You’re Being Lied To: When the Aggregator’s Quote Is Not the Trade

Transcript

Welcome back from lunch. I hope everybody is re-energized. We've got an exciting afternoon lined up for you. First up, I am delighted to welcome to the stage Connor Howe, who is the CEO and co-founder of Enso. Um Enso has been running now for just over 5 years. Uh has a fantastic track record. Uh Connor is here to talk about about DeFi. So, for years now, every DeFi user kind of relies on the number on the screen. Uh Connor has 2 years of data to say that you shouldn't. Welcome, Connor.

>> So, aggregators are lying to you, and today I'm going to show you how they're lying and how you can prevent them from lying to you in the future. I'm Connor Howe, one of the co-founders of Enso. I've been building full-time in Ethereum for the last 10 years. I wrote one of the UK's first research papers on Ethereum back in 2016. I built a fully collateralized stablecoin with a Swiss franc stored in a Swiss national bank 5-6 years ago, and I built cold storage infrastructure that holds over $5 billion in assets under management.

Enso was founded over 5 years ago. We have over We have enabled over $20 billion on chain this far. We pivoted two two times to find product market fit, which I think we have right now. We have over 300 plus B2B clients. We've raised over 40 mil, and we're the number one on chain development tool kit for building applications. A bit of uh context what we're going to talk about today. Going to give you a rough overview of Enso, the problem that we're solving, a bit about the Shield dashboard. Maybe some people saw this launch 5 to 7 weeks ago.

How over-quoting is happening, and the dirty tricks that people are playing in this ecosystem, and the solution to this over-quoting problem. Bit of context, Enso released an orchestration layer that allows fintech institutions and application developers to access the on-chain economy. Whether you're wanting to access real-world assets, DeFi such as lending, borrowing, swapping LPs, or even cross-chain, or even other on-chain actions like NFTs, you integrate one tool get access over 250 plus on-chain protocols with all the data available for this as well.

High-level, without using Enso, you need to build all this manual plumbing yourself, integrate each on-chain protocol, audit it, maintain it, and build all your infrastructure of how to combine these unique actions together one by one. Whereas if you were to use Enso, integrate one tool, and everything is out the box for you on day one. Each individual action such as um lending, borrowing, swapping, bridging can all be combined into one call data, one transaction, the one-click future that we all talk about in crypto.

So, for example, if I want to lend, borrow, swap, leverage, bridge that asset, and enter a new LP, you'll be using what we call execute. We have core four core functions in Enso that people are using us for. The second part is data. So, not only do you want to know how to execute on-chain, but a user and your internal operations need to be able to make a decision before you go on-chain. So, we give you all the API data, all user balances, all portfolios, all pricing to be able to decide what you would like to execute on-chain.

The third product, we bring this all holistically together into one product, which we called Enso Earn. Once you bring in stable coins and tokens into your product, the question is how do you these users? You don't want them to be leaving your application, using wallet connect and going to Ave's interface or Yearn's interface. One, because you can't charge fees. Two, you can't get user analytics what they're using. And three, quite frankly, the user experience sucks. If you want to retain users within your application, just integrate Yearn. We're able to offer yields and other opportunities to retain your users.

The last, but definitely not least, which is a primary talk of today and the premise of this, is the quote simulator. So, simulations is a trend now in the ecosystem for I would say maybe the last two to three months. We've been doing this for the last two years. So, if you were to make a request to Enso saying, "Hey, I want to go from this token into these DeFi positions." We will generate the call data, and then we will simulate this to make sure what the outcome is, if there's any malicious approvals, and to validate this can be executed correctly. So, this is now a standalone product that we launched a couple of months ago that people are using within their own product suite.

