The emergence of start-ups in the world’s economies and the massive arrival of investment funds have shaken up the business world since the 2000s, posing the challenge of information in an acute way. In this prolific and fast-moving world, the personal network and manual databases are no longer enough: new technologies, in particular artificial intelligence, are revolutionizing the search for targets in a business acquisition strategy (M&A). For example, the Sealk platform (in the pre-commercialisation phase) uses Kairntech’s AI to collect recent and relevant information on start-ups or SMEs, but also in a predictive manner: based on the analysis of merger or acquisition operations, a large group can anticipate which start-up to acquire if it does not want to fall behind its competitors; conversely, a start-up can seek out which large groups it might join. This powerful and original solution, intended for investment bankers, investment funds and large groups, requires no programming knowledge.

The world of M&A: personal networks and manually crafted databases versus new information technologies

In the traditional world of investment banking, new technologies are slowly making headway. 

Until now, the profession was essentially based on two pillars: a solid, often very personal, relationship network of senior executives at large companies; and manually crafted company databases created by an army of telephone operators whose subscription costs are often very high for databases with a solid reputation.

In a very simplified way, the first pillar creates the assignments, the second allows to identify the targets. 

The execution job requires strong skills in legal, accounting and tax matters as well as a talent for negotiation, even a very acute sense of diplomacy, as the merger of two companies is always an extremely delicate operation.

Over the last twenty years, the personal network of the investment banker has been enriched by  client databases of large commercial banks and by the sharing of local information between independent operators grouped together through partnerships or franchises.  

In this discreet, even secretive universe, whose operators are nevertheless very varied, from the giant global generalist bank to the highly specialized and sometimes local “boutique”, things could have gone on for a long time.

Two events have changed this situation:

  • The massive arrival of investment funds of all kinds, which are themselves in a way “buyers” of companies, but for their own account, and which do not necessarily need investment bankers because of their skills in this area.  
  • The emergence of startups in global economies. Indeed, they have become the origin of most technological innovations. Some unicorns become world champions, others remain independent, but most end their lives (sometimes very shortly after their creation) backed by a large group that finds an efficient way to boost its offerings or enrich its Research & Development. 

The challenge to obtain useful and strategic information

The world of startups is incredibly prolific, burgeoning, changing, and even the products or services they develop are sometimes the subject of a fad that can evolve from one year to the next.

To understand this world, you need to have not only access to a vast amount of information as so many companies exist, but also to up to date because their situation often changes: a financial investor takes a stake in the company, a strategic partnership is signed, new innovative features are introduced, an industrialist takes a stake in the company, a major contract is signed, etc… Everything is important.

The personal network of an individual or even paying databases are no longer up to this challenge. This is where Artificial Intelligence replaces traditional techniques and where the platform presented here takes all its interest.

If structured databases are still useful, the information available on the Internet, almost always free, exceeds in quantity several orders of magnitude the first ones and especially it is always much more recent. 

It is therefore essential to be able to access this formidable source of information, whose main flaw is that it is very noisy, i.e. it contains data that does not correspond to the object of the research.

This is where artificial intelligence comes in for the first time in the field of Natural Language Processing (NLP).

The Sealk platform: a solution based upon operational AI

The Sealk platform first performs systematic and targeted crawling of the internet to obtain essential information on all the companies monitored by Sealk (startups, SMEs, etc.). For each company, it is necessary to know, for example, its principal and secondary activities, the services or products it markets, its distribution channels, its business model, the names of its main customers, the location of its markets, etc. To do this, the platform uses the software solution developed by Kairntech, which makes it possible to create training datasets and customized AI models and to apply them to the documents collected in order to extract the relevant information. Kairntech’s solution does not require any programming knowledge. 

The solution can also be coupled with a knowledge base to improve the coverage and quality of the extractions. Thus, it is natively coupled with the Wikidata knowledge base. Kairntech performs a monthly update to have the most recent version of Wikidata and sends it to its customers. Sealk is now using this functionality in production.

All this information is aggregated with the information that Sealk already has, very often acquired through subscriptions, and which concerns changes in shareholders, acquisitions, fund raising, or investments made by institutional operators or large groups. This first step allows the creation of a unique and considerable documentary base on startups or SMEs from all over the world, and simultaneously on all investment or purchase operations whose targets were these companies.

Artificial intelligence then intervenes again, in a second step.

This second step consists of submitting this information to new algorithms whose objective this time is to discover whether these acquisitions or shareholding operations follow patterns, logic and therefore strategies, and which ones. It is therefore a tool whose aim is to be predictive: Indeed, by analyzing the recent history of a large number of acquisitions, the learning model obtained with its algorithms can answer the following questions:

  • For a startup: who is likely to invest in my company or to buy me out?
  • For a company: what technologies are my competitors investing in?
  • For investment funds: which companies are interesting for my fund? 

From a learning corpus to a predictive model

The main characteristic of AI is to be able to predict a behavior from a set of examples (the training dataset), by then creating, thanks to algorithms, a predictive model (the learning model). It is the use of this model on new data (those that are crawled continuously) that enables disambiguation and acquisition predictions to be made.

Kairntech platform: example of the level of detail for a company’s business that the Kairntech platform provides. This then allows for highly targeted searches for potential acquisitions.

Up-to-date and exhaustive information, describing acquisition strategies

With such a platform, investment bankers, investment funds, large industrial or commercial groups can have access to exhaustive and recent information, constantly updated, and possibly describing all the acquisition strategies deployed in the world’s major economies. 

This is mainly the result of a relevant use of artificial intelligence.

Of course, the Sealk platform described here also includes tools that allow you to interpret these results, i.e. the acquisition predictions, and thus to understand the underlying strategies. These tools are mainly statistical. They exploit the metadata collected in the first phase, that of research and disambiguation of information, and establish relations between the targets acquired thanks to these metadata. It is these relationships that allow the investment banker or investor to understand the acquisition strategies that are being implemented in a particular economic sector.

Finally, it is important to remember that all this is obtained at a much lower price than that charged by traditional players. 

Sealk platform: example of a company sheet, in this case the company Mirakl. The open tab shows all the information obtained: main and secondary activities, details of products and services offered, distribution network, types of customers, etc.