Discussing AI with the DN Capital team

I’ve been lucky during my 23 years at DN Capital to work alongside some real experts in their fields. One such expert is Raoul Oscar F. , who’s been working at DN for 2 years and is one of our AI specialists. In my first blog looking at the state of the market for 2024 I called out AI as one of the most exciting verticals for the coming years. To take things deeper and give more insight, I sat down with Raoul to discuss AI and our perspective on the sector at DN:

Nenad: Thanks for taking the time. To kick things off, what would you say are the key factors that make an AI company a standout candidate for DN Capital?

Raoul: We have a set of core investment theses and this is no different in the AI space, and applying this rigour means we’re able to distinguish genuine potential from hype in what’s a very overexcited space.

We are specifically looking for companies that can leverage applied AI capabilities to accelerate and improve their delivery of product and services. In general, we see AI as an enabling technology rather than being standalone and appreciate companies that can use this technology in clever ways to accelerate their or their customers’ product and business development, or to create a more lean infrastructure.

Our investment in Oxolo is a great example of this, we thought that its application of AI technology to deliver hyper-customized video messages in bulk could have a positive effect in leveling the playing field for smaller e-commerce players by giving them the ability to include video assets to promote their products at very reasonable costs, putting them to a similar level of larger players that have more budget or scale.

Nenad: In which sectors does DN see AI making the biggest near-term impact?

Raoul: It’s firstly worth noting that while the current AI boom is front of mind, AI has come a long way and DN has been making investment in the space for many years, starting from Shazam 20 years ago!

In the last 5 years, we made 10 investments in Applied AI companies across various sectors, mostly in the enterprise segment. Examples here are Incode which deploys AI for biometric identification, and Sanas which has developed a foundational speech-to-speech model used for improving understanding of call center agents operating in emerging markets.

An important theme for us at DN is what we call Data Integrity Management, which is centred around companies working especially on sensible data handling. New opportunities arising in the space are companies being built around the AI governance space as well MLOps companies specifically targeting industries that manage highly sensitive datasets, such as financial or health related data.

Nenad: Looking beyond immediate technological advancements, how does DN Capital assess the long-term value creation potential of AI ventures?

Raoul: In the long term, AI ventures should be able to deliver superior product while unlocking efficiencies at a company-wide level that should result super strong unit economics due to leaner and more effective organizations. That being said, in a scenario where products don’t have strong differentiation, it is hard to predict how customer acquisition costs will develop.

The differentiation between AI enabled products is one of the hardest problems to crack at this stage and is a large part of our evaluation focus, as it is hard to define what ‘moat’ or differentiation exists between products at this stage or how this will evolve going forward. We are also yet to see how the exit landscape will look like for the newer generation of applied AI companies.

Nenad: Which new AI trends or innovations is DN Capital particularly excited about?

Raoul: We are currently looking at a few trends that we think are very exciting in and around the AI space, namely:

Hyper-personalization: using AI to develop products that are tailored to individual needs rather than being generic. For example, as LLMs begin to include more memory features, the natural next step that they will begin to know individual users and be able to deliver greater personalization.

Levelling the playing field: fundamentally, generative AI is an enabling technology that has the potential to level the playing field and give access to products and services that used to be reserved to a small percentage of the population or of businesses to a much broader public. We may see small businesses more and more empowered by autonomous agents to achieve growth that would otherwise require costly headcount expansion.

Small companies with large P&Ls: Similarly to the above, with AI individual entrepreneurs and small teams of founders will be able to build profitable companies quite quickly and with limited resources.

Changes to education and training: Businesses often need to create educational and training videos but these place significant time burden on training teams and senior leaders, with AI-powered avatars able to accurately mimic leaders with which compelling training and educational materials can be created in a fraction of the time. AI can also be used to tailor educational programmes and the demeanour and approach of these avatars to a range of employee learning styles, increasing engagement.

Beyond this, a broader trend we’re watching closely is the diverging directions in which AI models are being developed. Many players are trying to build bigger and bigger models in the quest to create AGI (artificial general intelligence), but on the other hand, when you look at the builder community, one of the bigger problems they face is the cost of training and inference on LLMs.

As a consequence, many startups and companies are looking to create smaller specialized models that have very clear instructions to solve a use case rather than leveraging very large and expensive models.?