“Without unnecessary experiments”, or how we launched a bank accelerator

    The well-known American venture capital fund and startup accelerator 500Startups once asked 100 managers of large corporations about their work with startups. It turned out that almost all companies work with startups, but only one out of four pilot projects is transformed into a solution that can be brought to the market. And if we talk about banks, and even less. In this post, using our own accelerator as an example, we will show what the difficulties are and how we try to overcome them.

    Collaboration with startups is much cheaper and less time-consuming compared to product development with the bank’s internal resources. So we have the opportunity to embed almost ready-made technologies in client services and internal business processes.

    For several years, VTB has been testing various formats for working with startups. Every day, various departments of the bank receive requests from start-ups offering their services. But in order for such an offer to find its customer within the huge corporate structure, it is necessary to go a difficult and long way. Accelerator is a platform for the bank’s interaction with the Fintech market, where startups and internal customers quickly find each other. When selecting technologies for the accelerator from a large flow of applications, we focus on the real tasks of the business. On the other hand, we offer business solutions that can encourage the creation of new products and services.

    At the time of the launch of the first VTB accelerator in June 2018, not everyone in the bank clearly understood what we need startups for. In order to find interested internal business customers for piloting start-up solutions, the task arose of involving a large number of bank divisions in the project and developing an internal culture for working with innovations. An important stage in this work was the study of the foreign experience of large corporations in working with innovations.

    The experience of colleagues from Singapore helped us (a separate post is dedicated to this trip) The largest Asian bank DBS in 2009 was considered the most bureaucratic and inconvenient bank for customers in the country. The name DBS was jokingly deciphered as damn bloody slow - "damn slow." Then the new COO Paul Cobban came to the bank. He spent the entire training budget for top managers on hackathons and forced tops to participate in them along with startups - so that bank managers plunged into the environment of startups and became infected with their entrepreneurial thinking. He launched a digital training program for all bank employees, opened an innovation laboratory, taught all agile methodologies, created corporate innovation programs, held lectures and special events with entrepreneurs, and made an alumni program. And in 2014, DBS was recognized as the best bank in the Asia-Pacific region,

    How we selected projects for the accelerator

    Our task is to increase the expertise within the bank on working with startups, create an infrastructure for testing and integrating solutions into the corporate environment. We decided to trust professionals who have extensive experience in scouting startups and organizing acceleration programs, are able to balance the corporation's business tasks and startup opportunities. To do this, we have engaged the GenerationS team as a partner. Large corporations, venture investors and thousands of startups from all regions of the country annually participate in GenerationS projects.

    We collected applications for three months, and as a result, 190 startups wanted to participate in our acceleratorfrom Russia, the CIS and Europe. The accelerator was gaining momentum from the inside: three months later, 150 internal VTB experts from more than 40 divisions worked with us and evaluated and selected technological solutions.

    At the selection stage, bank experts have already confirmed that they are ready to conduct pilots with selected startups. Of the 190 startups, 32 were selected - they scored the largest number of expert votes. We invited them to Demo Day. Previously, the GenerationS team helped them prepare for the presentation, formulate a value proposition, and describe competitive advantages.

    At Demo Day, we gave the teams 5 minutes each for a presentation, which was evaluated by experts and top managers of the bank. At this stage of the selection, we evaluated:

    • The presence of MVP with a clear business model and value proposition for the bank - so far we only work with mature projects that already have a finished product, technology or even sales experience
    • The relevance and potential of a startup for VTB business
    • Innovation and feasibility of the solution
    • Competitive advantages

    According to the results of Demo Day, 12 startups for piloting passed to the accelerator . The accelerator does not have a budget for experiments, any technology testing is funded by an internal business customer, so we worked hard on the success criteria, which will include both the technological and the business component. Additionally organized workshops and seminars for business development for teams, joint work with mentors and experts.

    Under strict corporate procedures aimed at minimizing any risks, the task was to create a fast track for the implementation of pilots. In the case of work with startups, the risks are high, so the generally accepted procedures do not work here. Having gone all the way from finding a startup to pilot implementation, we have identified bottlenecks in current business processes.

    To quickly deploy the infrastructure for pilots, we organized a sandbox based on an external cloud. It’s convenient in the cloud: you can use only the necessary amount of resources and quickly deploy test loops. In addition, long approvals are not required for testing, since we do not affect the “combat” systems and processes of the bank.

    But in the "cloud sandbox" you can test far from everything. Some accelerator projects require working with personal data of customers, integration with internal data sources and internal banking systems. So you still have to work out the issue of rapid deployment of internal test loops and access to the necessary infrastructure.

    Who hit the accelerator?

    We were carried away by the description of the processes, and you are probably interested to find out which startups got to the pilot. In the first set of accelerator there were 10 of them:

    • Data Fabric - a software platform for the collection, transformation, storage and management of data based on semantic technologies;
    • FreshDoc - a designer of documents using artificial intelligence;
    • Synpatic - technology for isolating and evaluating information from sounds and speech, as well as intonation analysis;
    • VOCA-TECH - personal audio badge and voice analytics for bank branches;
    • WantResult - technology for generating target customers;
    • Ziax - a smart robotic system for call centers with speech recognition;
    • AIST - analytical targeting information system based on GIS (geographic information system);
    • TheWaay - a solution to personalize bank relations with customers;
    • Octopus - data analysis and management platform;
    • Fabrique.ai - a platform for processing streaming data based on AI / ML algorithms

    Now let's talk more about some pilots. With the technology developed by the Synpatic team , we assessed the satisfaction of corporate clients in telephone conversations. They did this on the basis of the definition of “emotional color” of the votes of customers and employees of the call center. Synpatic can collect and analyze many useful metrics, for example, NPS (loyalty index), Negative-total ratio (ratio of total negative time to total recording time), CommonOperatorClientRatio (ratio of operator’s speech volume to client’s speech volume), ClientEndAgression (negative presence of the client in end of conversation). Using Synpatic, we can spend much less time manually processing call records, increase the objectivity of their evaluation, and as a result, the quality of customer service control.

    Another technological solution created by the startup, the AIST system . It collects, stores and analyzes internal and external data on urban infrastructure facilities, individuals and legal entities, and ultimately helps to manage the financial results of the bank's branch network, assess the risks of network transformation, increase sales and solve other tasks of the bank. We launched a pilot in Novosibirsk and in one of the districts of Moscow. Using geomarketing tools, AIST assessed the location of existing points of sale (financial and integrated efficiency), and determined the best and worst locations for placing points. In the future, scaling technology will help optimize the existing banking network and make more informed decisions for opening new points.

    Based on a software platformData Fabric as part of the pilot, an artificial intelligence system was created to analyze legal entities and formulate recommendations for changing financial behavior based on data from open sources and transactions. The service helps legal entities to look at themselves from the outside and evaluate their “trustworthiness” in the eyes of the Bank and the state. As part of the pilot, a unique reliability assessment algorithm was developed, which in the future can be used by both bank customers and new service users. With the help of the Accelerator, we not only opened up new opportunities for VTB development, but also identified “complex” places in corporate business processes, learned to interact with startups and, most importantly, began to form a new culture for working with innovations.

    In May, a new set of startups started at VTB Accelerator. We invite startups to apply online vtb.iidf.ru .

    Also popular now: