I advise StarCIO Digital Trailblazers against forecasting an ROI when crafting vision statements for their initiatives. At my digital transformation workshops, we guide leaders in defining their visions, debating their benefits, and prioritizing which ones to plan. To quantify benefits, I prefer seeing explicit objectives and key results (OKRs) or key performance indicators (KPIs or improvement changes in them) as they are more explicit in what types of benefits the initiative targets.

I’m even more dubious about projecting ROI for AI initiatives, especially if the primary net benefits are productivity improvements. I suspect many teams rushing into their AI POCs did little to baseline productivity, and it’s hard to project the investment costs with any accuracy – much like trying to pinpoint timelines and lock in scope.  

My discussions with CIOs align with recent articles on how they struggle to find ROI from AI and that organizations may be expecting results too soon.

But recently published research challenges these assumptions. CIOs and Digital Trailblazers must recognize that CEOs and Board Directors expect to see realized short-term benefits from genAI investments.

Holy cow! ROI from GenAI

Here’s what the research says:

• IDC says that for every $1 a company invests in generative AI, the ROI is $3.7x, and top leaders realize $10x. They report that deployments take an average of 8 months and that the value is realized within 13 months.
• Deloitte’s 2024 year-end generative AI report states that almost all organizations report measurable ROI and one-fifth (20%) report an ROI excess of 30%. Similarly, nearly three-quarters (74%) say their most advanced initiative is meeting or exceeding their ROI expectations (43% meeting, 31% exceeding).
• McKinsey’s report on empowering people to unlock AI’s full potential is more pragmatic and states, “Gen AI has not yet delivered significant return on investment for enterprises.” Yet, their data shows the c-suite’s perception that 58% have increased revenue and 23% have decreased costs from genAI.

We can back into quantifying the target business benefits for genAI investments. According to the AI at Wharton report, the average enterprise planned to invest $10.3M in gen AI in 2024, with 43% of respondents stating they spend more than $10M. Deloitte also reports in their year-end study that 78% expect to increase their overall AI spending in the next fiscal year (2025).

Rounding these numbers, if an enterprise invests $11M in gen AI in 2025, it would need to generate $14.3M net benefits from it beginning in 2026 to generate the 30% ROI target. To hit IDC’s ROI, the organization must target $40.7M in net benefits.

Measuring AI of code assistants

Let’s look at calculating ROI from genAI and why focusing only on productivity improvements can lead to less desirable actions.

In the Deloitte study, 28% report their most advanced (scaled) gen AI initiatives are in IT (28%), much higher than the next two in their ranking of Operations (11%) and Marketing (10%). There are differences by industry. For example, the energy sector’s most advanced gen AI initiatives are in operations (23%), and consumer businesses have IT and marketing tied at 20%.  

So, let’s dive into IT. According to IBM’s Enterprise AI Development Report, 99% of respondents report using AI coding assistants in some capacity, and 41% of developers say it saves them 1-2 hours per week.

I used ChatGPT to calculate this net benefit for an organization with 500 developers. Assuming a $75 hourly rate, the net benefit is between $799,500 (1 hour of time savings) and $1,599,000 (two hours) per year. A smaller department with only ten developers would save between $15,990 and $31,980 yearly.

Using the 500 developer example, to get a 30% ROI, the CIO could invest up to $615,000 or $12,300 per developer.

Well – not exactly. You’d have to include the upfront implementation and change management costs. I’m not going to do the math, but I suspect IT departments with 500 developers that take six months to fully roll out AI code assistants will achieve the ROI. In other departments, the rollout and adoption can be longer, which is likely why the most advanced AI rollouts are in IT.

When the CFO looks to realize GenAI’s net benefits

Now, here’s the problem. You promised a 30% ROI, and productivity generates the net benefits. What does the CIO do when the CFO wants to cash in and financially realize the business value of the gained productivity?

A CFO holding the CIO to that number will want the productivity translated to a business benefit. Did the productivity yield faster delivery timelines? Did the IT department take on more projects? Did the team address production issues faster? Was the AI-generated code of higher quality, which resulted in fewer production defects? Was less technical debt created through the process, and is the code easier to maintain?

Pause for a second. If the CIO answers yes to any of these questions, these are separate net and potentially higher business benefits than the productivity gains.

But, the CIO didn’t express or measure these OKRs, and the CFO drew a straight line from productivity to cost savings. The CFO will likely look to reduce headcount, probably starting with the other 59% of developers who did not adopt gen AI coding assistants into their workflow.

Successful GenAI, but failing to communicate the business value

The bottom line is that you can achieve ROI from gen AI investments, and the big X factor is your timeline and cost in planning and executing a change management program that drives adoption. However, investments in gen AI can target and demonstrate more valuable business benefits, which CIOs and Digital Trailblazers should focus on, especially when communicating with executives and stakeholders — starting with their vision statements.

I recommend against forecasting ROI. It requires too many assumptions, and stakeholders must tune into how it was calculated, or they’ll draw their own conclusions. Debate the net benefits, express them as OKRs, set a baseline where possible, and calculate them as key deliverables for the initiative.

Where to find other net benefits? See my other articles on 5 tips for better business value from genAI and winning with genAI by surpassing productivity drivers.

Soon, I will publish a genAI business benefit checklist on the StarCIO Digital Trailblazer Community.

Tune into my recent Driving Digital Standum for more on AI governance and business value.