A simple guide to driving business impact

In this short point of view Ernest explains how do start the implementation of a new technology like Gen AI and how do you make sure it truly delivers the business value. 

Problem statement

Too many AI programs start with technology rather than value. Organizations launch proofs of concept without a clear link to business strategy, without defined use cases and KPIs, and without an execution plan that includes people, process, and change management. The result is familiar: scattered pilots, unclear outcomes, stakeholder fatigue, and stalled adoption. Technology alone does not transform performance; disciplined selection, validation, and scaling of value-creating use cases does.

Author’s perspective and expertise

Observations and learnings from recent projects

The answer lies in starting with the business value. Rather than jumping into technology deployment, leaders must first understand how generative AI supports their strategic goals. What problems does it solve? How will it improve efficiency or drive growth? What KPIs will demonstrate success?

Begin with concrete business use cases that show how GenAI can enhance processes or empower teams. Define measurable goals and focus on areas where AI can make a tangible difference. Experiment with pilots and MVPs to test assumptions and prove value. Once results are clear, scale up investments thoughtfully.

• Start with business value: Ask better questions up front: Which specific strategic outcomes should AI move—revenue, cost, efficiency, risk? If your CFO asked, “What will this change in our numbers within two quarters?” could you answer?

• Identify targeted, high-impact use cases: For example, a customer-service copilot that reduces average handle time by 20%, or contract intelligence that cuts review cycles in half. If you’re a retailer, consider dynamic product discovery; if you’re a bank, think about smarter document processing for KYC; if you’re a manufacturer, explore exception handling in supply chain operations.
• Define KPIs before you write a single prompt: Cycle time, conversion rate, cost per ticket, defect escape rate, compliance exceptions—choose metrics that matter to the business, baseline them, and commit to targets.
• Execute in small steps: Build MVPs and run controlled experiments with real users and real data. Prove feasibility and value quickly. Then, and only then, place a bigger bet.
• Prepare your people as much as your platform: Train teams, set clear guardrails, and build human-in-the-loop workflows. Generative AI is variable by design; confidence grows when experts can review, refine, and approve outputs.
• Scale what works: Productionize successful MVPs, integrate with core systems, implement governance and observability, and expand adoption in waves—by process, team, or geography.

A word of caution

Importantly, technology adoption is only half the story. Equally critical is supporting your people through the transition—training, change management, and building a culture of AI adoption drive sustained impact.

• Beware novelty bias: Don’t pursue GenAI because it is new—pursue it where it creates measurable advantage. If you turned off the AI tomorrow, would anyone notice a difference in the KPIs that matter?

• Don’t boil the ocean: A portfolio of focused, high-clarity use cases beats a sprawling backlog of experiments. Start where the data is ready and the business is motivated.
• Data readiness matters: Poor data quality, unclear ownership, and weak governance will derail even the best models. If your data is incomplete, inconsistent, or inaccessible, fix that first.
• Governance is not bureaucracy: Gartner and Forrester both emphasize the link between responsible use and successful scale. Policies, access controls, evaluation frameworks, and auditability are guardrails, not handcuffs

• Change management is not optional: Technology adoption without change management is wishful thinking. Coach leaders, train teams, and make it safe to learn out loud.
• Control costs early: Track total cost of ownership—model usage, infrastructure, integration, compliance—and compare it to realized benefits.

Concluding point

The key takeaway for business and IT leaders is clear: embrace generative AI by anchoring projects in business objectives, focusing on measurable results, and scaling iteratively. This pragmatic approach aligns with the experiences of successful organizations that are already realizing meaningful GenAI-driven transformation.

Supporting perspectives

To explore how this value-first approach plays out in specific domains, read these complementary points of view:

• Adam Rogalewicz’s point of view on applying GenAI to strategic initiatives like Revenue Optimization explores how to turn pricing, offers, and customer engagement into measurable growth. Read Adam’s perspective.
• Wiktor Witkowski’s perspective on accelerating and automating software development shows how engineering teams can boost velocity, reduce defects, and improve developer experience with GenAI—without sacrificing quality or governance. Read Wiktor’s perspective.