In today's enterprise landscape, something concerning is happening with the way AI initiatives are being evaluated. Many organizations are working hard to understand the value AI can deliver and the costs involved. Yet, by applying traditional ROI models, they risk setting themselves up for disappointment.
As a boutique consulting firm with a track record of helping global organizations unlock value from their digital transformation, we’ve seeing this challenge repeatedly. The problem isn’t the technology – it’s the MINDSET. Where the traditional ROI measures, relying on hard metrics like cost savings and revenue growth and soft metrics like customer satisfaction, assume all benefits can be forecasted a priori.
However, for measuring your ROI for your AI need a different approach. Evaluating ROI should extend beyond immediate gains. As projects progress, employees develop AI skills, data quality improves, and models evolve through learning and interaction. These iterative, compounding effects create long-term value that traditional ROI models overlook.
In a traditional ROI approach, High-potential AI projects may be prematurely abandoned or left unfunded, missing opportunities for business value and competitive differentiation.
To avoid this, organizations need to rethink how they evaluate AI. In this article we’ll explore why traditional ROI frameworks fall short and how shifting your approach can help you identify and prioritize AI projects with true transformative potential.
Let’s dive in.
The Limitations of Traditional Value Metrics
Traditional ROI metrics are designed for projects with clear, tangible benefits like cost reductions, efficiency gains, and process improvements. However, they often fail to account for AI’s transformative potential. Evaluating AI projects using this lens is like judging the value of electricity solely by comparing installation costs to the price of more candles.
Consider the example of a retail bank deploying an AI-powered customer service chatbot. Traditional ROI measures would focus on immediate, measurable outcomes such as reduced call centre costs and faster response times. But as the chatbot system evolved, it unearthed deeper customer behaviour patterns, enabling personalized product recommendations that drove proactive issue resolution that reduced customer churn.
So, what started as a cost-saving initiative transformed into a revenue-generating engine, redefining the bank’s customer experience strategy. Remarkably, the initial ROI calculations captured less than 25% of the actual value delivered – an oversight that applies across industries.
The core issue with traditional ROI is its inability to capture the ‘flywheel effect’, where businesses improve their AI capabilities through interaction, and the AI systems, in turn, become more effective. Using traditional ROI to evaluate AI projects is like judging an iceberg by what’s visible above the surface – completely missing the mass beneath the surface. This oversight is precisely WHY understanding the waves of impact is crucial. Let’s dive in.
The Waves of Impact
Imagine tossing a stone into a pond. The impact of the initial splash creates waves that expand in concentric circles. AI initiatives work the same way, creating ‘waves’ that amplify one another. Let’s explore these ‘waves’ of AI value.
Wave 1: Measured Returns
The initial splash represents traditional ROI - clear and quantifiable benefits such as reduced costs, faster processes, and fewer errors. For example, a chatbot might be measured by call centre savings or reduced response times. These are important metrics, but they only represent the first layer of the AI’s potential impact.
Wave 2: Intelligence Evolution
The circle of impact widens as AI begins to generate insights and uncover hidden value. For example, the retail chatbot from earlier starts revealing customer behaviour patterns, enabling service improvements, product insights, and value creation.
Wave 3: Business Transformation
The compounding effect of these waves has the potential to fundamentally change the business operating model in long-term. Once a simple chatbot use-case evolves from a cost saving tool into a strategic business catalyst, influencing product strategy, customer experience, and market positioning. At this phase, the AI is not about just support the business it becomes the catalyst to reshape it entirely.
Yes, the initial splash - traditional ROI matters. But focusing solely on the initial splash creates the risk of missing the transformational waves.
Navigating the AI Value Journey
Now that we explored how AI creates waves of value, the real challenge is how we harness this potential on the ground. Based on our experience, here’s a five-step approach for capturing value:
1. Set Adaptable Objectives: Begin with clear targets both quantitative and qualitative targets (e.g., "reduce customer churn by 20%") but remain open to unexpected outcomes.
2. Layer Your Metrics: Start with immediate impacts like cost savings, then incorporate deeper measures such as decision quality and team productivity. What you can't directly count often counts most.
3. Create Value Capture Systems: Use tools or customized dashboards to track both direct impacts (quantifiable savings) and indirect benefits (customer satisfaction, cultural shifts). Balanced scorecards combining financial and operational metrics can help.
4. Build Learning Loops: Conduct quarterly reviews with the team who directly work on the actual development to identify unexpected AI benefits. Use these insights to refine your strategy and spot new opportunities.
5. Align Stakeholder Communication: Craft targeted value narratives for different audiences: financial metrics for CFOs, process improvements for operations teams, and strategic insights for executives. This helps for a strong execute buy-in which is one of the key metrics for the project success.