As artificial intelligence moves beyond the hype cycle into practical implementation, enterprise organizations face unique challenges in scaling their AI initiatives from promising pilots to production-ready systems that deliver measurable business value.
In a recent interview with AI Today, Vijay Guntur, CTO of Ecosystems and Practices at HCLTech, shared valuable insights from his 30 years of experience guiding enterprises through technological transformations.
Broadly, Vijay recommends AI implementation follows this structure:
Stage 1: Addressing Data Quality Challenges
When organizations move from pilot to production, data quality emerges as one of the first major hurdles. Limited, curated datasets that work well in controlled pilots often fail to represent the complexity and variability of real-world scenarios.
Key Challenge: “When they’re doing pilots, they have limited data sets and the quality is reasonable, but when they scale and when they infer, they find gaps. Ensuring that the quality of data is high and it’s representative of what we expect to happen during inference time or during usage is a big gap in getting things into production.”
Implementation Strategy: Before scaling any AI initiative, conduct a comprehensive data quality assessment.
Identify gaps between your pilot data and production data environments. Develop data governance frameworks that ensure consistent quality across all sources feeding your AI systems.
Stage 2: Managing Organizational Change
The second critical phase involves preparing your organization—particularly your workforce—for the changes AI will bring to established workflows and processes.
Key Challenge: “People underestimate the change that is required to adopt systems. It’s a matter of managing that change in a very methodical and systematic manner in organizations and taking people along.”
Implementation Strategy: Guntur recommends focusing on two core aspects:
- The ability of people to understand the technical aspects of it, the business impact of the change that we are making… preparing the organization by up-skilling, building the capability.
- When you actually deploy these systems, there will be some disruption… How do you equip people to do that work in the new manner?
A real-world example demonstrates this approach: “We built a clinician system that will help clinicians become more effective in how they use existing knowledge bases when they are diagnosing and prescribing a course of treatment. We’ve been able to get up to 15% efficiency in this process, but making sure the clinicians understand what the system does is an important aspect to land this into production at scale.”
Stage 3: Measuring ROI and Business Impact
As AI moves beyond experimentation, organizations need clear frameworks for measuring return on investment and business value.
Key Metrics Framework: According to Guntur, AI investments typically deliver value in three primary categories:
- Productivity and Efficiency: “People are seeing productivity and efficiency gains… that is one aspect of it.”
- Risk Management: “People are also wanting to manage risk better. So whether you think about technical risks or business risks, especially in financial services, regulatory risks, AI is having a big impact in reducing that risk.”
- New Business Value: “We have newer systems that can change and rewire the way you do things, re-engineer the process to create new business value, which was not there earlier.”
Implementation Strategy: We are seeing a simple breakeven analysis for some of these builds to be of the order of two years or lesser as well,” notes Guntur. Organizations should establish clear metrics in each category relevant to their business objectives, with realistic timeframes for measuring returns.
Stage 4: Implementing Governance and Ethical Safeguards
As AI systems scale across the enterprise, governance frameworks become essential to ensure responsible use and regulatory compliance.
Key Challenge: Balancing innovation with responsible AI practices across the entire AI lifecycle.
Implementation Strategy: “I think that full life cycle, if you think about build, test, and then deploy and continuous monitoring… there’s regulation that’s coming about making sure based on the criticality of the system, what kind of norms are expected, what kind of implementation rules are expected,” explains Guntur.
He adds: “Businesses are adopting to this… to be able to make sure the systems that we have built are responsible, they are ethical, they provide the necessary security that systems are expected to have.”
Stage 5: Preparing for Emerging AI Capabilities
Looking ahead, organizations should begin planning for emerging AI capabilities that will likely become mainstream within the next 12-24 months.
Future Trend: “This is supposed to be the year of the agent tech,” predicts Guntur. “Today we have point solutions for agent tech and not much standardization about how agents communicate with each other. But I think a lot of that will be in place in the next year or so. And we will have agent to agent collaboration, agent to agent communication, which means you can build custom workflows for business process to use agentic technology and AI, which will make sure the whole AI adoption is faster.”
Implementation Strategy: Organizations should begin exploring potential use cases for agentic AI within their workflows and processes, preparing both technical infrastructure and team capabilities for this emerging capability.
Conclusion: The Shifting Business-IT Relationship
As AI becomes more accessible and integrated into business operations, the traditional relationship between IT departments and business units is evolving significantly.
This transformation is…going beyond IT departments. Businesses are getting comfortable because as consumers, they are getting a lot of the exposure to this tech.
Businesses are influencing the decision-making and sometimes even funding projects for this change to happen. It’s just not the IT organization.
This shift creates new opportunities for collaboration between technical and business teams, with AI serving as a bridge between technological capabilities and business outcomes.
Organizations that successfully navigate this changing dynamic will be better positioned to realize the full potential of AI across their enterprise.
