What is our primary use case?
We are using it in two ways. One is by analyzing our current clients to create more business by deciding if we can offer them new products or if there is a risk of their leaving us or stopping use of our credit lines.
The second side is to prevent the risk of default. Our credit clients, because of the economic situation or internal decisions of the company, can go into default and stop paying their credit lines. We use it to prevent that risk. If we see a deterioration in a client, we can decide to stop lending money to the client and prevent risk in that way.
So on the one side it's to create or attract more clients by identifying certain trends or certain characteristics and offering them more products. And on the other side, it's to prevent the risk of credit default.
How has it helped my organization?
Both of our use cases are valuable. On one side it creates more business, but on the other side, it prevents impact on our company's equity. We have seen positive results on both sides, but especially on the risk prevention. Still, we have seen positive results in the generation of more information for our commercial team to attract clients as well. We have managed to reduce the percentage of high-risk clients from 8.9 percent to 5 percent. That's very significant in the overall quality of our clients. That's very good news for everyone, especially our investors.
Darwin is also helping us to make models faster and increasing the efficiency of the team. Instead of trying and trying to prove various models, we get a model fast. It's making data science faster but we would like it to be even faster still.
It has helped our team to convert data into knowledge and to more deeply analyze the results.
What is most valuable?
The solution helps with the automatic assessment of the quality of datasets, such as missing data points or incorrect data types.
What needs improvement?
We have used Darwin as a complement to other tools like R and SPSS to get the accuracy we want. This is one thing we've told the people at SparkCognition and they're working on it. In these kinds of situations, we don't use Darwin 100 percent because of that limitation.
The solution does help us towards making models operational but we are not at the place we want to be. We want models giving the answer to whether we should make a loan or not, but we are not at that point yet. It still has some limitations. These are things we have given as feedback and they're working on them.
Also, it would be great to have a solution that can organize the models. Right now, when there are a lot of models, they are disorganized. In the future, when we have more models, it will be more complicated to find the things that we are working on. It's about the user interface. You have the screen where you can look for other models but you can't organize models by name or by date.
Something they are working on, which is great, is to have an API that can access data directly from the source. Currently, we have to create a specific dataset for each model. But it would be great to have an API that gives us the opportunity to have a connection with our datasets or data lakes for each one, and a specific file for each model. Sometimes, you find that you have to add a new variable and you have to create a new file with that variable instead of having a connection via API to your datasets.
We have also asked SparkCognition that instead of automatic suggestions for addressing dataset issue, things should be defined by the user instead. There have been occasions where we have numerical data and Darwin has suggested using a nominal variable. We would prefer to define categories ourselves, instead of the recommendation that Darwin makes.
For how long have I used the solution?
We have been using Darwin for about six to seven months.
What do I think about the stability of the solution?
The stability of the platform has been okay. We've been using it regularly. Anytime we have a problem we call our contact and it's solved fast.
What do I think about the scalability of the solution?
We need to finish some internal changes and some internal processes so we can go to the next step and scale it. Our dream is for even our commercial team to start using Darwin — not making the models but using different applications. Our dream is to scale it to check all our possible prospects. We have work to do on our side on the scalability.
How are customer service and technical support?
Tech support has been good. When we report something or when we need help, they have responded really fast. We are usually in touch with a particular person. When we call him or tell him what our problem is, we immediately have our response.
Which solution did I use previously and why did I switch?
Our CEO started everything. He brought Darwin to us. We have also been using R and SPSS.
There are a lot of differences between each of them. In R you have to code, but you can do more things. The nice thing about Darwin is that you don't have to code. It's user-friendly. But there are some things that you cannot do with Darwin that you can do with R. It's a good solution for basic information. We are using more unsupervised models because we are using clusters. That's the reason we are using the other two solutions.
How was the initial setup?
The initial setup was complex because it was the first time we used this kind of AI tool. They visited us here in Mexico, which was good, and they prepared us. They explained what Darwin is, what we can do with it, some use cases, etc. Since then, they've been helping us and answering our questions and helping us with any issue we have. At first it might have been difficult because if it was something new, but we adapted really fast and started using it immediately.
Our deployment took about four or five weeks.
We didn't have a specific implementation plan other than just to begin using it and to give feedback to SparkCognition on any issues, recommendations, and thoughts about the product.
There were four or five people involved with us from SparkCognition at different points, people who came here or took part in video conferences. On our side, overall, there were between 12 and 15 people involved, including our BI and IT teams. But not all of them use Darwin. We have about five or six regular users. The rest got some general knowledge about Darwin and what we can do with it.
What about the implementation team?
At first, we did not work with a third-party. Right now we are working in an alliance with Xprtica. It's an arrangement that we have just started. We are ahead of them in terms of knowledge about Darwin so we're waiting for them to also get used to Darwin and to learn more about it so that they can start helping us with the different things we have in our plans.
What was our ROI?
We haven't seen a return on investment yet, but we can start calculating it regarding the clients we have detected who may fall into default. That's one of the things we want to do to determine if we've been effective or not in that prediction. And we also want to calculate what would have happened if those clients actually would have gone into default. What the impact would be on our equity.
Which other solutions did I evaluate?
We didn't test any other AI solutions.
What other advice do I have?
One of the most important things we learned, and that we also recommend to other companies, is to have a data link; to have all their data ready. Without data you cannot use Darwin. You really need the data to start using it and to take advantage of Darwin.
You also need people who understand data science. They can help you understand how to use Darwin and to interpret the results that it gives you.
Right now we are not measuring the accuracy of the models. We are using it to give some insight and some answers. We're on our way toward that.
Which deployment model are you using for this solution?