but to start off the day i'm very very pleased to have a very special guest for our keynote speech because he is somebody who is one of the most successful vc investors in the european maybe even worldwide ecosystem dr christian nagel from early bird ventures christian has more than 20 years of experience he co-founded early bird he ipo'd companies like tip 24 or interhoop he even invested in unicorns that were exited this year like peak games and he still sits on the board of number 26 and has invested in uipath i can have a very long list he's also a very very successful sailor so basically he's a role model for me because i love sailing and i love vc investment and unfortunately he will tell us now how to just how our industry gets disrupted so christian welcome to the rise of ai and please enlighten us uh what is the threats that are coming at me and you maybe in the future yes thank you very much for the for the kind invitation
um um and yeah let me maybe maybe start with a little anecdote so how what why do we how do we why do we see um the industry is is changing it's it's mainly because nothing has changed for a long period of time and we actually have been there was about two years ago have been called by a friend a usbc who called us and asked us well do you know this entrepreneur he has a great idea he's from berlin you must know him he's around the corner and we said we have no clue and we thought we have a good network and and no things popping up and and interesting uh topics interesting entrepreneurs who come up with great ideas and we haven't haven't met him but this usbc have met him and then that happened again and then we realized well obviously they have something what we are missing they have support uh tech support so to say that i come to this in a minute what it means what we are not
have had at that point in time and that basically triggered more we have been discussing this and what needs to need to change what we can do to change let's trigger the idea of basis that we do that we that we need to act but before i tell you a whole story let me start with with some some support on the slides to give you a little bit overview of uh first of all um how vc works and many of you may know this but there's a value chain like like in all value chains and a lot of businesses and this is the venture capital value chain you see here so basically it starts with we call it sourcing basically getting to know the entrepreneurs um that's very manual um this is mostly mostly inbound and sometimes as you reach out also we do like like in many cases um then comes the screening then obviously from the many ideas we receive or or look after then we have to basically screen them down and
and see what is it what does it fit what do we like where do we have experience um where do we like to invest what fits into our our strategy investment strategy so this is kind of the things which defines basically how we screen what we screen and what we do then comes and everybody knows this who has found it then comes the annoying due diligence um basically where we try to find out um any liabilities so to say we turn every stone well not every stone but at least we try to get a good picture of where we are where is the company are the numbers correct and so on are the contracts so to say um perfect and then comes the closing um after this there's obviously much more work to do because it's just the start of the journey which takes a couple of years and this makes cases many years it's it's rather rather more tennis years than five years so it's a kind of a long journey where we do
then support the companies we have in our portfolio help them grow make them raise the next round then comes the next round and finally there comes the exit and as you see this process is is very manual um it takes time sometimes annoying time especially when it comes to closings can take place a long time and we thought after this experience something is changed or has to change over time so we believe um also with when if you compare where we where we have in the past and we see the future we see the past as being pretty much pretty basic inbound oriented so basically the vc uh i wouldn't say kind of arrogantly sits in the office and waits for good ideas to come that's a little bit a little bit over exaggerated here but this is exactly a little bit the picture maybe which which is usually outside and there's a little outbound so we sometimes reach out and try to try to get to known get to know entrepreneurs and then and
see what they're doing we believe this will drastically change so there will be basically it will be outbound so the money is looking for the best entrepreneurs and the ideas and not the other way around and especially in europe it will change in the usa it has already started to change because there's much more money available um looking for great entrepreneurs great ideas in europe it's a little bit the other way around now you have we have basically a lack of capital and this will trigger and and this change more to basically that we see uh successful vcs they have to reach out to the entrepreneur get to know them build a relationship and then may finally find something which is a fit and has to be fit from both sides and then the investment may happen so this is this is our clear belief and also this is supported by some experience in the past we um we saw now um to do this in an efficient way um because we see so only
limited capacities so and and we see a deal universe if we look into europe something in the range of 12 000 investments being made in this intervention space there are many more more ideas out there we believe it's in the range of 40 50 000 great ideas out there basically some find fine to find financing some not but this is the pool basically which is out there of a combination of entrepreneurs who have developed something um and in many cases are looking for for for funding or at least to look for build the business and look for in many cases for support to build the business now this is a lot per year it's i'm talking per year so kind of 50 000 per year currently we have in our database something like 12 000 which is also a lot and our task is basically to find in this pool the best combination of us being a good partner and the best entrepreneurs this is what we do and obviously with um with this becoming