Gaining traction 300 plus large B2B clients, 20 billion dollars settlement on chain. Our record was 3.1 billion in three days for unique 80 DeFi combinations. It still, as of today, is the most complex on chain interactions in the whole ecosystem for this amount of money. By one combination, I mean lending, borrowing, and going into an LP, swapping into two tokens, and entering an LP. That's just one unique combination. This was for the Bear Chain launch, Bycol Rhicol, that was fully powered by the Enso team.

And again, we've had zero loss of funds in five years operations. You know, touch wood that this keeps going. And we work with large projects like Alchemy Uniswap Etherfi Yearn Curve and many, many more. Five years building, 13 audits. Laying some context for you because what you're going to see of the dashboards is very indicative of how people take security and user funds as importance within our own company. So, 13 audits, other companies maybe do one, two.

We've done We've had zero hacks, again, touch wood. Zero loss of funds. We take security as high priority within Enso. We're also contributing to the open-source environment where we created Temper, which many people know. We created it 3 years ago. It's an open-source simulation tool that we actually took inspiration from and the learnings from that to create a quote simulator that's powering many applications as of today. Temper is used by players like CowSwap, Gnosis Pay, Gelato, Yearn, and many, many more.

We also created Enso Wearer. Um, again, not said this publicly, but people like Uniswap, their universal router for saturation was a fork of Enso Wearer. You've also got Mekina that's based on Enso Wearer. You've also got Centrifuge, the portfolio management, asset management, that's also based on Enso Wearer. So, all the core functions that's operating and powering the Enso infra is actually been open-sourced and used in many different areas to ensure that our core is is is moving forward in the right direction. So, Enso Wearer is this tool that allows you to build on-chain applications really quickly and easily and safely.

Now, not only do we have our own proprietary swap algorithms or our proprietary how to execute into DeFi, we also rely upon others. So, for example, maybe 0x is really good at curve stablecoin to stablecoin swaps. Maybe you've got 1inch that's very good at verbose tokens. So, not only whenever you make a request to Enso, do you get our own routes, we also call over 15 other providers to see who has the most efficient route for the particular action that user wants to do. This ensures that you have guaranteed quotes. You're able to do what you would like because we don't want, you know, our infrastructure partners to leave using our infra if they don't have an option what they would like to do on chain. They've got execution reliability

and also redundancy if our own execution algos go down. We also validate and simulate every single response and it's been battle tested now for over 3 years in production. As we do this meta aggregation, we are in a very unique unique position that we want to ensure everyone that uses us has the correct output. And by doing that, we simulate every single response that we get from other people. By simulating responses from every other people, we launched this dashboard called shield.enso.build.

This basically goes through all of the aggregators and checks who's overquoting, who's underquoting, who's not even able to execute a transaction, which caused simulation failure, and the latency for this. Now, obviously Enso is at number one. Now, I'm not just shilling Enso, but it's the data proves that Enso is the most reliable and the the provider who doesn't overquote. Whereas you've got some other parties that go from 0.41 all the way down to 31.66. Now, this is an overquoting average rate, meaning if I request you for 1,000 times, how many times am I going to give you the wrong amount? So, let's say the going rate is 1,000 1 ETH to 2,000,

I Enso will get it wrong four times at 1,000 times. Other parties are getting it wrong 310 times out of 1,000 times. This dashboard also allows you to go each individual aggregator into us if you want to, you know, go more detail into it. Again, it's all public and it's all available to use for free. And to really drill down into what overquoting is, say let's say I want to swap one ETH into USDC. The going rate is 2,000 USDC, right? Other people will say no, it's 2,450.

That's what overquoting is. Whenever they say it's 2,450, and when you get the call data back and you simulate this, you know it's only 2,000, but people have increased it to 2,450. And to really drill down again on the analytics, if you're requesting these parties 1,000 times, they will get their amount wrong 300 times. So, they will increase it whether it's one, whether it's 50, whether it's 500, it's wrong over 30% of the time. Whereas ours is only 4.1. Again, we're improving day-to-day right?

