BizOps: realizing business value from data! In data and AI, organizations are increasingly challenged to harness these technologies to drive business value. In this interview, we speak with Somil Gupta, Founder and CEO of Algorithmic Scale and Co-Founder of Cerebellum Platforms! Somil brings experience in bridging the gap between technical implementation and strategic business outcomes.
In his upcoming presentation, this time at the Data 2030 Summit 2024, Gupta will delve into BizOps, an approach designed to enhance operational adaptability and resilience in the face of complexity and uncertainty. Join us as we explore Somil’s insights on leveraging real-time data, automation, and AI to drive operational efficiency and decision-making within organizations!
Hyperight: Can you tell us more about yourself and your organization? What are your professional background and current working focus?
Somil Gupta: I am the founder and CEO of Algorithmic Scale and Co-founder of Cerebellum Platforms AB. With a background in engineering and industrial management, I have had diverse roles in product management and business development. For the last five years, I have been working to bring the two worlds together and focus on realizing the business value from data and AI. My efforts have been centered on delivering the promised business outcomes and ROI from Data and AI investments.
Most people only focus on business value at the beginning of the project. And it is often forgotten later on during the technical implementation. However, our approach focuses on value throughout the project. We have created frameworks that enable us to prioritize everything by value – be it use cases, data, models, platforms, operating models, etc. To that end, we strive to find viable Data/AI use cases and work out the business case and potential outcomes. This requires adapting the business functions and processes to develop new products, services, and business and operating models that are Data and AI-compatible. Sometimes, I work on AI literacy and tactical AI skills as part of training initiatives. I also collaborate with development partners on Enterprise Data and AI strategy and platforms.
Hyperight: During the Data 2030 Summit 2024, you will share more on BizOps and realizing business value from data. What can the delegates at the event expect from your presentation?
Somil Gupta: The environment and business landscape is changing very rapidly. There is a lot of dynamic complexity affecting the business operations. The current enterprise tools, processes, and ways of working are nowhere sufficient to deal with the disruptions that frontline units and business leaders experience. The current ‘predict and plan’ systems and processes are too slow and rigid to respond to operational complexity. Adaptability and resilience are the most coveted and critical abilities. So, naturally, there is much more expectation from data and AI technologies to help.
In my experience, the current state-of-the-art in data and AI, except perhaps Generative AI, is not sufficient. It does not provide the level of adaptive decision support needed for high transaction volume, high uncertainty, and time-critical situations. In my talk, I want to introduce BizOps as an alternative approach to solving operational challenges and building adaptability. I will break down BizOps into its core components to discuss the new set of expectations and requirements from Data and AI. Finally, I will highlight some gaps in our current approaches. I will also suggest a roadmap for where data and AI need to be by 2030 to make an impact on business performance and outcomes.
Hyperight: Can you explain how BizOps helps bridge the gap between data-driven insights and operational execution?
Somil Gupta: BizOps bridges the gap between data-driven insights and operational execution. With this, BizOps enables organizations to make high-quality, context-aware decisions rapidly, even under conditions of uncertainty. Unlike traditional decision-making, which often relies on complete information and static reports, BizOps uses advanced tools and automated decision frameworks. It integrates real-time data and embeds analytical insights into operational processes. This allows businesses to adapt their strategies based on the latest information. Thus, making it possible to navigate complex, fast-changing environments efficiently. Companies like LinkedIn, Yahoo, Mastercard, and Amazon have used BizOps to transform their operational execution. They have leveraged data and AI to optimize everything from customer interactions to resource allocation and risk management.
By ‘engineering’ decision frameworks that incorporate AI, machine learning, and dynamic feedback loops, BizOps teams empower businesses to not only react to changes but also anticipate and shape them. These teams play a crucial role in ensuring data insights are not just consumed. They ensure the insights are quickly acted upon in a way that aligns with business goals and outcomes. BizOps reduces the lag between information and action, enhancing operational intelligence, and resilience, and driving smarter, faster decisions.
Hyperight: How does BizOps leverage real-time, contextual intelligence to enhance operational efficiency and agility within a business?
Somil Gupta: BizOps leverages real-time, contextual intelligence to enhance operational efficiency and agility. It integrates and embeds advanced data analytics and automated decision-making frameworks directly into business operations. By accessing and analyzing real-time data, BizOps teams can generate powerful insights into current conditions, customer behaviors, and market trends. They can also quickly assess the business impact of new events on critical customers and operations and develop the next-best actions. These insights are contextualized within the business environment, allowing for quick adjustments in strategy and operations. This capability enables organizations to respond proactively to changes and optimize resource allocation. It also streamlines processes, ensuring agility and minimizing the time between insight and action.
