Announcing Our Investment in Axyon AI
We’re excited to share that Montage Ventures led Axyon AI’s $4M Seed round.
Axyon AI has been at the forefront of delivering advanced AI-driven predictive solutions.
By integrating cutting-edge AI-powered insights and active indices, Axyon AI enables asset managers to navigate complex market dynamics with greater precision and insight. The company helps uncover alpha opportunities with AI through its optimized indices on equities and ETFs and AI-powered asset rankings. Axyon AI is proud to be one of the world’s most innovative AI technology providers for financial services, being one of the few to experiment with Quantum computing applications to the investment management sector.
We at Montage believe that asset allocation is undergoing a transformative phase, and AI will play a pivotal role in shaping its future. Axyon AI is well-positioned to capitalize on this opportunity by leveraging its state of the art technology to deliver alpha opportunities via AI-powered assets rankings and model strategies on their preferred asset classes.
A key factor in our decision was the exceptional team behind Axyon AI. Their deep expertise in both finance and AI, along with a track record of successful innovations, instills confidence in the company’s ability to execute its vision and deliver tangible value to the financial industry.
Axyon AI’s solutions demonstrate scalability and adaptability to meet the diverse needs of financial institutions. This flexibility positions them as a strategic partner capable of evolving alongside the dynamic requirements of the financial ecosystem.
We’re excited to partner with Daniele and the team at Axyon. Our investment in Axyon AI is a testament to our belief in the transformative power of AI in the financial sector and our confidence in the company’s ability to execute.
We sat down with Axyon AI founder Daniele Grassi to discuss Axyon AI, their unique value proposition and how he sees the transformative impact of AI in financial services evolving over time.
What does Axyon AI do and what is unique about it?
We work with asset management companies to improve performance via the delivery of AI generated alpha, which comes in two main forms:
Our rankings of assets in terms of expected performance from several time horizons in different investment universities.
Model strategies, and academics that can be easily integrated into their investment products.
What differentiates us is that we have remained focused on the process involving advanced machine learning and artificial intelligence into the investment process. In particular, everything that has to be done correctly to ingest several sources of data and use those to create machine learning models capable of providing alpha and an edge on the market.
Why is what you’re doing difficult to build?
There’s quite a few companies out there providing consultative solutions where they try to build custom made models for specific investment strategies.
We are different because we are not model focused but process focused. We believe that there are certain complexities you have to handle when you employ advanced machine learning to financial services.
We are working in a world with a very small signal to noise ratio. So there is a lot of unexplained behavior in the market. And this makes the task of using machine learning very complex because the edge you can add is very small. You have to discriminate between everything that is noise and random from what you are really understanding from the market. We believe that this can only be done if you are extremely obsessive in terms of recklessness of the statistical and machine learning techniques that you employ.
We believe that this is only possible if you really define the process that has to be used and then lead the machine to actually execute it. As opposed to going out and handcraft individual models for specific investment strategies.
There is a commonality between models employed of different investment strategies given different asset classes when it comes to how you should treat data, and how you should allow a machine learning system to learn from it. So that’s why we’re always focused on the process itself because we believe that that’s the only thing that you can really control — how you handle data and how you go from there to a machine learning model that over performance market.
What has been the most challenging part of implementing AI in the financial services sector?
From a technical perspective it’s the data and the data quality, because unfortunately in our sector even when you work only with premium data providers the cleanliness of the data is not always what you want. It becomes extremely important that data is of very high quality. We understood from day one that we had to rely only on sophisticated established data providers with a large breadth of data and professional ways of gathering data.
The challenge has been to detect those data quality issues that are not immediately apparent. We have spent a lot of time defining this step of our process — a sophisticated statistical process focused specifically on handling this has become an advantage for us.
What is AI explainability and why is it so important to what you are doing at Axyon AI?
According to McKinsey:
“Explainability is the capacity to express why an AI system reached a particular decision, recommendation, or prediction. Developing this capability requires understanding how the AI model operates and the data types used to train it.”
I think transparency is crucial as it provides humans with a sense of control and understanding, particularly in our sector, wherein the outputs of machine learning directly impact individuals. Therefore, establishing and respecting operational boundaries in, for instance, an AI-powered investment strategy is significantly essential. This is achieved by making the algorithms and their analyses transparent, creating a narrative to understand what signals are influencing AI models’ decisions and market signals.
However, we must consider that we cannot wholly explain a machine learning model. If you could, then you don’t need it because a human could come up with the same conclusions. As an example, properly trained dogs have been used in security, defense and rescue activities without them being able to explain their behavior precisely. Similarly, AI algorithms can be used to build quantitative investment strategies even when their internal functioning is not completely interpretable in the eyes of the investor, provided that they produce performing, reliable, consistent results that can be generalized in different market scenarios.
Highly effective techniques have been developed to identify which characteristics of a piece of data have contributed most to the decision made by an AI model. However, it is essential to state that the difficulty in explaining the behavior of AI models need not prevent their use in real applications, even when the stakes are high.
What are you most excited about in this next phase for Axyon AI?
We have solved all the steps that had to be taken to get into a position where we can scale in terms of coverage, performance, data ingestion, etc. This fundraising allows us to have the resources now to push to sustain the product quality and its growth. We’ll be able to position ourselves in the market effectively, and to scale it from a commercial perspective, which is, in the end, what will determine the metrics for our next step.
I’m really happy to be working with Montage and, specifically, Todd. Apart from all the things I heard, this process has demonstrated that Montage has a different approach in the VC space. Todd has been amazing and we are very happy to build this company together.