An example of the last week, the worst overquote, think there's a Ah, there we go. There's a trade size of 500 Yeah 955k. They say that's what you'll get as an output. The reality was 903k. A user is losing 51 grand in one trade. The question is, is this intentional? Is it not? How is this actually happening? One perspective and and viewpoints, and again, this is up for you to decide whether this is, you know, intentional or not, is actually artificially inflating quotes. So, again, I'll make a request one ETH to 2,000, and they'll say it's

2,100. Why would they do this? To get more order flow. If you get more order flow, you get more fees. Once you get more fees, you can get more You get more revenue, you can hire more people, and kind of play this game of fake it till you make it. You also have the opportunity of um UIs. You know, if you have these UIs, which many aggregator UIs do, you'll see many different options. And if you're a user, and I think most people even in this room, even at the amount of years of experience we'll have, you will click the best one.

Even though it may be completely incorrect. And maybe you won't even verify whenever it goes on chain. So, there's a massive incentive to do this. More volume and more private fund raise and this is a endless loop of deception to users. The second thing is, maybe they're just not simulating, right? Some people don't. If you want to generate a quote and the amount out for particular token trades or even on-chain execution, maybe they actually just have an error in sight, right? The fourth, but definitely not the last last what we've seen in the last couple of weeks, the last last last, is Uniswap hooks, malicious Uniswap hooks, and Curve pools are malicious as well. But, the question you really need to ask yourself is,

is it a combination of all of these? Some people are playing tricks to deceive the users or if it's just one or not at all. For example, a single toxic pool in the last 2 weeks for Curve, had a total volume of 117.8 million. They were able to generate 59.1k in the last 18 days, right? This is just 2 weeks. You can generate 60k and that's just one pool. Imagine a scale of these toxic uh toxic pools that are on chain in combination with over quoting.

And really, what does this result in of over quoting, under quoting, deceiving users? Is it if you're an infrastructure provider like Enso that works with entities that have hun- have millions of users and hundreds of thousands of users, you're really shifting the responsibility to the integrator and they're losing a lack of trust with their users. You know, I was thinking about this last night making making this presentation. Imagine my mom came with 10,000 bucks, right? It's a lot a lot of money and I would say in the crypto world right?

And goes to Revolut and wants to buy a particular token. So, out of these 10,000 bucks for 10,000 tokens and if it goes to $10 for these tokens, I make a 100 grand, right? And then she makes the trade, she's done all the research, and she goes, "Why did I only get 8,000? What did I do wrong?" You've already lost a user in crypto. You've already lost the trust. And she most likely will probably never use that application again. You've also got a degree of financial deception, right? We're we're being more institutionalized, we're turning more as a regulated entity. If people are intentionally or malpractice changing the quotes than what it would

be just to get a user to agree to it, it's like I go into a store and I buy a loaf of bread and it's two bucks. And I go to the counter, they scan it, and they go, "It's two bucks." The card machine says two bucks. And I check my my bank statement and say $3. That's deception. That's You know, there's going to be a lot of legal consequences in the future for these types of actions. Execution quality, reliability, you know, really we want more people to come to crypto as an industry, but if we're not able to give them a reliable quote, they're going to start dropping off more and more.

If you're an integrator, for example, if you use other parties, your customer support costs are going to go through the roof. You're actually going to most likely have to re-refund your users and to keep them happy with your with your service. Dirty tricks, um I'm not going to say all of the tricks that we found, um and I'm not going to say all of the solutions that we found because it is a a cat and mouse game. Um it's cats and robbers cops and robbers essentially. So, whenever you do simulations, um, some on-chain smart contracts are able to detect that you are running a simulation. And it will give you a higher greater amount out. So, let's say for example, run a simulation that says I'll give you 2,000, right? Whenever I

go to execute, they've got an FL statement on-chain. They'll give you two they'll give you 1,800. They'll take the full slippage amount, whether that's 2% up to 5, and take that all for themselves. You've also got setting fee based on a timestamp or block. So, that's called, let's say, quote execution delay. Um, and I actually don't think there's much research has been published on this that thus far. But it's just saying, "I know the user will maybe not make the decision in the next 10 to 12 seconds. So, I'll show them a really good quote for block 100, and I will inside my contract take the full slippage at block 101."

