Chandramauli Chaudhuri leads the Data Science initiatives across Fractal’s Tech Media & Telecom vertical in the UK & Europe. He works in close collaboration with senior business stakeholders and CXO teams across some of the leading global enterprises, enabling the development of long-term strategic AI solutions.
Being in the field of Artificial Intelligence and Machine Learning for close to a decade and working across a wide range of industries, his primary area of interest lies in R&D, algorithmic customisation, capability enhancement, and MLOps deployments of solutions. Analytics India Magazine interviewed Chandramauli to gain insights into AI transformation at the enterprise level.
AIM: What are the key factors when it comes to driving successful AI transformation for an organisation? What are the emerging AI trends enterprises are capitalising on?
Chandramauli: As a business leader driving AI transformation across an organisation, it is critical to understand that Artificial Intelligence is just the means of value realisation and not an end goal by itself. Thus, the factors differentiating success and failure lie in its synergy with the company’s core principles, value proposition and customer-centricity. AI adoption is not a plug-and-play solution that yields overnight returns. Businesses need to think beyond just the cutting-edge software, high-end infrastructure and skilled coders. Alignment of the company’s culture, customer expectations and ways of working to support such transformations need to take equal if not greater importance. The companies that are doing well, especially in banking, finance, media, telecom, and tech, are those that have integrated AI into their day-to-day functions. They are moving it away from being a siloed and ‘specialised’ initiative undertaken in small pockets, to broader cross-functional collaboration.
As far as emerging trends are concerned, organisations have started focusing a lot more on two key areas – execution excellence and risk management. This means nurturing an agile mindset across teams, pursuing the right use cases, developing a strong data foundation, investing in the right skills, and having a robust strategic roadmap. There has also been growing acknowledgement of the challenges associated with cybersecurity, user privacy, and digital consent. Issues like lack of explanations, absence of audit trails and presence of bias in AI systems have gained far greater prominence from the global community in the last couple of years than in the past decade. It’s true that we still have a long way to go and yet to fully appreciate the complex socio-political and economic implications. However, we have started looking in the right direction, focusing on building greater transparency and trust. The early adopters of these practices stand to reap the rewards in both the short and the longer term.
AIM: Where does the industry stand in terms of scaling AI solutions? How can AI/ML become a differentiating factor for companies?
Chandramauli: AI opens new frontiers to solving real-world problems. We have already seen some great examples of AI powering decisions in almost every domain, from climate change to the choice of songs. Add to this, the ability to augment with new-age technologies like 5G, IoT, AR, VR, etc., and scale through cutting-edge hardware, open-source architectures and cloud computing – what we gain is the prowess to redefine the limits of end-user personalisation and engagement. The resulting increase in efficiency, effectiveness and productivity has a direct impact on the top and bottom line. It is redefining the way successful businesses look at their strategic and operating models. Naturally, there is a lot of excitement around the future and a rush to seize any competitive advantage.
But, this is only one side of the story. For the vast majority of companies that are trying to drive innovation and scale their AI operations, progress has not been at the scale or pace that people might assume. Many are still struggling to push past the pilot and experimentation phases. This is mainly because, the traditional mindsets, legacy technologies and ways of working run counter to that needed for a full-scale AI transformation. Studies suggest that, currently, only about 8-10% of the firms engage in practices that support widespread adoption. For the remaining ones, enabling AI transformation at scale still very much remains a work in progress.
AIM: How does AI help organisations add value and manage risks?
Chandramauli: Large enterprises need thousands of decisions to be made every single day. Doing so effectively, requires a combination of process automation, contextual insights and cognitive engagement. Not only that, shifts in the industry landscape can trigger quick changes in customer behaviour, rendering past insights completely useless. Such dynamicity poses a significant challenge for the traditional software and analytical solutions, which depend largely on pre-defined logic and set patterns. The COVID-pandemic has been a reality check for many organisations in this regard.
AI, on the other hand, is particularly well suited to address such real-life challenges. Consider a business problem like content piracy as an example. Despite decades of effort, it continues to be a perpetual problem in the media and entertainment industry. It causes billions of dollars in losses every year. The volume of pirated content across the globe, in fact, reached record highs during the lockdowns. This is because, the consumers and distributors of pirated content are continuously changing, updating and scaling their operations beyond the limits of traditional anti-piracy systems. AI-powered solutions, learning and re-calibrating in real-time from online trends, network logs and consumption data feeds, can be much more effective at identifying and mitigating such behaviour.
