Artificial intelligence (AI) methods and processes fundamentally transform the economy and society. The potential future business potential of AI startups is correspondingly high. It is to be expected that many of them will turn out to be “unicorns”.
Thus it is very important for venture capitalists, banks & insurances to deal timely with often capital-intensive AI startups.
However, the procedures in use are new, complex, and often intransparent. It is of correspondingly high importance to be able to judge their correctness.
We have long-year experience in the areas Deep Learning / Artificial Intelligence and have supported well-known banks and insurances in these topics.
On this basis, we have developed a guide enabling the efficient pre-audit of AI start-ups concerning the most relevant topics.
With that, its is possible to assess the solidity of AI start-ups in an effective questionnaire-based way and to identify potentially critical issues.
Risks can be mitigated and losses from “wrong“ engagements correspondingly be reduced.
The processes known as “artificial intelligence” (“AI”) are already in use in many ways and transform the economy and society in a fundamental way. Their fields of application are as diverse as they are promising for the future and include, among others
- recommendation systems for customer-specific products
- autonomous driving for cars and drones
- automated recognition of clinical pictures
- automated high quality translation services
Despite progress already made, the number and quality of new developments remains strong and the number of start-ups founded for this purpose continues to be high.
Though the components often available “out-of-the-box”, the complexity of AI procedures is generally very high and requires relevant expertise. Thus, e.g., it is imperative to take care
- which AI model to use
- how to select and prepare the data
- how to ensure that potentially self-generated data leads to realistic results
- how to ensure that the quality of the results is sufficient
This is made more difficult by the fact that many procedures are ultimately “black boxes” and are difficult to validate.
In addition, some special features must also be taken into account in the implementation and the chosen infrastructure (like the use of GPUs, Cloud solutions, etc).
The necessity to enter into fast engagements in order not to get left behind and the at the same time high complexity of the issue means high risks for venture capitalists as well as for banks and insurance companies. These are reinforced by the facts that, e.g.
- many startups take advantage of the hype and (let) call themselves “AI startups”, although this is not always the case. (According to a study 40% of all European “AI start-ups” have nothing to do with AI.)
- many procedures are still very new and their effectiveness is doubtful
- complex AI procedures are often used, although they are not necessary and/or the data situation is not sufficient; often “classical Machine Learning” is preferable to Deep Learning
- many founders are very inexperienced and therefore initiate fragile processes
- unauthorized data or at least data from sources that are not bound by contract to provide these data is used
For potential investors and contractual partners, there is thus a considerable risk of possible misinvestments and unrecognised financial losses as well as damage to their reputation.
Our Range of Services
We bring in our guide, which covers the following relevant topics
- data for training, testing & production
- methodology use and validation
- processes in place
- used systems and hardware
- license situation
For each of these topics, we have identified relevant issues for which we provide a list of questions and risks as well as short descriptions.
In addition, we can introduce the guide in the scope of a workshop and support in overall questions concerning AI / Machine Learning / Deep Learning.
If desired, we also offer advice on special cases and support you in identifying possible risks and formulating and evaluating questions to be asked.
- insight into the critical points and the questions to be asked regarding AI start-ups
- effective prior clarification of critical points
- identification of start-ups where commitment is out of question
- use of freed-up time resources to review more appropriate cases
- make quicker decisions and stay connected to promising technologies
- if required, in-depth analysis of specific start-up in question
- reduction of risks of bad investments and damage to reputation
- saving of money and resources
Depending on customer requirements, a flexible procedure is possible.
You are welcome to contact us, preferably via email.
Dr. Dimitrios Geromichalos
Founder / CEO
RiskDataScience UG (haftungsbeschränkt)
Theresienhöhe 28, 80339 München
Telefon: +4989244407277, Fax: +4989244407001