Translated by
2019/06/04 16:21:20

TAdviser Interview: The chief information officer of Alfa-Bank Sergey Polyakov About the future digitalization 2.0

The new Chief information officer Alfa-Bank Sergey Polyakov told in an interview to the Chief Editor TAdviser To Alexander Levashov about new — difficult and expensive — a stage of digitalization and the views of approaches to management of IT.

Content

Sergey <br /> <b> of Poles </b> <div> the Idea of bank digitalization in general consists in transferring clients to remote links of self-service </div>
Sergey
of Poles
the Idea of bank digitalization in general consists in transferring clients to remote links of self-service

About a role of IT in the strategy of Alfa-Bank

Recently the new strategy of Alfa-Bank was approved. In its basis three key directions of development: mobile technologies, failure from paper, digitalization of departments. It is stated that as a result of its implementation Alfa-Bank will service clients without passport, without papers, only in the presence of the smartphone. What calls are put by new strategy in front of the IT block?

Sergey Polyakov: Let's begin with paper. The Bezbumazhnost is a character and a criterion of level of a digital maturity, and not just a task, useful from the practical point of view. However the question of achievement of this purpose is connected by the most part not with IT, and with regulatory and legal requirements and practicians, well just with human habits. Technically it is not so difficult to make processes paperless - there are also databases which store images, and applications which will recognize and will transform images to texts. There are methods to protect electronic documents electronic similarity of "watermarks" or, say, a steganography. There are storage systems of unstructured data with a possibility of indexing and quick search. There are processing chains which use similar objects – all this can be made, it is a lot of work, but it is inzhenerno not really difficult. So a call not in technology.

As far as I see, now in different governmental bodies there is a process of digitalization, own obezbumazhivaniye, and probably, there think much of how to change requirements, legal regulations, regulation, approaches. It means what can happens the following: we invest in a bezbumazhnost proceeding from the current regulation and commonly accepted the practician now, and regulation and habits will exchange for few years, investment can be premature. Therefore in the solution of a problem of a bezbumazhnost strategy, how many tactics is important not so much: the general vector on elimination of paper documentation, but with focus and prioritization of processes in which there is already a clear regulatory, legal and behavioural picture. As a funny thing, in many sales points in the States, after input of the PIN code you will also ask to sign the check. You never know …

Whether you consider what work with clients will mainly overflow to the mobile canal?

Sergey Polyakov: Work of bank with me personally precisely overflowed completely.

The idea of bank digitalization in general consists in transferring clients to remote links of self-service. Mobile bank — one of such channels. Good, in any case, by sight, mobile banks are not done today only by lazy. I use many of them, all have some features, prikolchik, but generally differences cosmetic. Today, banking mobile application is a mature product and even if you are not able to do mobile applications, you can address to the external organization which to you will construct it.

Value from the point of view of the client moves from the magnificent interface towards reliability of work and an opportunity to execute more and more wide range of transactions. And it is already a merit not so much of mobile application, and all that stands behind it – bank middle and a back office.

On the first wave of digitalization before banks there was a task to transfer basic one-step transaction relationship with the client to mobile applications — to look at the statement, to transfer money. Came, made, left. Everything was simply and quickly, edzhayl-commands baked the microservices supporting similar functions as pies, time-to-market was calculated for hours, all rejoiced and thought that they so will be always.

Then simple "tsifroviziruyemy" interactions ended, and everything became unexpectedly difficult. We want to do more interesting things in mobile bank, to receive services which on the entity in a second not to perform — we will tell, mortgage to receive. We want that the bank guessed our requirements, thought for us. Similar problems are not solved in mobile application, it is solved in deep layers of technology and operational infrastructure. Somewhere in a subsoil of middle office many elements and processes, plus, perhaps should interact, it is necessary integration with external resources — with state authentication tax administration TRAFFIC POLICE bureau of credit histories. And all these deep processes should issue you result in a type "yes", "no", or still somehow, like "yes, and still here that we for you have, if you here …" in mobile application.

In terms of mobile application everything is simple — I asked a question and received the answer, it is not more difficult, than balance enquiry on the account. But unlike service of obtaining balance which several programmers can make in two weeks in the form of microservice the advanced query starts twisting process in a subsoil of banking systems which a snake can creep for hours and days, to come to a surface in the form of additional customer interaction in the same mobile, and maybe a call, and tens or even hundreds of people work on it already, and many systems are involved. Whether these systems are ready to such interactions and integration — and there are the central issue and a call of the second phase of digitalization.

