.png){kind=logo}
The tech-driven industries are undergoing a major shift, with companies going from experimenting with AI to implementing AI solutions that make a difference in their performance. Although businesses have spent millions on big data and analytics in the last decade, the challenge they currently face is executing decisions rather than gaining insights.
In this episode of the Analytics Insight Podcast, Priya Diyalani speaks with Amit Shrivastav, Global Head AI GTM and Director Product Management at Kellton, about how agentic AI and intelligent systems are reshaping enterprise decision-making. The discussion explores how organizations can move from reactive analytics to autonomous, goal-driven operations while maintaining trust, governance, and strategic alignment. Here are the key excerpts:
Decision Velocity refers to the ability of an organization to act on insight without sacrificing accuracy and relevance. Even though most organizations are able to get their hands on dashboards and other analytic tools, it has been difficult for these firms to act quickly based on those insights. In dynamic environments, decisions made late will end up creating missed opportunities. Artificial intelligence makes it easier to shorten decision cycles; however, speed is not enough. The right balance between speed and data quality is required for value creation.
The problem is not about the availability of data, but about the design of decision processes. Most organizations are confused about decision responsibility, approval procedures, and execution processes. Despite all the sophisticated analytics capabilities, decisions are frequently held up in approvals and within silos. While the traditional approach centers around the delivery of insights, it does not help much in implementing them efficiently. As a result, organizations suffer from inconsistencies in moving intelligence into action. To resolve this problem, workflow restructuring and governance become critical factors.
It is necessary to incorporate domain knowledge into any AI system right from the beginning. This includes involving experts on the particular subject matter in model creation for accurate capture of the business environment. Intelligent escalations should also be put in place such that the AI operates autonomously within specified parameters and gets human intervention in case of anomaly detection. It is also important for organizations to align their AI systems with proper KPIs rather than relying on proxies. This ensures that the need for speed is never compromised.
If AI tools are not aligned to support business goals, they may end up producing results that are right from an operational perspective but wrong on a strategic level. For instance, focusing too much on transactional margin optimization without thinking of lifetime customer value and brand value could result in strategic damage down the road. This lack of alignment would reduce the amount of trust in AI tools and affect customers. This means that there is a need for clear metric definition and control mechanisms to counteract the challenge.
In conversations about enterprise AI, experimentation has turned into scalability and return on investment. Companies are not asking anymore if AI is effective, but how to utilize it effectively and responsibly. Value first is imperative: Begin by understanding what issues your business faces, what challenges arise in the process, and if AI is actually needed. There may be better approaches in some scenarios. Other characteristics of enterprises using AI well include ethics, explainability, and scalability. They create autonomous outcome-based systems through the integration of AI into their operations.
Listen to the complete discussion on the Analytics Insight Podcast.