Another one which is really, really comical, quite frankly. Um, from this dashboard, people know that we're tracking them. And they know that we're publicizing their dirty, dirty little tricks. And they are changing all of their quotes just for 1inch's API. They're gamifying the public perception of what their quotes are. We have part of a partners, integrators, that have came even friends came up to us and said, "Yeah, man, you're showing this value, but I know it's like 10 15% more on the over quote rate, right?"

Which validates more just integrate 1inch so because you're not going to have these games that played. But it's really interesting to see that people are actively gaming this this this dashboard. Um, do do do So, yeah, cat and mouse game. I kind of like that picture to be honest. Good. Cool. So yeah. You're able to use Enso Quoter and Enso Shield to avoid avoid all of this. So, generate any call data from any source from any execution on chain, send it to our API regardless if it's for Enso execution, whether it's one of these aggregator, right? So, let's say they're gamifying our UI and you want to validate their gamifying, then go and

prove it. Take the call data and call our Quoter API. Whether that's from Uniswap or even your own contracts, and there's no vendor lock into this, right? The second part is this is tamper protection, the tamper detection. So, for example, whenever you generate the call data, you get these responses, send it back with the responses you got to second step validate that this is correct. There's actually no funds required for simulating these transactions. So, let's say you've got a user that wants to do a $10 million trade, and the user doesn't or at least you want to simulate a $10 million trade to present it on the interface, instead of you going and finding the top whales balance and override the balances for it. So, you

can do this for any any user and any token across all EVM chains. Well, it'll give you the exact output precision prediction in which you're able to validate what the outcome will be, and you're able to simulate this all all reliably. Again, really how it works, simulate send your transaction to the simulate endpoint, you'll get all the simulation ID, amount out, gas, and status. Status is actually quite interesting. Again, I'll go back here. I've spoken a lot about over quotes, but you also have simulation failure.

Which basically says, let's say this party who's this Who am I calling at some of the time? The call data just won't get executed, right? Imagine you're going into large bank, large institution, and 5% of the times you try to do an execution, it just doesn't go through. So, you're able to check this in all of the quarter as well and get the status if it's able to be or not. Again, validate all of this. And last but not least, present this all to the user for them to make their own choice.

Today, we're also noticing new features for this. So, we've also got automatic balance top up. So, if you don't want to define what tokens require a top up for this, we'll take the token and the amount in and automatically override the state to increase the balance so the transaction can be simulated. If you also want to build your own dashboards, again, I love to go back to this dashboard cuz it shows you the truth. Um you can also use a new feature we call batch quoting. So, you're able to call all these individual parties, including us.

Please, you know, help us improve as well. Take the call data and then call this endpoint in AnzoShields with all the individual call data sets and do a batch quote. So, you're able to see the response from all of them within one call, which increases, you know, efficiency and reliability. Toxic flow detection for Uniswap hooks and curve pools. Again, this is a cat-and-mouse game and we're not going to explicitly state every single detection mechanism because we basically have solved all of the ones that are out in the market as of today.

And from day one of using this, you will get this already within your product. Hidden fees detection. Um yeah, there's there's other mechanisms of changing it based on block time. Um yeah, also, there's many many others for how people are adding additional fees inside of this. Just to summarize, you know, just use Anzo really for your on-chain execution. We've been around for 5 years. Again, touch wood, no hacks. To go back to this do do do do.

To go back to this, over 50% of these parties, I think it's actually 60% have had hacks or loss of funds and some, you know, eight figures and higher. And it really shows you that, um, you know, the the trust in your own ecosystem, even as of today, it is sadly just just not improved. So, just use Enso, you know, even players like Yearn, even people like Yearn, they've been using this continuously within a product on a day-to-day basis. Um, before Enso, you're a sad little pancake. Before Enso, after Enso, you're happy.