In department the similar difficult service is rendered by the operator or the client manager — this person also is your "technology" of integration. He can come into one application, into another, to write something on the sheet to paper, to enter to the third application, and here you received the answer. When the digital bank works for you, this process should be end-to-end, automatic, it should not be attached neither to the person, nor to paper (you remember about a bezbumazhnost?). And the result should be exposed through mobile application. Mobile application is only an iceberg top, all the rest inside. Reformatting of the internal, deep systems for such types of tasks, problems of the second phase of digitalization, is also our radical task, our call. It is what we will work the next two-three years on.

What functions will enrich mobile bank in the future taking into account that you will overwork, you proapgreydit middle-infrastructure and will make what you are going to make? In what will key differences of that channel of self-service from the current channels there be?

Sergey Polyakov: I do not know it. And such good "I do not know this". I do not want to know it because my task as head of the block of technologies – to construct the scheme, architecture which allows to solve the whole classes of tasks, but not to find separate judgments for momentary problems. Therefore I should know not about future functions of applications, and about classes of the forthcoming tasks, and their display to the interacting technology stacks. Let's tell, between the moment when something comes to the head of the business line to mind, and the moment when we are able to implement it, there would pass some reasonable, limited time. Let's tell, three months. But if I tell it "forgive, but on it we are not ground, we should build at first the bridge between the data storage module, the credit module and compliance system, and it, this system, refaktoritsya sorry, therefore it will be necessary to wait half a year more", so I did not complete the work — technology stacks should be not only are ready, but also are ready to interact with each other.

Danger of your question here in what. If I know what specifically functions business wants to implement, I have a temptation to begin to build solutions under these functions, very quickly, but is very specific. It quite often happens to edzhayl-commands. But I do not want to be in this situation because then I will be stuck in that the most monumental task of cross-system integration about which we already spoke. Do not misunderstand me — at IT naturally the type of tasks bank will eat what understanding to solve the next several years in what direction business lines will move. For example, we will develop a mortgage and cash loans, we will improve scoring, we will improve the system of recognition of the client for secondary offers to make them more relevant, to monetize data, to obezbumazhivatsya. Amount of the different ideas on the party of business huge — but it is impossible to form technologies specifically under everyone and therefore I look not at tasks, and on classes of tasks, and I build technology architecture under classes too, but not under specific objectives.

The third direction of strategy — digitalization of departments. In what the main calls here?

Sergey Polyakov: It is other party of the same coin. If in the beginning we spoke about remote links of self-service then we speak about approximate channels of service — through operators in departments, well and on phone.

Both the theater begins with clothes, and bank department begins with arrangement of digital queue. When the client came to department, at first it needs to be identified, then, having identified, to obtain the additional information — as obvious (demograficheky and legal data), and enriched – all that will allow us to create the appetizing, personalized offer, and even two or three, for that time that the client paces from a threshold to a window. Well we will also not forget, it is necessary to make that why the client actually also came to bank department — to render him service.

These tasks — to identify, know better, to find out why came, to offer, to execute — can be executed by different IT systems. But ideally all these functions should be reduced in the uniform information and operational ecosystem feeding the employee of the bank interacting with the client in real time. This system should prompt ready offers, allow to perform right there them. The similar multimodal uniform workplace can be called differently. Someone calls it the uniform front office, someone — an expanded CRM system. We far promoted on the way to such expanded CRM, I think, in the next year we will complete its creation, synchronously with the strategy of bank on creation of network of new generation.

How many applications are used by your operator now?

Sergey Polyakov: Two main, one of which is connected with CRM, with demography and a personal straight line and enriched information, and another – operational. More and more functions of the operational application remove to demographic, to CRM ++, our United front. Why all these functions cannot be taken also out to other canals, for example, in ATMs? I do not know. For me, for example, a riddle, why the screen of the ATM such what it is. It is some remnant of the past. Why it does not look as the screen in your mobile application? Same all functions. I think, it should exchange.

About outsourcing and insourcing, target and inappropriate platforms

Different banks differently build contractor relations, with vendors, about developer outsourcing, and differently create internal competence of the IT block. What your approach? Who writes your uniform front office?

Sergey Polyakov: Where it is possible, we use ready technologies and ready applications. And we do it with pleasure. We are able to integrate them and we will continue to do it. I have no hard line concerning what it is necessary to do only by external forces and that — only internal.

What principle is applied when choosing this or that model?

Sergey Polyakov: You know, the market Russian technology workers so hard that you can come with any principles, but reality right there will make sober you. And when you will tell: "I want to employ 100 programmers on Java, on Pega, on Camunda, on SAS, on Scala", the market will laugh and will answer: "it is absent anybody and will not be in the next year — or teach people independently, or here the contractor who is ready to offer you these specialists".

