IBISA: Agricultural Insurance with Earth-Observation Data and Blockchain

Transcript

available and we extract some data on that like index or animal before going back to the Earth Observation part more about the principle of Ibiza so as I said Ibiza is a b2b solution that is built on local partners the local partners are the one that Honda mutual they collect the contribution and they pay the benefit and in order to make it more efficient and more resilient we use we can use Mobile Money payment micro payment once the farmer have contributed and are unbothered into the system every

months we will extract the satellites database of the cup covered those satellite data will be assessed by people we call Watchers I will come back to this after and those who are sure will assess the if there is a loss on that and what is the level of loss if there is a loss they will log it into the automated back office and this automated back office will trigger a payment from the local partner thanks to what I've described before we can offer a protection that is affordable for farmers and this protection is also easy because there is no need to fill a claim file as I say because of this automatic

remote assessment they don't need to fill anything and it's a low first payment and in season payment this in season payment is important for them because it allow them to put in place mitigation action during the season and not wait until the end regarding affordability thanks to the use of those two technology we can offer a solution with 20% of fee I remind you that the traditional insurance you and I use for our car flat we have 25 to 30% of feet and in micro cop micro insurance some product have up to 70% of

so on 30% of the money is used to cover the cost the risk here we have 80% of the money that is used to cover the risk a bit more about this con watching I just described before what we do is that we use all our observation data available from multiple satellite from multiple data providers and we put it in the platform and then we use crowdsourcing so we ask multiple people to assess those those data and to assess the loss of the plots or the groups we will see that just after they will collectively vote or let's say decide on a level that will be the input of the

process we just see before and they will be rewarded with reputation token based on the quality of the answer what'd I just say is that we don't assess plots one by one for two but we assess them by groups of two by two kilometers or something like that so it's like village scale we do that for two main reason first is that satellite data are not always as good or resolution of out this one as this one and the second is to avoid moral hazard because if you are a farmer and you are assessed like your neighbors you don't have any interest to try to cheat the system because anyway if you abandon your field your neighbor will not and

you will be assessed like your neighbor so we are currently mainly deploying to a country the first one is India so we are deployed in India with than than people mutual it's a local mutual that have two million members today we are deploying with them since one year with paying customers and we planned to arrive to five thousand farmers this year with them in India the other partner we have is a IBM in Nigeria so it's also co-financed by the

government affliction boy and here it's an association of pastoralist all over West Africa we also have a technical and financial partnership the first one in terms of size and in terms of history is with the European Space Agency so they support us of course because of our work on the Earth observation research second is the Luxembourg government as I just say that co-financed our deployment in ninja and last but not least is consensus con because so consensus you know here is the only place I think where I don't have to present them and

we were a part of the Tachyon program last year the last point of course like every project is mainly made by people so we have a complementary team with different type of knowledge insurance blockchain and space mainly or sub servation space satellite banking inclusive insurance all those to pick that up in Ibiza and that's all I promise you that it will be shot I take my promise but if you have any question of course feel free to ask are you working on a model to extract some

insight from directly yeah so that's a that's a good question that's the idea of I will rephrase your question to be differently is why we choose what shows count watching and not something else for us there is three way to be the Oracle let's say in blockchain sense the first one is traditionally insurance is to have an in-house expert so the in-house expert will take all the image and also assess on behalf of the company so first it's centralized and second it's not scalable at all because you double the number of farmers and the expert explode the other solution is did indeed is to automate that with machine

learning algorithm and things like that we to our point of view it's not that scalable because you will need to retrain your model in every region you deploy it's not that resilient also because you can have what whatever kind of problem in the data clouds or whatever other kind of problem and if you did not train your model to detect those kind of problem then it will not manage it correctly and the last is it's still somehow centralized because you will need to execute the machine learning algorithm so that's why we went to this town watching solution because it's

decentralized it's scalable and it's resilient at least that's the pet of the of the company but also one of in your question also if you are for example an AI expert you can become a watcher and ask your algorithm to assess so also we are making a sort of market where if people think that they can automate it in a clever way they can do it and they can earn reputation token yes so who can be a watcher so that's

also a very good question so until now it was more Earth Observation expert because we were work so we are doing the watcher since one year since the beginning of the deployment in India so at the beginning it was more our observation expert our knowledgeable people because they helped us to build the platform and to build the content of the platform what kind of image we show what kind of index what can because we can you can compute a lot of things on satellite image so we could construct the platform with them but now we are opening it more and more to not earth observation eligible people but people that just need sort of sort of training to say okay this is type of data you

need to look at and Exeter how and the idea is that the Watchers can be anyone that just have time and energy to put in that how expensive are these earth observation services and is earth observation enough to assess the state so there is two question here first the earth observation image mainly they are available for free they are put at the disposal of the community of the world by the NASA and open space agency mainly so the only things you have to pay is

the technological stack to have an API to request the data you want the time you want so that's cheap I mean when you compare it to cost of the rest so that's very emotional cost and the second question was okay what can we assess so indeed we cannot assess everything from space this is what I say I say if you have a caterpillar that will eat the maize do I mean the fruit of the plants that I cannot see from space but what we can see what we see well from space is drought and also excess rain that we can see it well because it's also in the wall region it's not on in one field and that's why also it don't make sense to assess plot by plot so found so that's

why when I say that Ibiza is covering crop it's wrong the reality is that Ibiza is covering drought so today we are explaining in India we are deploying saying Ibiza is covering drought last question do you have some company willing to share with you images but concerned about privacy so they don't want to give access to their data to anybody just for your services the earth observation data I mean to be honest we

don't have you have private companies that sell their data but mainly when they do that it's because it's really accurate date I mean with a resolution as I say we don't need that anyway because we assess in two by two kilometer zone so it's better for us to use NASA satellite because first there is a lot of scientific literature about it so we have a lot of feedback that it's working it's working well and it's well calibrated and everything so clearly there is no need for us to go to private satellite data rather than to compute to complement

rather than private satellite data which we use sometimes is more sophisticated data computes based on free data because on the free data available if you mix them correctly you have again a lot of scientific literature on how to compute drought model drought modeling and things like that so that's more something that is interesting for us than really 5 meter resolution image okay and do you already have something like no yeah so we are running since one

year in India with paying customers so we are currently 300 farmers covered in the platform and yeah and as I say so they are unbought that by our local partner that is the one who go in the field because we don't we are not b2c because we don't speak their language we not have their culture we are not in this situation in situ so that's very complex so that's why we rely on our local partner to go in the field to and also we are we have an Android app they can undock them directly in the field and when they come back home with the Internet in synchronize and etc so they are doing all the management the last mile and we are doing the the asset really the last

assessment and the back office of the instrument and they are doing the distribution ok I promise you that it will be shot [Laughter] [Applause]