With each passing day, the need for digital transformation is getting more significant. Given the complexities behind it, a successful digital transformation is a critical task, but it has the potential to drastically reshape an industry or a business and gain many competitive advantages. Successful digital transformation is a priority for every business today, and to get a clearer picture of its roadmap and criticalities involved, Shekhar Jitkar, Senior Director, Quadrant Knowledge Solutions, interacted with Alain Veuve, CEO & Founder, Parashift AG. Apart from being the CEO of the machine learning company, Mr Veuve is also the Founder & Chairman of the Swiss digital tax consultant Account AG. He has strategically advised various global companies toward successful digital transformation and business undertakings.
Pre- & Post-COVID Digital Transformation
The conversation started with a discussion based on the pandemic, the war and the disruption in global trade. Shekhar pointed out to Alan how these events made business leaders rethink and replan their business strategies and asked him about how he thinks business leaders are evolving in their leadership roles and changing their business strategies toward successful digital transformation. Alan pointed out that he has exclusively seen a transformation era of pre- and post-COVID time. Before COVID, he experienced fewer initiatives and convictions from business executives toward digital transformation. The pandemic, Alan feels, came as a brute force to push these business executives into placing serious thought about investing in the digital transformation of their companies. They have now realized how critical it is to have a successful digital transformation strategy for sustainable growth.
Roadmap to digitalization
Alan then went on further, suggesting a possible roadmap toward a successful digital transformation in an organization. Alan highlighted a few steps that any organization can follow:
- The first step towards digital transformation is the strong conviction of the top executives and other associates to transform everything into a digital format. He highlighted that like electricity, which is a necessary tool in today’s time, the digital network should be considered of equal importance and business executives should take more initiatives into the overall digital transformation process and not in bits and pieces. Alan also feels that primarily it is important to not digitize existing processes but re-build the whole digital network from scratch.
- Secondly, companies starting with the whole digitization process should start small and not involve all departments of an organization at the same time. The cost of starting digital transformation should be less as that reduces the chances of error and complications.
- Lastly, companies starting with the digitization process of their system should leave about 10 to 20% part of their process before its actual realization. This part is the hardest and according to Alan, this is the part that can be left up to humans based on cost-benefit analysis.
Response to digitization
After this thorough guidance, Shekhar asked him his opinions on industry leaders’ responses to the recent changes bought by technologies like cloud, machine learning, and artificial intelligence in the software and the user industry. To this question, Alan jokes that he is apparently a kid in the cloud technology movement and if it was up to him, he will put all software programmes and files in the cloud system. He emphasizes that there are a few functions that are only limited to cloud technology and his team at Parashift combines data, use-case learning, tenant learning, cross-tenant learning and the use suitable ‘on-premise installations’ to upload the diverse data on cloud. Alan also states that even if it is challenging to work on cloud, it is doable and in recent times almost all software has some components of machine learning integrated to improve user experience. This reduces time the user spends on the software learning it and if these software’s are already coming with features that are automated beforehand, it is reducing the time spent by users learning it along with an increase in efficiency. This user experience can also be termed as cost of use, which is a separate term from license and operational computing cost and Alan pointed out that reason for digital transformation can be long but in the same line of thread connecting the facts that cloud computing enables machine learning which in turn enables low user experience time.
He gave the example of Parashift wherein they were having difficulty convincing banks and other financial institutions to convert into cloud space, but over the past six or seven months of the time these institutions have realized its importance and are now investing time into understanding why cloud computing is necessary and what benefits it offers. Alan gives the remark that most compliance officers in these institutions have to be given an in-depth knowledge of the need for digital transformation and are unaware of the developments happening in this area even when the IG teams of the same institutions are aware of the benefits and risks. Thus, in the near future, Alan feels that several banking institutions are going to transform into cloud computing, and with the available sensitive customer information in such a case, there will be risks involved, but ultimately, the automation and optimization rates in big industries will rise with the reduction in infrastructural complexity and that is the gain that people should focus on. He only hopes that over time machine learning will become more integrated into business operations and Alan gives some advice to software developers to build their products with a machine learning the first attitude.