bigger it's complicated and that's why we started basically to think about okay what can we do how can we use tech to help and support the process of basically of finding the right the right ideas and also then helping screening them in a way that we very efficiently come to basically the combinations of the vc and the entrepreneur being being a perfect fit and what i what you can see here on this chart is basically that we try to identify and how do we identify this is tech supported we have developed our own tech in in this respect now we try to first identify the the the best ideas and this goes via basically scrolling and identifying the sources we have web crawlers basically crawling uh github calling product hunt we go through public registries like uh hundred star company houses house and so on we take social media so we have a lot of sources it's it's more than 100 sources basically we are tapping into to um identify first of all um and
this is we do so then we know we know something we know the company we know the name we know we know a couple of other things then comes the next step where we basically enrich this data where we start making this more valuable by by using other sources like news like payment data traffic data we look into crunch base into the investors being behind the financing round and we look pitch book cb insights so we lose we use other sources basically to enrich this data so out of these kind of just information about a company comes more information which are relevant uh to us to us in the sense that we um basically have diff have certain criteria of what we look at and this is different from vc to vc some like to invest more in that sector some like to do more fintechs some like to do more logistic tech some do others so but this was very simple obviously then we have other criteria like which phase is the company in
uh which geography is the company in um and so on so a lot of criteria basically we fill into our engine now and start to enrich the data by this by information which we can can use basically to identify the the right um right companies here and this process with this process basically the outcome is that we know much more that we have a much enriched profile of a company so we already know much more and this goes not just into companies and can be something like a developer um publish something on on github and see saw some traffic um jumps or something like this um camping app being in the app store can can be something like this very simple examples now but just to show that um these signals we take and then it can identify um a potential interesting investment opportunities for us or interesting ideas into an interesting entrepreneur behind these ideas um so now we have this enriched data and now the point is that with this enriched data basically
we um we we benchmark those so we try to basically find out um what is what is relevant and for this we do take these sources and this is in this is where this algorithm starts to learn so basically we say well this bench this database is better than the other database and we this is a research study which was done at the technical university in munich and they basically compared different sources and compare them and rank them so what you see here is the ranking of different sources in terms of how valuable are these information how precise they are how actual they are and so on so we start ranking basically the different sources we have for them to the to the importance basically to consider them and make them relevant for our engine and this is what we doing come out come with the ranking and this is also learning basically because this can change over time but at least you have you have a starting point from which you can uh refine
the the the importance of different sources and the the um the ranking relative ranking um of those with regard to basically the evaluation of certain data points we we receive out of those data bases so what what in the end it comes to this this metric this is this is interesting to understand so in the end you want to be um and on the on the on the horizontal axis you have the kind of the how efficient you are are you manual menu is hard to handle the data flow we have currently or are you completely automated and then the question is how do you trust these these these well sources i just mentioned but how you also trust any algorithm and um obviously you like to be very automated and like to be have have reliable a reliable um outcome the issue only is that today nobody really trusts trust the engines so you would always say well i rather trust something where i have objective control where i have um grip on the
data where i've seen it where i have basically have a must trustful basically source a direct source not indirect sources and don't trust an engine which base combines different data over time and so this is what you expect but you like to be basically in the right upper corner here where you have basically a high efficiency meaning highly automated and you can trust the results here and then and we also refer here on the study which is done on the technical university of munich and which comes to come came to interesting results and what this shows here is that um basically you can trust the engines so um we have reached a state of machine learning and capability of ai algorithms that you can start trust these engines and this is basically again relying on this study done on technical university of munich and they compared basically the outcome of different investors in terms of if you give the investors the same set of companies 10 companies and the investors should choose the ones which