If you remember, SberTech a few years ago took and skhantit developers at the contractors …

Sergey Polyakov: In my opinion, such thing can be made once. Next time or at you in the agreement something about it will be written, or the contractor will not work with you any more. So it is impossible to do business.

Now we freely stir teams of the employees and specialists from contractors, but the general principle such: if it is the target platform, then we would prefer to employ people to the state. If it is the inappropriate platform, we would prefer to have the external counteractor who supports it. And it is not important for us whether the mission critical platform is. We have inappropriate things which, nevertheless, mission critical.

For example?

Sergey Polyakov: For example, one of the platforms used now in departments. Finally we will leave it therefore it is inappropriate. Still we have one of mobile banks which is quite widely used, but which inappropriate. It is good, it perfectly works and if it does not work, then there will be a discontent at a large number of people. But finally we will transfer it to function to our standard mobile platform.

So the critical systems can quite be developed and supported by external forces?

Sergey Polyakov: Yes — if there is an examination if contractors know the case why and is not present.

About problems of digitalization 2.0 and reform of the IT block

Until recently in bank there was a selected technology division Alpha laboratory which as I understand, was defunct and crept away on business blocks. How now management of informatization is arranged? What is the IT block engaged in? How do you interact with business blocks?

Sergey Polyakov: "Alpha laboratory" worked by the principle of "client ways" (or customer journey), it is a certain form of the organization of development. There is no Alpha lab now, but approach to building of new functionality for the end customer and for business remained. The most part of business functionality will be made using similar approach, it is a part of strategy of bank, and this approach likely is today de facto the standard for the advanced organizations. But there is on this way and a danger – the IT organization "from the client" but not "from a technology platform" can lead to crushing of technology platforms.

Usually each client way is agglomerate from several edzhayl-commands on 8-9 people who work on some wide subject, for example, a mortgage or salary projects. If these commands do not look at each other and do not follow any general architectural view, then absolutely easily technology stacks can creep away.

One of reasons for criticism of Alpha laboratory was the fact that it often implemented technologies which were required in the moment for execution of specific objectives. As a result, it led to extraordinary fast accumulation of functionality from the point of view of the client, but left the whole field of wonders from technological solutions.

In the beginning we said conversations that on the first phase of digitalization rather simple functionality connected with fast one-step transactions and communications between clients and bank is implemented. But on the second phase it is already necessary to integrate difficult processes, and it already occurs in middle-and back office. At this stage the cabinet of curiosities of technologies, prototypes, MVP becomes not just expensive entertainment, but also begins to break for the sake of what we went in edzhayl — Time-to-market.

With similar difficulties, I assume, practically all organizations deal. At my many colleagues who manage technology departments a similar impression — the breakthrough using flexible methods of development left us with a large number of technology semi-solutions, MVP which it is necessary somehow to integrate and force to interact now, and it long, expensively, and the main thing, is absolutely unexpected for the customer.

The axiomatic offer on the solution of this problem which assumes involvement of the strong architect not really works because any architect cannot tell the customer: "sorry, the release of your service will be delayed for half a year because it breaks my architectural principles". A certain contract between needs of IT, long-term needs of business and what is necessary in the moment is necessary.

Certainly, it is necessary to have the good architectural drawing, but not dogmatic, but practical. Besides, it is necessary to understand that the problem of creeping of technologies is also a problem of project management. If you or in workshop distribute to all crews on building the drawing, and they on it will begin to work, then each of these crews interprets the drawing in own way. Therefore on building, in addition to the architect, there is also a foreman, and in workshop there is a master which monitor that interpretation of the drawing was identical.

Business divisions understand problem depth?

Sergey Polyakov: Yes, understand. Digitalization 1.0 went approximately from 2006 to 2012-2013. Quickly enough in the Agile mode there were technology artifacts, mobile banks, Internet-applications propagated. But then suddenly unexpectedly everything slowed down, and IT managers began to say: "oh, you know, works here for two years, and here for two million dollars, and here it is necessary to purchase licenses …". But how? So it was good still quite recently! What happened? And there was what was produced ingredients, technology stacks, and the following class of solutions of digitalization 2.0 already requires internal integration. Integration is a long, unpleasant and dreary thing.

What in the context of the designated problem will occur in IT of Alfa-Bank?