Cloud computing user experience:
With this perfect observation, Shekhar continued the conversation by asking Alan whether the user experience has improved with these new technologies and what type of response the users have regarding the change. Alan remarks that this mostly depends on the comparison of things, which means people often underestimate progress in the long term and overestimate it in the short term.
From the outside, customers will not say that there has been significant improvement in machine learning over the last few years, but professionals in this field know the improvements that have occurred, even if they might not look the same for a user. However, if we look further back twenty years, the software has undergone significant change. Everyday items in a household can now be equipped with some form of software and comparing it to similar products twenty years back there has been a huge development. Alan thinks that soon the concept of software from a user’s point will disappear as most products will have this software integrated in them. There have been huge technological leaps in the field of automation, Robotic Process Automation (RPA) and document processing. Parashift is working in the same direction of building huge data networks that could eventually become Generative Pre-trained Transformer 3 for document extraction. In Alan’s opinion, building this software is a difficult and costly affair with many unknown variables, but even with the difficulties, there will be a huge lead in the next five to ten years.
Globalization & change in entrepreneurship
With this, Shekhar asked about Alan’s opinion and observation on the entrepreneurship opportunities and challenges in today’s competitive and volatile business environment. Alan replied that there are a few things he doesn’t like, especially the marked development in the IDP (intelligent data processing) sector. Process automation comes in a variety of colours and shades and for every use case and vertical, there is a niche solution out in the market. Even four years back the market was mostly occupied by legacy vendors doing rule-based templates following parsing, OCRDUMP. But nowadays, there are a lot of new players with increased development in the IDP sector helping educate decision-makers on the benefits and uses of IDP technology and that it should be among the toolsets of every enterprise company. Alan believes that in the technology sector, people are still in the education phase and every entrepreneur and his competitor is trying to make their industry topic more important than the rest. In the global scenario, industries in the IDP sector are lagging and Alan gives an example of Parashift where it was evaluated by a client alongside 12 other global vendors and they got the order from the client. Even with Parashift and many other companies in the same domain, catering to their clients globally, there is high evaluation because of differences in time zones, location, and culture. After this reasoning, Alan concludes the answer to the question by stating that in the short-term, the pace of globalization will be slow, even reverse sometimes, but in the long-term, international businesses will be open to the diversification of job distribution helping overall globalization. There will be a decrease in inequality and people from foreign countries will be open to more business and collaborations on a global scale.
Future technology trends
On the question of emerging technology trends that can be leveraged by society to benefit it, Alan highlighted that even though he is not a machine learning engineer and mostly into in-house consulting and strategy, the technology development in the field they are working on, mainly document swarm learning, is gaining traction in today’s fast-changing technology market. It is a blend of active learning and transfers learning at large and at Parashift, people are working towards mixing thousands of models working on individual tasks of document extraction with a zero-touch process. Here, Alan narrates the story of how Parashift was developed from an idea, where he wanted to send many documents through API without configuration and validation of results. Here, he realized that there was almost no such product in the market capable of doing such a thing. Even if there were a few products, it would depend on different capabilities, projects, etc. This led to Alan thinking of starting Parashift as an API first product which would have thousands of document types, ready-made, pre-trained and globally accessible for enterprises which would be available just by connecting with API. Even with multiple challenges and tasks, people in Parashift are constantly burning the midnight oil to form Parashift into making a company adapted to document swarm learning and available to all. Even with all the breakthroughs, there are always training and maintenance challenges involving thousands of documents which are costly as well. Irrespective of all these challenges people in Parashift are constantly trying to reduce the cost of their product by investing in making better technology that goes into their software.
Shinjini Sarkar is a Senior Content Specialist at Quadrant Knowledge Solutions.