Measuring the ROI of Generative AI Projects in Enterprise Settings

Calculating the return on investment (ROI) for artificial intelligence (AI) initiatives is more complex than you’d think. There are dozens of hidden cost factors and intangible benefits. How do you ensure your organization’s project is financially successful?

The Challenges of Attributing Improvements to AI Initiatives

It seems just about every business is adopting AI solutions lately — and most are not worried in the slightest about securing returns. According to a 2024 KPMG survey, 78% of senior business leaders are confident they will see an ROI for generative AI by 2027 at the latest.

But how do enterprises determine whether or not their financial and operational improvements stem from their AI initiatives? Doing so may be more challenging than decision-makers think. According to Deloitte, around 41% of firms have struggled to measure such impacts.

The intangible benefits of generative AI implementation are extensive. How do firms quantify improved decision-making, enhanced creativity, or refined innovation? Even if they can define and measure their generative AI project’s outcomes, estimating ROI may still be challenging.

The Importance of Defining Metrics Before Deployment

Enterprises must be as technology-oriented as possible to deliver unparalleled business results. However, most struggle to develop effective strategies due to the rapid pace and massive scope of transformation. Moreover, many find managing IT infrastructure expenses challenging.

Many undergo those same struggles when adopting generative AI since it is evolving exponentially. This rapid model development rate complicates ROI formulas.

Every month or so, generative technology makes massive strides. This means the latest solutions will require updates within six months at the latest. Businesses investing in them must factor these changes into their cost analysis.

It’s worth noting that many information technology (IT) teams won’t have enough time to complete implementation before updates are required. In fact, it takes an average of eight months to go from prototyping to production when adopting AI initiatives.

For this reason, defining clear metrics before deployment is crucial. Revenue growth is a given. Other options include client retention rate, cost savings, and employee productivity. However, the specifics vary depending on the model’s application.

Methods for Quantifying Generative Technology’s Returns

Senior business leaders must quantify returns. Although experts project generative AI will contribute $4.4 trillion to the global economy annually, the largest gains will be concentrated among very few firms. Even for large enterprises, an ROI is not guaranteed.

Chief financial officers (CFOs) and chief information officers (CIOs) must define and measure their project’s financial and operational impacts to determine its value. There are several ways they can approach this problem.

1. Benchmarks

How efficient are staff members? Are customer retention rates above the industry average? How much does the workplace spend on labor? Establishing a benchmark using financial records, client satisfaction surveys and employee productivity levels is crucial. CFOs can only tell whether they are seeing returns if they have marked a starting line.

2. ROI Formula

CFOs and CIOs can calculate their project’s ROI by subtracting the implementation cost from their investment’s net gain and then multiplying that number by 100. For example, if they spent $24,000 but made $32,000, they’d have a return of 133%.

3. Cost Modeling

On top of spending money on computing infrastructure, large enterprises must pay for quality datasets and AI talent. Other indirect costs include employee training, compliance, and disruption to business processes. Cost modeling throughout all implementation phases will increase the accuracy of ROI projections.

Best Practices for Measuring a Generative AI Project’s ROI

According to the International Data Corporation, organizations realize an average return of $3.50 for every $1 invested in AI. Interestingly, around 5% of firms worldwide are seeing an $8 return on average.

Well-equipped enterprises are poised to reach this upper limit if they are strategic — which they can only be if they crunch the numbers.

Quantifying downstream outcomes for intangible benefits is a good rule of thumb for calculating an AI initiative’s ROI. For example, higher customer satisfaction typically translates to better lead generation. Alternatively, improved productivity leads to a higher project completion rate.

Stakeholder communication is another best practice for capturing a comprehensive overview of ROI. It is particularly beneficial for large enterprises with multiple generative AI applications.

According to a McKinsey & Company survey, 72% of firms have adopted generative AI for a minimum of one business function. Around 50% use it for two or more applications, while about 27% have deployed it in at least three functions.

In scenarios like these, communication is everything. Continuous reporting and analysis lend to problem-solving and optimization. The C-suite can only see the bigger picture if each department contributes its piece of the puzzle.

The Bottom Line of Calculating Returns for AI Initiatives

Even though simplistic formulas exist, there is no one-size-fits-all solution for calculating an AI project’s ROI. It’s in your best interest to gather comprehensive, enterprise-specific data points to make informed decisions.