AIM: What are the things organisations should keep in mind while building an AI roadmap?
Chandramauli: To build successful AI roadmaps, leaders need to devote attention to four key aspects.
First, define a strong and clear narrative, that explains to everyone what AI is and why it is so critical to the future of the organisation. This needs to start from the top – the board and the executive team, along with the key decision-makers including managers and team leads.
Second, dedicate time and effort to address the unique barriers when it comes to such fundamental shifts – apprehensions of the workforce about becoming obsolete, difficulties in adopting the agile ways of working, etc. The leadership team has to provide a vision that brings everyone together and shows how they fit into a new AI-driven culture.
Third, budget for integration, adoption, and training. This must be over and above that allocated for merely procuring the technology and infrastructure. Most successful companies, spend more than half of their analytics budgets on activities that drive adoption, such as workflow redesign, communication, and training.
And finally, identify and prioritise the right use cases. This means striking the right balance between feasibility, investment, time and value. While the focus must not be restricted to just gaining quick wins, undertaking initiatives that may not get deployed or provides no returns in sight can severely jeopardise both current and future AI prospects.
AIM: How do you choose relevant AI use cases that will deliver maximum impact for your organisation?
Chandramauli: This is a particularly important question. Depending on where the organisation is in terms of its analytics maturity, the approach toward choosing the right use cases needs to change.
For enterprises that are just embarking on their AI journeys, it is better to start with a few well-recognised and well-understood business problems. The primary focus should be on learning, identifying the gaps and gaining experience. The solutions need to be technically feasible within the current organisational constraints. Most importantly, while these initial use cases do not necessarily need to drive the biggest monetary benefit, the impact must be measurable through some established business KPIs.
On the other hand, for organisations that are further along the path, factors like potential returns, time to value, cost of deployment, infrastructure availability, training requirements, etc., need to be accounted for. A good practice is to start by collaborating across cross-functional teams, including people from strategy, analytics, IT and operations sides, to understand the current needs and state of the business. The aim should be to arrive at a set of clearly defined annual objectives and accordingly prioritise the execution of use cases at a monthly or quarterly level which can help achieve the same.
AIM: At what stage of the project should companies ideally look for AI-based digital transformation?
Chandramauli: Whether it is the B2C or B2B space, digital transformation is about streamlining and improving customer experience. AI aids such transformations by identifying the areas that can maximise value. Thus, generally speaking, digital strategy should move hand-in-hand with AI transformation, right from day 1. It allows teams to be better informed, focus on the important processes and achieve better results.
However, if we look at some of the larger organisations, this is not necessarily the case. In such situations, it is worth taking a step back and first assessing the need, effort, cost, time and value of undertaking an integrated digital-AI strategy. The key here is not to force AI into the mix just because other companies are doing so. This may be counter-productive, disrupting the normal functioning of the business.
A more prudent approach would be to decide on a case by case basis, and depending on the company’s and customers’ needs, implement AI as required. Gradually, as the capability matures over time, the business can start focusing on bringing the two together under a single roof.
AIM: Should companies invest in AI solutions developed in-house or hire third-party vendors?
Chandramauli: This depends largely on the organisation’s needs and current level of maturity.
Partnering with the right vendors has its advantages, especially in the early stages of AI adoption. These organisations bring years of specialised experience, learning, and innovation to the table. Outsourcing to partners with such deep technical knowledge can drive much-needed momentum and build the foundation for long-term AI success. It can also lower costs and mitigate long-term risks.
For organisations that are more mature and have dedicated teams of data scientists, partnerships and collaborations at some level still make perfect sense. Many such companies opt for a hybrid approach – build some solutions in-house, licence or buy some SaaS platforms and outsource the rest to other partners. It provides the flexibility to quickly scale efforts up or down without deprioritising critical objectives. Most importantly, it brings in fresh perspectives and cross-industry learnings which can be invaluable in terms of driving innovation and building competitive advantage.