Furthermore, BizOps frameworks use feedback loops to refine decision models based on outcomes and new data. This allows businesses to react to real-time events and anticipate and adapt to future challenges and opportunities. The integration of real-time contextual intelligence into day-to-day operations helps organizations maintain high performance and operational resilience, even in complex and volatile environments. This proactive, data-driven approach to operational management is key to maintaining a competitive edge and driving business growth.
Hyperight: In your experience, what role does automation play in the success of a BizOps strategy, and how can organizations effectively implement it?
Somil Gupta: Automation is a critical component of BizOps and it is a very broad spectrum. It ranges from simple rule-based automation tools to RPA to sophisticated bots and agents that execute decisions in real-time. The complexity of the process determines what type, scope, and level of automation is needed. BizOps teams need to trigger automated and coordinated execution of complex decisions and workflows based on real-time information and insights. Automated systems continuously monitor key metrics, trigger alerts, and initiate actions without manual intervention, enabling faster, data-driven responses.
To implement automation effectively, organizations should start by identifying low-complexity business outcomes. These outcomes often contain repetitive, rule-based tasks and processes such as data extraction, report generation, and basic decision-making workflows. Next, they should integrate these automated processes with existing BizOps frameworks, ensuring they are aligned with business goals and can adapt to changing conditions. The change management around working alongside automated systems is also critical for maximizing the impact of automation on BizOps success.
Hyperight: What do you believe is the most significant challenge organizations face when trying to integrate BizOps into their existing workflows, and how can they overcome it?
Somil Gupta: The most significant challenge organizations face in introducing BizOps into their workflows is over-reliance on planning processes and systems like ERPs to drive operations. A disproportionate share of Data/AI investment, resources, and management time and effort goes into trying to predict the future and optimize plans based on assumptions instead of analyzing what’s happening right here, right now, and optimizing decisions based on that. As a result, organizations have been ineffective in translating plans into execution. The rigid structure of planning systems creates data silos, unnecessary complexity, and operational delays. The events that could signal an opportunity or risk get lost in aggregation and get stuck in planning cycles.
Instead, organizations should strive to cut through all that complexity and focus their resources and effort on building situational intelligence to understand what is happening right here, right now, and how it impacts the business. They use these insights to drive their decision intelligence enabling them to adapt decision-making and operations. This is the foundation of BizOps. The best way is to start small with a subset of data and simpler models and slowly experiment with more advanced solutions.
Hyperight: How do you address the challenge of data silos within the context of BizOps, and what strategies can be employed to ensure seamless data integration?
Somil Gupta: Addressing data silos in BizOps is about creating a streamlined approach to data management that supports effective decision-making. Instead of working with volumes of raw data en-masse, BizOps focuses on developing specific data domains, each representing a distinct business aspect, such as customer behavior or operational efficiency. These domains produce targeted data products, which are curated and enriched to provide precise insights relevant to the business context. This strategy reduces complexity and enhances the agility of decision-making processes by ensuring data products align with the business’ operational pace.
Furthermore, BizOps can also leverage existing data assets. Assets like pre-enriched datasets or business-critical information stored in data warehouses, provided they meet the necessary standards for latency, quality, and reliability. When these assets aren’t sufficient, direct data extraction from source systems allows for real-time processing of business event streams. This ensures that decision-makers have access to the most current and actionable data. This holistic approach enables BizOps to break down data silos, facilitating seamless data integration and operational efficiency across the organization.
Hyperight: How do you foresee the role of AI evolving in BizOps, particularly in terms of real-time decision-making and operational efficiency?
Somil Gupta: AI is a core component in generating context-aware situational intelligence and generating decision-intelligence to select the next-best action. One big change we expect in enterprise data and AI strategies is, instead of hauling the data across for analytics, we must focus on embedding decentralized AI models into BizOps Decision Frameworks closer to operations.
Firstly, the model design needs to align sharply with business drivers and outcomes. Secondly, the model needs to be fed real-time data from many different systems to ensure it is sensitive to events. We will need a robust data infrastructure to distribute high-quality, low-latency data to distributed models. Lastly, the model needs to be more reliable and deliver a much higher degree of operational performance. This makes automated monitoring and troubleshooting extremely important. Lastly, we will need federated learning techniques to train models with new data at the edge.
BizOps is about putting data and AI to work together with people and process to solve complex decision problems. So there is also a literacy angle to it. We must demystify AI as probabilistic reasoning devices and accordingly interpret the predictions and risks. We need to be very transparent about what model can deliver and what’s the confidence score. It determines its utility in a given decision problem. We are tracking the development of foundational decision models and Gen AI because they excel at contextualizing insights and generating response strategies and next-best-action. They could be excellent co-pilots or assistants to decision-makers.
Tune into an On-Site Interview with Somil at the Data Innovation Summit 2024!
In an on-site interview at the Data Innovation Summit in Stockholm, we speak with Somil! Somil was one of our esteemed moderators at the 9th edition of the Data Innovation Summit back in April.
Add comment