Over the last four to five years, we have been pleased to get to know innovative companies developing Artificial Intelligence offerings, applied progressively and successfully to an increasing number of sectors. However, we have also seen many start-ups as well as more mature companies claiming to integrate AI into their products. The term AI is often used as a must-have buzzword for investor pitches rather than a genuine technology development.
AI is not a new market segment: it is first and foremost an academic discipline that has given birth to new families of techniques and tools now applied across the different segments of the high tech universe, and that are leveraged by a new breed of companies to challenge the incumbents or to create new markets.
As a result, AI has become a trendy yet noisy space where it is complex for investors to identify the actual value around this technological change.
We have recently advised 5 AI-centric companies from different verticals and at different stages of development:
- The sale of Qopius, a French Computer Vision specialist in the Retail industry to Trax, a Singapore-based retail tech company;
- A €11m Series A for the Israeli pioneer Ibex Medical Analytics which deploys the 1st ever AI-based digital pathology cancer diagnosis system in a live clinical setting. This financing round was led by Israel-based HealthTech VC aMoon together with 83North, Dell Technologies Capital and existing investor Kamet Ventures (AXA’s investment arm);
- A €14m Series B for Paris-based Dreamquark, a software provider of industrial-scale AI solutions dedicated to the financial services industry, brought by a group of new investors with strong expertise in the Financial Services space including Alma Mundi Ventures (lead), NewAlpha Asset Management, Keen Venture Partners and AG2R LA MONDIALE, with the participation of historical investor CapHorn Invest;
- A growth equity round for Datawords, a leading provider of technologies that aim to help global brands deploy their digital strategies internationally. This financing was designed to finance the development of their AI platform to automate and dramatically streamline the workflow with their customers;
- A Series A for AI-powered digital consumer lending platform Finfrog, an investment from Raise Ventures.
In this article, we share our views on a few takeaways for these transactions on several topics such as today’s prevailing AI-based models, their applications as well as an opinion regarding value potential around this technology. Of course, those are bound to evolve as the technology further matures.
1. AI-centric companies can be segmented around 3 main different business models and offerings.
(i) Data science platforms:
This category deals with AI/ Machine Learning platforms enabling customers to become data-driven companies by leveraging AI outcomes internally. A few well-known players include companies like Databricks, Datarobot or Dataiku. They offer cloud-based data science software platforms who enable data analyst teams or other users to perform data science projects.
Most recently, a new generation of players has emerged, more industry-focused, who bring actionable industry-wise insights and contextualization.
For instance, DreamQuark offers Financial Services players with a scalable software solution strongly focused on delivering explainability. i.e. empowering users with the capacity to explain the key factors influencing the decision pushed by the artificial intelligence core. This explainability feature is key for Financial Services customers, and is often linked to regulation and compliance constraints, as decision rules on credit or investment decisions for instance, must be auditable.
Through its software platform Brain, Dreamquark users can load and process data to deploy and integrate AI models generating insights and outcomes for financial institutions in the fields of marketing (CRM, churn detection), fraud detection, risk detection or asset management.
(ii) AI vertical application software companies:
These AI companies offer customers a fully-packaged and verticalized software aiming to address very specific applications, goals and use cases.
The list here is long, and the universe almost infinite, with a huge diversity of topics, as shown with two of our recent customers:
- Qopius is offering a computer vision in-store technology for the retail industry, addressing challenges such as out-of-stock, sales analysis, real-time planogram adjustment, etc…
- Ibex Medical Analytics provides an AI-driven pathology cancer diagnosis system deployed worldwide: Ibex’s Galen™ Prostate solution is the first-ever cancer detection tool used in routine clinical practice in pathology labs, with demonstrated success in detecting missed cancer cases. The application helps pathologists enhance cancer diagnostic accuracy by automatically analyzing slides and alerting in case of discrepancies with high clinical significance (e.g. a missed cancer) – thereby providing a safety net that reduces error rates and enables a more efficient medical workflow.
