We realized that what we had built to solve our own problem was a missing piece of fundamental infrastructure across every sector of the digital economy

So we set about the task of doing that…

[Bastian]: Yeah, the crypto space in particular is really interesting. On the one hand, smart contract execution and other txs are trustless. But then, how many among potential users are really able to read and verify smart contract code themselves? Or the UI code of a d-app to ensure it’s calling the right contracts as advertised? So at the end of the day, most users rely on someone telling them that some protocol/d-app is safe to use or not.

And that’s exactly what UTU’s mechanism makes much easier and seamlessly intergatable into any (d-)app.

[Nico]: What a great story! To be able to provide a trustworthy environment in any scenario is essential. Way to go guys!

Q3: So what are you focusing on at UTU Trust  now and what’s on your roadmap for the next few months?

A3 [Bastian]: Right now, we’re actually giving some clients hands-on support with the integration, to speed this up but also to learn how we can make it easier still.

Perhaps most notably for a crypto audience, we’re integrating UTU with Ocean Protocol

R3 [Nico]: Impressive!

A3 [Jason]: We launched our trust API for all sorts of marketplaces and our creditworthiness API for financial services marketplaces in Q2 and as Bastian says we are integrating a whole crop of clients across multiple uses cases, sectors, countries….

[Bastian]: Most notably on the product roadmap (see also http://utu.io) is our upcoming UTU Protocol launch, and an Android SDK.

[Jason]: we will start sharing some case studies soon demonstrating just how much value these clients get from integrating UTU

[Nico]: Can’t wait for that 🙂

[Bastian]: Just like us ?

[Jason]: We’ve also got some super exciting developer community engagement programs coming soon that will give people lots of cool ways to get involve in the future of digital trust in significant ways

[Nico]: Excellent!

Q4: I’m sure our community is dying to know this…can you talk about the partnership with Kephi Gallery and how we’ll work together?

A4 [Bastian]: Kephi gallery being a marketplace, it’s a natural place to integrate UTU!

For instance, it will be able to tell people which artists are liked (or not) by their network.

A4 [Jason]: for sure…this is why we’re here!

By integrating UTU’s Trust API into the Kephi  marketplace, all kephi users will  be able to get personalized recommendations and ratings from similar and familiar buyers and sellers.

R4 [Nico]: This is great! you will be contributing for us to provide a better and personalized experience for our users!

[Jason] I’d note the concept of familiarity, beyond similarity, as a defining feature of UTU’s model by the way

Even more, you’ll have the ability to add badges and video stories for your favorite NFTs and creators.

Artists selling their NFTs will benefit from increased sales and uptake… which I’m sure will make them pretty happy.

Buyers will have an improved customer experience, as they can make purchase decisions in a more trustworthy environment….

[Nico]: Of course! And this is a great way to match potential sellers with potential buyers!

[Jason] exactly, optimize for trust, reduce friction, everyone is happy

[Bastian]: Exactly. Plus, our system is flexible enough to support app-specific signals. E.g. for Ocean data assets, we can show how well certain data assets’ liquidity pools are supported by the users’ network, and I can imagine similar signals on Kephi. As always, we’ll figure out the details based on user feedback as we go.

[Nico]: Thanks! We cannot wait to start working together!

At this point Nico selects some questions posted on Twitter.

Q5: One of UTU’s principles is to move away from aggregated ratings to a “human-friendly personalized narrative and data points. What kind of data points will be provided besides some sort of “creditworthiness” score?

A5 [Bastian]: The creditworthiness API is actually a bit of a special incarnation of our more general Trust API. Both are able to ingest any kind of context data. For example, this might include borrower tx histories for lending apps, or the info which car you own on a car mechanic platform.

Right now this requires a bit of setup work for us in the backend if a new kind/format of data is provided. Over time, more and more of this will be fully automated when onboarding new client platforms/apps.

A5 [Jason]: We can consider which ever and as many data points as our platforms feed to us. The idea is not to create one, universal model of trust for everyone, no such thing exists. Rather we build individualized, contextually relevant models of trust for every user, context, transaction – the goal being best fit rather than highest rating. So for one user a recommendation may be based on a certain set of features like price and quality, for another user an entirely different set of features may play more prominently in their decisions – so we’d show those

The foundation of trust however, is relationships, so we start with those.

R5 [Nico]: Good point!

[Jason]: we can map relationships from on-chain transactions, social networks, user-uploaded data, and data from our clients that we use to filter that huge universe of data points by the ones that will be most meaningful for you.

[Bastian]: Always with getting consent from the user btw. We’re building this privacy-first.