Sergey Polyakov: We will continue to work, naturally, on Agile-model and with client ways. It is really very effective method to deliver functionality to end consumers. But at the same time we return the idea of a platformennost. We designate target platforms and technologies with which we will live. We will reorganize IT, creating functions and processes which in many cases will be grouped not around business process, and even not the business customer, and around the most technology artifact — the stack of technologies, the platform. The development movable by the interests of owners of a product will be balanced with long-term requirements to health and efficiency of IT infrastructure. It can seem to someone old-fashioned, but, in my opinion, it is good and sober approach which will allow us to continue to create quickly value for business today, saving compact, coherent technology construction and rational economy.

What do you mean under technology artifacts?

Sergey Polyakov: Let's assume, I have at least three technology platforms for creation of mobile banks now. And several mobile banks work at different technology platforms.

Different mobile banks for different audiences?

Sergey Polyakov: Yes, different commands worked on them, they at fast speed, sometimes within three months, created this or that functionality for different business divisions. The financial service could have a question — "why there is a lot of you and why you cost so much?". It was possible to answer this question so: "you know, it is inexpensive in comparison with that how many money we earn using these applications", and till certain time this answer of all arranged.

But then there were more difficult tasks — will tell, to connect salary projects of legal entities with individuals which work for these legal entities. The same business divisions are expected fast solutions, but already quickly cannot make it IT and block, it is necessary to cross the technologies which grew in isolation from each other.

If we do not want to get to such traps, we should integrate applications on a single platform, on the uniform architectural concept. Then I will be able to move people from one project on another, from one command to another. I will be able effectively to manage the resources, I should not train 100 people in 100 technologies, and then we will return to high performance again.

And therefore in IT strategy we say about a platformennost, and that in addition to client ways as a method of creation of the user functionality, we will grew and enlarged platforms. We will monitor that our solutions which we will do on client ways stood on the large supported target platforms which will become number less. And we will refuse some platforms which will not be necessary to us.

For implementation of this strategy we need the architectural picture and professionalizing of project management. The architecture already is, I inherited it from previous leaders of IT, and it is pleasant to me, by the way, I do not see the reason something strongly to change.

What do you mean, speaking about professionalizing of project management?

Sergey Polyakov: When I say that the IT of Alfa-Bank needs professionalizing, I do not mean that here nonprofessional people work. Level of competence and talent reads off scale here. I do not joke, I a lot of things saw. But if, for example, you are well trained fighters, and you will be delivered in a caret, and will order not to descend from the place, I pound from your abilities and talents any will not be. I want to place people so and to manage them so that to take the maximum potential which in them is put

IT in Alfa-Bank very much accelerated in terms of delivery of final functionality, reached extraordinary flexibility in terms of delivery of this functionality, in particular, due to dispersion of project control and easing of architectural requirements. IT also strongly grew, and now it is necessary to work on management is always cyclic process, time to be focused came now. This approach is congruent general strategy of bank — to do less, but the main thing. We will see the first results this year, and system changes will take roots during few years.

On what platforms you are going to be focused, first of all?

Sergey Polyakov: I will not call an abbreviation because names are not really important. In fact, we will have a single target platform for all mobile applications. It is already selected, it is absolutely modern, there are all necessary words — microservices, containers, dockers, management of containers. A task — to transfer to it all applications and not to create applications on other platforms.

We have a single platform for employees of departments which should be transmodal, integrate both CRM, and transaction, we spoke about it already a little — in the work with the help of the uniform application the client manager should have an opportunity both identify the client and understand it and sell it a product and execute its request.

Everything that the client does in the mobile application, in principle, should appear both at the operator, and in contact center. The running term for this purpose — "omnichannel". It is a hard task not because all functions should be provided and in general are identical in all canals of customer interaction and because process which began in one channel can then get over in another. You began to do something in mobile bank, and finished using the operator. Or, on the contrary, the operator began to do something, and you asked to confirm something in your mobile application.

Provide omnichannel not easy, under all channels there has to be uniform infrastructure of middle office which remembers what the client made there, there, and there. From there are problems of creation of the general CRM, the general layer of storage of a status, the general reference books, and these are expensive tasks.

What number of IT in Alfa-Bank?

Sergey Polyakov: The general resource base on which we rely is about 3000 people. Part of them works in the state, a part — for contractors. Detailing is not really important. Something is given outside, something is transferred then inside, and it is not really essential to me.

This digit includes IT specialists who work in business divisions? If I correctly understand, Alpha laboratory crept away...