These players can offer fully SaaS or on-premises solutions, depending on customers’ IT systems and data privacy rules. They often start to gain customers with a strong added value on a specific and sometimes narrow use case, before broadening the scope of the offering for the same customers groups on new applications.
(iii) Service providers delivering AI-powered B2C or B2B services:
In this third category, AI is not packaged and commercialized as a software but is the underlying technology powering a service giving an edge to the company to better deliver its services.
For instance, Finfrog is among the new-gen companies that apply AI to revolutionize consumer credit. Via its proprietary machine learning scoring solution, the Paris-based start-up has the ability to decide in real time whether to grant or not a credit to a customer.
Finfrog leverages a mobile-centric user experience and the EU PSD2 directive allowing users to easily and safely share their bank account information (through a trusted third party). It enables a fast process for small personal loans meant to avoid high overdraft fees incurred for instance by freelance workers and lower income households.
Datawords, which has been providing for more than 20 years digital asset management services to global brands, is another example of a company having implemented AI to innovate in the way they power their services. Deploying globally a consistent digital communication strategy has become increasingly complex, due to the booming volume of digital contents, formats and people involved on one side and the need to perfectly tailor digital communication to local markets. Datawords therefore developed a proprietary technology based on AI to optimize both quality and cost of digital assets localization and streamline internal and external workflows.
2. AI becomes a transformative asset when targeting niches and vertical areas where it can rapidly reach an extremely high performance regarding very well defined use cases. Because performance is key.
One of the main benefits of Artificial Intelligence is to unlock productivity and value by automating data analysis, business intelligence and optimizing traditional operational processes (reducing costs and/or time). By adopting this technology, companies can obtain a differentiating factor and create a competitive advantage. But achieving the positive outcomes described above thanks to AI technology requires an almost perfect level of algorithms performance: accuracy is a critical success factor in order for AI to provide reliable and actionable results.
Reaching a best-in-class accuracy requires several ingredients:
- a flawless and powerful set-up regarding data preparation;
- a constant test and learn / experimental approach with training on various datasets;
- a deep business understanding;
- A human guidance, intuition and support in order to find the perfect chemistry.
Time is of the essence to reach this objective.
While it is extremely difficult to create real value with AI on applications addressing too generic issues, an acceptable level of accuracy and performance is achievable more rapidly by addressing a targeted industrial or vertical use case: Qopius in retail tech and Ibex in healthcare are considered technology leaders in their respective categories as their algorithms significantly outperform traditional approaches. But this reasoning could also be valid for other areas such as fraud detection, military intelligence, industrial maintenance, etc… and more generally to many industries that have hitherto remained relatively non impacted by technology.
3. Artificial Intelligence is still an attractive investment thesis for investors
“AI” is no longer considered as the magic word to peak investors interest. It remains a central theme for most investors but they now need to be truly convinced by the ability of a company to leverage AI to actually solve a business challenge.. That’s why they look primarily for verticalized solutions, and for companies that have a clear and measurable advantage in terms of performance delivered by AI (accuracy, efficiency, etc.). This superior performance level often stems from an early focus of the founders on solving real business challenges. With an iterative approach applied over time to different and higher datasets volume, such teams have the opportunity to create long lasting product differentiation.
Besides, AI is not a subset of SaaS and needs to be approached with a refreshed methodology. For instance, the economics of AI software players can be quite different from traditional software businesses: lower gross margin due to higher infrastructure costs and higher human support requirements, scaling challenges and the risk of weaker pure technological entry barriers due to the commoditization of AI models. Nevertheless, companies overcoming these challenges will be big winners for their investors.
To know more about our track-record, please find some of our business cases.
To further discuss this topic, feel free to contact our dedicated team:
Clipperton is a leading advisory boutique dedicated to the support of fast-growing tech companies and founders in their strategic transactions (M&A and Private Placements). Founded in 2003, the firm today has offices in Paris, Berlin, Munich,London, New York and Beijing.