The potential of Generative AI to drive transformative growth and revolutionize innovation across people, operations, data, and technology is undeniable. However, adopting any transformative technology comes with inherent risks, and Generative AI may present even more. Proper oversight is essential to accurately identify, assess, and manage these risks.
Many leaders, including CEOs, CTOs, CDOs, CFOs, and COOs, have the same question. So where do you start?
/01Where do you start?
This article looks at multiple use cases and provides a simple framework to help leaders decide the starting point. The right first project is rarely the most ambitious one; it's the one that compounds — a focused, measurable application that builds the data foundation, governance muscle, and organizational confidence for everything that comes next.
/02Evaluating a use case
To effectively harness generative AI, leaders must identify use cases that have the potential to make the most business impact. In my experience, there are three critical factors to consider when evaluating the potential of a use case for AI implementation:
Business value. Assess the business value of the proposed use case. Can the benefits be quantified in terms of increased revenue, cost savings, risk reduction, or improved decision-making?
Data viability (barrier). AI models heavily rely on the quality and quantity of the data they are trained on. Evaluate the availability of relevant data for the specific use case. Insufficient or irrelevant data can lead to unreliable model performance. Essentially it is garbage in — garbage out.
Implementation effort. Consider the technical feasibility and effort required to implement the AI solution. Factors such as the complexity of the model, integration with existing systems, and resource availability influence the overall effort involved.
Use cases that score well on all three tend to ship. Use cases that fail on any one — value, data, or effort — tend to become cautionary tales.
/03AI-enabled people
Whether it's customers seeking updates on their orders, employees inquiring about their vacation balances, or CEOs seeking consolidated sales figures, people require information. This flow of information plays a critical role in determining the efficiency and effectiveness of the work performed. Focusing on how people accomplish their work effectively would uncover Generative AI projects that could serve as a starting point.
Customer service (external & internal) to query information has become the top Generative AI priority for organizations. Coupled with Conversational AI, many organizations use this as a strategic stepping stone to ensure a successful enterprise-wide implementation of Generative AI.
Content generation can be an overwhelming and burdensome task in many departments (marketing, sales, HR, finance) often leading to high-value strategic work being neglected as teams struggle to meet constant deadlines. With Generative AI, content creators are freed from the relentless pressure of churning out content.
/04AI-enabled technology
In various sectors, including Banking, Financial Services, Insurance, Healthcare, Life Sciences, Automotive, and Retail, companies of all sizes have come to rely on technology as the backbone of their operations. Generative AI can be a shot in the arm for this backbone.
Democratization of software development. In the past, coding was primarily confined to software engineers, but Generative AI is changing that. It empowers domain experts to contribute to software development by assisting in writing new code, testing and debugging existing code, and identifying security risks. This adoption by domain experts will spark innovation across various fields.
For software engineers, Generative AI enhances productivity by optimizing software delivery time. By leveraging DORA metrics, the impact of implementing Generative AI on the software delivery process can be precisely measured, allowing for continuous improvement.
/05Trusted data as backbone
Generative AI has increased the value of data, leading to competition among companies to exploit its potential. Companies with abundant and trustworthy data tend to achieve higher ROI from their AI investments. Identify required data sets by working backward from the customer experience to determine what a Generative AI platform should offer.
Data mining of unstructured data is crucial for developing competitive advantage and differentiating platform value propositions. Most enterprises sit on substantial unstructured content — emails, documents, transcripts, tickets — and most of it is not yet feeding any model.
/06G-FORCE — governing the program
Three implementation principles separate programs that scale from those that stall. Ground the model: retrieval against your trusted content, with citations, beats raw generation for almost every enterprise task. Keep humans in the loop on anything customer-facing or consequential — not permanently, but as the default until measurement says otherwise. Establish guardrails early: data classification, allowed use cases, evaluation harnesses, logging, and a clear policy for what the model is permitted to see.
Package the program scaffolding as G-FORCE — Governance, Financial controls, Objectives, Robust technology, Culture and Education. Six headings, one per executive owner, so the program has clear accountability before the first pilot.