Sergey Polyakov: It was already slipped back, in other structural divisions. There is, certainly, an IT enclaves which are located in business divisions — in investment bank, in a processing center. But in the main IT it is consolidated. Total number of 3000 people is approximately halved between maintenance and development. In the direction of development, I think, we will come that about a half of people will work in "the client ways", and other half will be connected with platform, deeper developments. It is, probably, some version of two-speed IT (a bimodalnost or 2-speed IT model — a comment of TAdviser). There are organizations which say that two-speed IT is dead and so it is not necessary to work that all organization should be Agile. I do not trust in it.

Why you do not trust?

Sergey Polyakov: The Agile-organization is that which uses Agile-approach up to the depth of a technology stack. Results, in my opinion, are ambiguous. In particular, creeping and reproduction of technological solutions. I want to return IT platformennost. I want that platforms became less. I want that they were enlarged. I want that their development went a little bit on other process. And I want that IT artifacts had IT owners, but not business that there was a person from IT who would be responsible for the fate of this or that platform.

These are the same relations, as between the construction architect and the customer. He builds the house, puts a floor, lays pipes, and you to it at the end: "and I want so, and so". At this moment someone should tell: "yes, but here what will be". At me as the technologist, certainly, first of all, has a responsibility to the customer. But there is also a certain metaphysical responsibility to a product. I cannot make an engineering and nasty thing.

Otherwise the house will fail?

Sergey Polyakov: Perhaps the house will also not fail, but the customer finally will be dissatisfied. You will not be glad to your solution, you just do not understand all its effects yet. Therefore it is important to find balance and understanding. The person at restaurant can ask: "give me omelet, I want in a minute". It can answer: "please, or eat a crude dish, or it is necessary to wait". I as the good cook, I refuse to issue a half-cooked dish even if the client demand, I have a responsibility to my product.

With omelet clear, all have a personal experience. And here with microservice infrastructures of such experience at the customer, as a rule, is not present. Therefore I should explain why it is sometimes better to spend time, money, efforts to platform development in order that then not to get poisoned with this thing, continuing omletny analogy.

About clouds and DPCs

What your approaches to development of computing infrastructure, telecom infrastructure? Whether there are perspectives of its output in external clouds?

Sergey Polyakov: The cloud is a fashionable word, it came to us to life by efforts of many, it is most known likely to Amazon, and to it there is a place. But it is necessary to understand here that. The cloud purpose — the infinite stepless scaling of computing infrastructure following growth rather simple, homogeneous business of service like social network, online store, etc. Business and IT models of such organizations are fixedly studied now, are rather afraid of them supposedly now they in finance will rush, their methods undertake an example. Colleagues from some banks go to Silicon Valley, look as Facebook or Google works, admire infrastructure. It is healthy, but at the same time it is necessary to understand that the universal bank is much more difficult thing, than Facebook, Google or "Yandex". The complexity of processes is incomparable. Finally, the same Facebook is in principle one very advanced digital service, decisive, with IT of the point of view, a problem of scaling on one billion users. Tasks universal banks absolutely others, structure of service and a product it is incommensurable more difficult, and just accumulation of cloud capacities does not solve our problems.

It is a question of reliability?

Sergey Polyakov: In particular. Is not present in Facebook of a problem of reliability, to our measures. And in Google is not present — well you do not receive the message from any friend or will not see it in a tape — and what with you will be? Or you will see different contents of the film from different devices? Or to you some message — and what's next will come to mail twice? What happens to you? And here if you have a different account balance in phone and the ATM, or salary came twice — here you will start …

Approach of cloud scaling is used by the Internet companies which quickly grow and collapse. The bank seldom has a problem of fast scaling of some uniform service. We have diverse services, and we have difficult tasks of integration. I am not sure that in a cloud it is solved well, quickly and conveniently. Check hypotheses, test applications, experiment is please.

The second complexity with clouds are rules of storage of personal information. As the industrial solution for data processing of external clients it is unacceptable also by the legislation.

It does not mean that we reject cloud computing as such. In reality if to dig, many "cloud" organizations actually just use cloud stylistics, a cloud paradigm for management of own infrastructure. This reasonably useful occupation, it helps to smooth function of growth of infrastructure, to use it more effectively, but it "we in a cloud" not to call absolutely fair. We widely apply solutions of cloud type for management of internal infrastructure, but about overcast we do not blow, we quite on the earth.

Whether consolidation of own DPCs of Alfa-Bank is planned?

Sergey Polyakov: Everything is driven by economy. If the project on consolidation of DPCs makes sense, then we consolidate. Unlike, for example, exchange technologies where all external infrastructure of market participants is tied to the location of DPC and to move DPCs very hard, with bank DPCs it is simpler — none of clients put the high-frequency trade robot in our DPC. We have no specific plans when it is required — are consolidated.