are the best and this is what we did here or what they did they basically they did this and had basically um investors choose among those and then they also ranked the investors in terms of others very successful investors is proven over time so they really did a lot of data data analytics here on on that end to find out um how does the um the investors so basically us compare to what the engine would would recommend and the outcome is quite surprising in the end so the engine is as good as the best the best industry what you can see here so you have in terms of accuracy and in terms of recall we have we are at 80 percent here um meaning so to say that the best investor has an 80 hit rate on the best um best investments um done and the engine um the extra boost comes to the same same level so also the 80 percent and this shows um or at least this indication um where you now
can say you can trust somehow the engines more than maybe in the past so that that doesn't mean that you basically the engine is taking the decision but at least shows that the algorithm can support you in terms of finding interesting according to your profile what you are looking for interesting investment opportunities and obviously the investment decision still stays with the investor and is not being replaced by an engine but at least you can you can say well it's heavily supported and properly will be more supported over time so heavily maybe not as of now but over time will be more supported by an engine who um gives you more transparency uh gives you more accuracy and in terms of what do i really look for and where do i find it and probably shows you things you have never discovered before and um and i mentioned this um so far we have been seeing um something like 8 000 investment opportunities per year um with the auto sourcing basically so the engine sourcing adding
to this we add this up to 30 30 000. um and this is um just by the way the western european part of um of of europe so it comes to thirteen thousand now the question is or the point is not that we always try to achieve and get more interesting opportunities the point is that we would like to also have support in terms of more efficiently filter this down to the one which are really relevant which i meant the perfect matching between investor and entrepreneur and this is also something we see being supported by it can be supported by an engine so that we in the end are not flooded with even more investment opportunities but it's being basically we see even less or busy active deal with less but those are more relevant and that's a big difference from from the past that you have basically tools which can help you and source and scrape and so on to increase the number of potential investment opportunities but this is not it you can
do this and there are other ways of doing this but with the smart machine learning engine with smart basically implementation of ai algorithms you can basically qualify this qualify this to an extent that you find again much more efficiently and you basically be brought you basically you see almost everything or have at least access to everything and the engine helps you making making this pulling this down to a more size which you can then manually go after and and and focus on which is obviously much more efficient uh for for both so it's not much more efficient for the entrepreneur um and also obviously much more much more efficient for us um if you then look into the value chain um what is what is happening so we believe that there will be a increasingly support on the um on the sourcing side you can you can describe it as as kind of we argument the bc so that we see get something which gives us more insights into what is out there where are
the interesting opportunities and helps you and this is the next next step in the value chain which is marked with a with a with a check mark check mark somehow i wouldn't say this is completely solved but at least there are solutions you can build solutions we have built a solution which helps us also with the screening part of it we believe that this will continue we believe in due diligence we see also something we haven't basically have gone to the next step as early bird now but we believe that it will happen um it will also be supported by an engine um the closing contract thing will also just as a point smart contracts will come why do we need to basically put this out of contracts why is there not something especially as a kind of standard um bonus i think this will also become much more efficient um and and but we haven't again we haven't done any any effort in this direction but we believe this will come to also make
these steps more efficient um if you look to support for the portfolio companies we have invested we believe that there also will be support with regard to competitor analysts where are they basically um so that also we see there that they can be can be very valuable um use of of machine learning and and ai algorithms and so tools basically which can support even further across the the entire value chain so um and we hopefully in the end also everybody will benefit from this so the entrepreneur who basically gets a better match in terms of finding the right pc at the right point in time and also basically with regard to the process in terms of getting getting closing and everything faster done and so the entire traditions process process being much more efficient and then also more and better support in the course of the lifetime which is mostly time-wise the most of the time you spend together um this is basically what i wanted to to present here um so in terms of