Um, and yeah, that's a bit of overview of how people are lying to you. And I think now, hopefully, if you're building any application in crypto, quite frankly, that's doing any call data on chain, you should be using this tool or you should be simulating transactions. I think you're going to get into more legal consequences than most likely you're aware of. Um, because you are deceiving users by showing information is sadly just not not correct. So, this is one, uh, uh, a legal protection. Number two is also user retention protection. And three, it's providing the best possible service to your user. So, if you're offering like an AI agent bot, you

should really simulate this. If you're doing a, uh, uh, card processor and you've got like crypto payments with cards, you should really simulate this to show the users. DeFi, if you're fintech, if you're an institution. And again, you don't need to use Enso with the execution and data, you can solely just use this product within your environment. We saw one of the largest, um, cross-chain bridge aggregators actually start to use this in production and this is over 100 million quotes that we're handling, um, at a scale. And they're incredibly happy because now their users aren't coming asking them for for refunds. So, yeah, we had tomorrow have a workshop uh, 1:00 to 2:00 upstairs.

In this workshop, you'll be able to build a DeFi application but over 250 plus protocols on day one. Ask me. I think we've got some questions. >> Okay, we've got some time for Q&A. Does anybody have any questions for Connor? We can take two questions. Anybody else? We'll have to find him after. >> Oh, please no hard ones. >> We've one up here. >> Be nice, Marcus. Ha, ha. A really good question.

Um most people were happy and some people thought we were just out like outright liars. Um but we have the data to back it up. So, we were very surprised when the response of we're very happy with this and now you've found our faults in our our business that we can improve on. And the latter was you're just an outright liar. Um which was quite shocking, right? Because it gives you a perspective in how people want to improve.

And then the other component was how people are just starting to game this. Just for other people's perception, right? So, they know that we're tracking it and they know our certain IP addresses, they know our particular API keys, and they will intentionally change their quotes to be more realistic than what they will provide other people. That was really surprising to me. Um to the extent that again, you start asking your yourself is this intentional or is it not, right? Um but some of these people jump from you know, 25% all the way down to sub-10 within a couple of days, right? Um and

then people reached out and said, "Hey, that's not really the truth." And we go, "Well, they're changing it." So, that's I would say the the biggest shock. >> One more question. Any more questions for Connor? Okay, well Oh, yeah. >> Yeah, I'm in on this uh topic of under quotes and over quotes, we see the industry moving towards um intense and solver infrastructure. Yeah. Near intense um and the solver takes all the risk of ensuring a quote is uh you know, is given to the exact requirement. I mean, the volatility is entire risk is taken on by the solver. So, I mean, are your like plans to uh

you know, build out or taking solvers to your to your kind of infrastructure to kind of solve this problem? >> Yeah, I mean, so An Solvers and infrastructure that's agnostic of what it in integrates, right? This is just the aggregator dashboard. We also integrate directly with RFQ systems, also integrate directly with solvers and market makers that do some some do hold inventory, right? Um so, this is not a true representation of every other party. Um we were, I believe, you know, top three solver at some point in CowSwap back in the day like 2 and 1/2 years ago.

And most solvers still don't hold inventory. So, I think the narrative of intense is the ultimate solution to this still needs a little bit more massaging, right? I think market makers are in a good position to hold the inventory. I would like to see in a couple of years people holding more inventory. I just don't think it's really industry-wide and just yet. I I think it's a good solution. But again, we saw this in the last 2 and 1/2 years. Maybe it changes in 2 and 1/2 years. Um yeah, it's it's a good solution to the problem.

But I don't think it's the infinite solution. >> Great. Well, please join me in thanking Connor for his time today. Let's give him a round of applause. Thank you very much.