[Jason]: I always give the example of choosing a babysitter… imagine  considering hiring a babysitter with 50 5-star ratings from 50 random people and 1 1-star rating from your sister…

you follow 50 random folks who could be bots or your sister whom you know and trust?

[Nico]: Sister, most likely.

[Jason]: probably you follow your sister… this is at the core of what we do…the idea that large numbers of anonymous data points don’t compare to the right data point from the right person, to convey trust

Q6: From a technological and commercial perspective, how can traditional companies be integrated into the UTU ecosystem? What are your plans for traditional companies that continue to exist with their own systems?

A6 [Jason]: This is a great question! I’ll take the commercial side of course… 

UTU was built with this challenge in mind…That the ideal trust infrastructure for the internet is a web3 solution but it must be backward compatible to web2… since this is where the majority of the market currently is. We don’t believe in building it and they will come…we built UTU to be a bridge for web2 companies to become familiar and comfortable with the transformative power of web3 first through UTU’s role in their platforms as an honest broker of trust and steward of their users’ privacy.

A6 [Bastian]: Technically, we’re building APIs and SDKs that make it very easy to integrate UTU recommendations and feedback into any app, centralised or not. We started with a web app SDK, but as mentioned, Android will follow, and then others. We’re using the experience of helping clients integrate to learn the remaining pain points, to then remove those.

Basically, we’d like to make it as seamless as intrgating e.g. the Paypal button. But there’s some integration work particularly for the data provision to be done.

R6 [Nico]: Very interesting guys! For the looks of it UTU can be integrated into almost any type of business.

Q7: Aside from personalized recommendations, the other purpose of this partnership is to ensure ratings/reviews can’t be bought or manipulated. 

There are a lot of giveaways in this ecosystem. Most of them need to meet some requirements such as providing positive reviews on other platforms in order to qualify which is a type of manipulation. So, How does UTU’s Trust API which is AI-based is going to interpret such positive data reviews that comes from said participants?

A7 [Jason]: UTU’s role is not really to make anyone’s decisions for them about who they should trust. We often say that we take a descriptive, as opposed to prescriptive, approach to trust, meaning that we will show users all relevant information, good, bad, and ugly, and then let them decide. 

As the model interacts with you more and more it will learn how you feel about such an interaction.

A7 [Bastian]: Yes, we recognised that problem very early. There’s actually a lot of manipulation and faking going on with recommendations in general. E.g. a lot of Amazon 5-start ratings are bought.

Therefore, we designed the reward structure of UTU Protocol to counteract that. Users can endorse service providers, services, products etc. and stake some UTU Trust Tokens on them. When that leads to a recommendation to be shown to another user who then follows that and also endorses, the 1st endorser is rewarded. OTOH, if too many give bad feedback, the 1st endorser is actually penalised.

So basically there’s a punishment for giving bad endorsements.

At this point Nico opens the AMA for the group members to ask their questions.

Q8: Did you consider community feedback/requests during the creation of your product in order to expand on fresh ideas for your project? Many projects fail because the target audience and clients are not understood. So I’d like to know who your ideal consumer is for your product?

A8 [Bastian]: Yes, we encourage community and user feedback very much, and actually seek it out. Being a B2B2C business, that’s the only way to make it work!

Q9: So many projects just like to speak about the “long term vision and mission” but what are your short terms objectives? What are you focusing right now?

A9 [Jason]: At the moment we are really focused on customer acquisition and onboarding as well as getting our testnet and Android SDKs out. We expect to have more than a dozen clients spinning (and paying) quite soon…so all our energy is going there

Q10: How many team members do you have? Do they have enough experience in the blockchain field? Do they have any experience on working in crypto and non-crypto project?

A10 [Jason]: UTU has a team of about 25 people including PhDs in AI, Decentralized Systems, and other fields…we also have R&D affiliations with the University of Southampton, Turing Institute, and UKRI Trusted Autonomous Systems Hub and are backed by global AI powerhouses like SoftBank Group’s DeepCore and Zeroth AI…so yea…we feel well prepared.

Q11: One of the points that is crucial for the adoption and massification of your project is to be simple, therefore I would like to know how optimized the work processes are so that the Platform is truly simple to use for users?

A11 [Bastian]: We’re putting a lot of work into that, up to being industry partners in some research projects on this. And of course listening to the user feedback, as mentioned before.

That marked the end of the AMA. We hope you learned something new. If you did, please share it with the share buttons below so others can discover it. 


Stay tuned for the next AMA by following our updates on Twitter, Telegram, LinkedIn, Reddit, and Facebook.