AI auto-generates M&A candidates

Traditional approach: Company X wants to expand in a particular technology area and wants to prepare a list of potential acquisition candidates. How does one identify these companies? How does one rank them? One hires an expensive investment banker to prepare a shortlist. Bankers love buy-side mandates.


New approach: Hire a machine.


Acknowledged that M&A is a lot more than shortlisting candidates but let’s peel the M&A onion bit by bit, shall we? Target list generation is a key activity and most conglomerates maintain an active list and spend hours of CXO time on it.


Here is how a machine can help - illustrated via an example… Let’s take the Electric Vehicle sector. The “biggest electric car maker” (acquirer) wants to identify companies (targets) that have the closest matching technology portfolio to buy.





Vectorize: 12,571 electric vehicle patents (recent ones) are accessed. This covers 1809 companies (patent assignee). The machine vectorizes patents (see vector cloud figure).





A 2-dimensional representation of patent landscape


In this process, the machine understands what each word means (See Inset: Machine understands text via vectors).


Interpret: The machine next understands what areas is the acquirer owning technologies in. Below is a plot of the focus areas which span from battery packs to torque control to thermal charging etc. 8 areas are identified by the machine.



A 2-dimensional representation of focus areas in vector space


Shortlist: For each area, the machine maps every patent of the acquirer with every patent of the potential target (1808 companies) and given our definition of finding closest technology targets it uses “relatedness”/ “closeness” metric to shortlist candidates. The figure below has 8 diagrams for the 8 areas plotting every company (1808) basis the patent vector analysis. The shortlist is in front of us. A shortlist for every area.



We defined the closest technology as criteria… we could as well define it as complementing/ most core/ most cross-connected/ etc… a metric change would change the shortlist.


Acquisitive companies maintain an active target list. The machine can repeat this analysis in hours. Compare this with going through a procurement process to hire a banker and the costs associated with this.


In the “digital” era where increasingly everything is a vector do we peel the onion or push the boundaries of what is possible with a different form of intelligence?








If competitor 1 acquires competitor 2… what happens, what technology vanishes… 1809 companies are a network where anyone can buy anyone else… should we acquire someone today given potential changes in the technology landscape?

Should an organic technology research project become inorganic (acquisition)? How do we model this?







Existing human forms of intelligence will struggle to simulate such complexities… machines can… albeit artificially.

0 views

Contact us

Decodem Technologies Pvt Ltd

402, Building 9, Shanthi Park Apartments, Jayanagar 9th Block, Bengaluru 560069

CIN: U72900KA2019PTC128245

  • LinkedIn
  • Twitter
© 2020 by decode'm.ai. Proudly created with Wix.com