the use of of um or the change um i won't say it's a disruption of the vc but at least it's something which which is a change and we haven't seen it change for a long period of time so this value chain i presented is almost unchanged over the years and now we believe that this will will change and will become much more efficient and hopefully beneficially beneficiary for for both again as for as mentioned for the entrepreneur and um also for the vcs making use of it and we are not completely new in this um the us basically gave us the final trigger to develop this and now we are the stage that we're really using this and but this will be i think where we believe the norm for vc thank you very much thanks a lot christian for being exactly on time that was perfect we have six minutes left for questions and if i may i would pick something from the chat we have one question um from torsten um sorry
i have to read this kahufa aren't you afraid that clever ai founders will understand the algorithm above all make themselves valuable by tuning the data sources of the algorithms maybe provoking github stars could be an example well there's always way to cheat cheat engines then the engines will come better so it's probably it's race always like like everywhere um and it's not so that we basically blind tapping into something it's not basically that that the algorithm is finally writing the check or that the the the contract will be automatically executed i think that's not the case so it's still something where we can where we can interact then and and obviously find out and and check whether it's it's it's uh it's correct or not um and i'm sure we will be tricked that's that's that's a given so maybe one advice to the audience if you want to get attraction by vc just change your title to founder and everybody will jump on you um and another question is um since you you mentioned
the picking phase or selection picking phase there are some there's some data um like market size would you also look for that or is that something your analyst would do in the from the ai side sorry yeah i think um well the i can help here i think there's still a lot a lot to do i would say since sometimes in some cases not so easy to to grab market sizes um especially since we like ideas or or or products and developments which are not not have a defined market size so um this is interesting that you basically open up new markets in those cases obviously different uh difficult to to grab market size and analyze market sizes but this is something we are um are we like somehow um because opening up new markets is very interesting because you basically have no competition in these markets and find many cases kind of blue ocean type of things so but um so we haven't made our mind how this can be basically supported by the
engine but but again this is so i think we can't completely um uh eliminate um okay also make this manual support and also support and within ubc but i think to a large extent you can optimize the process what is your feeling like based on the current results that you have and i think you started also this year probably also before that digging deeper into the topic what are the main drivers for for selection criteria is it the people where they come from universities is it more the product so the recognition on product hunt what is really helping you that you can already analyze today well it's always the combination of the the entrepreneurs or the the the people um the person uh we can we like to get to know and the product so this combination always and sometimes obviously the person you can't you can't probably the engine can't help too much on the other person um only maybe with regard to a background is it a repeat entrepreneur has it been backed
by others so there's still so little little support which is which is helpful uh with the engine which you can do to the engine but then it's about the product and um good products basically um they just show traction and normally ideally traction without spending a lot of money so can can you can see early traction early on the question is does it mean a lot can but not necessarily but this is something we can we can also get support from from the engine basically to find out the these um these spikes you see and basically in take up or usage of products this is something you can see in the end it's a combination of of a fantastic entrepreneur who has um has the right guts and has the vision of realizing something and the product and since this is an ai conference can you shed some light what kind of ai algorithms or areas of ai are you usually applying at the moment is it more the nlp part where you have maybe
it helps you discovering the right companies or is it more on the data crunching part where you automatically look for for certain patterns um currently it's basically the the first of what's the widening the funnel and as we train the engine because that's a manual part of it basically to set the right um to write basically measures and the right and to fine-tune the the relevancy of of different parameters this is exactly where we earned the process so now we have just a very much increased funnel and now it's about basically making them more official and finding out what is relevant for us and this is exactly where the in the face where this engine gets trained and where we receive early results quickly with the early wins you find currently in terms of making this more efficient and so basically guiding us to the to the right um in terms of spot on the focus areas we we like to invest and um and um but this is exactly the phase we are in
basically that the engine gets gets trained over time now with regard to our criteria so what are we really looking for in terms of which sectors which geographies and then obviously triggering down much more much more granular over uh over time cool thanks a lot christian i think there is a disruption coming i recently spoke to kochu and they told me they have 30 million annual budget for data scientists and i think at least our fund doesn't have that um yeah so we have two people um so so we're all getting there um let's see what happens thanks so much for for telling us some of your insights and uh maybe we can continue this on a sailing trip one day um so thanks so much for joining thank you very much