EBOOK

The Radical ROIof Gen AI

Global research shows that enterprises are winning big with generative AI. We dove into their strategies and measured their outcomes.

Based on research by Enterprise Strategy Group:

Nearly three in five organizations worldwide are already using gen AI
92% of early adopters say their gen AI initiatives have paid for themselves
Those who have calculated the ROI are seeing a return of 41%

GEN AIIS A SUCCESS

Though a new and complex technology, gen AI is making inroads across IT, cybersecurity and multiple lines of business.

Respondents use gen AI in IT ops (70%), marketing (44%), customer support (56%) and cybersecurity (65%).
55% have prioritized employee-facing solutions to improve productivity and efficiency.
88% report that they’ve seen a material improvement in efficiency as a result of their gen AI efforts.

93%

say their gen AI initiatives have been mostly or very successful.

DATA STRATEGYDEFINED

How many LLMs do early adopters use? Commercial or open source? Which data challenges lead to the most frustration and cost overrun?

96% of early adopters are training, tuning or augmenting their LLMs.
59% of orgs will have deployed three or more LLMs by next year.
64% say integrating data across silos is difficult.
Only 11% say their unstructured data is AI-ready.

of early adopters expect autonomous agents to take over some tasks by the end of 2025.

INDUSTRYSNAPSHOTS

The research breaks the data down to examine how six key industries struggle and succeed (but mostly succeed) with gen AI.

Healthcare orgs more often report that gen AI is a game-changer for HR.
Marketing firms are 50% more likely to use gen AI internally than customer-facing.
Retailers are least likely to struggle with unstructured data (45% vs. 55% across industries).
Read the report for more, including on financial services, manufacturing and tech.

68% of ad/marketing firms deploy employee-facing chatbots (vs. 50% across industries).

GLOBAL PERSPECTIVE: 8 REGIONAL LOOKS

Bad actors will start to move into the ML infrastructure layer itself, forcing the model to provide the wrong answer, or worse, to divulge private information — like the data the model was trained on. Organizations will be forced to establish a rigorous security approach to operationalizing AI.

UK companies are 9 points likelier to use gen AI for software development and debugging.

SEE HOW GENERATIVE AIHAS CHANGED THE GAME

Researchers at Enterprise Strategy Group surveyed 1,900 business and IT leaders to find out how the first wave of gen AI adoption is changing business.

Short answer: It’s making them richer.

For more, download the full report today.

Download the eBook

Methodology

In preparing to create this report, Enterprise Strategy Group conducted a comprehensive online survey, which was fielded between Nov. 21, 2024, and Jan. 10, 2025. All respondents represent organizations with 500 or more employees. About 39% of respondents represent organizations with more than 5,000 employees; 47% represent orgs with 1,000–4,999 employees; and 14% come from companies of 500–999 workers.

The intent of this report, and the research that underpins it, is to better understand the experiences enterprise organizations are having with generative AI technologies. As such, the survey sought to find organizations meeting two key characteristics: First, organizations already using gen AI to augment and execute business processes in production. Second, the gen AI tech being used must go beyond consumer-grade “off-the-shelf” solutions — such as subscriptions to ChatGPT — to commercial and open-source models that can be trained and tuned with proprietary data to improve accuracy and relevance.

The process of finding this cohort of early adopters allowed us to observe how broadly gen AI has been adopted by enterprises to date. Of 3,324 respondents, 1,900 (57%) said they are using commercial or open source generative AI solutions.

The 1,900 early adopters who completed the survey were drawn from IT and cybersecurity (52%), software development (13%), data operations (8%), and other lines of business (e.g., marketing, customer support, manufacturing, 27%). To qualify, respondents must have reported both knowledge of and influence over their organization’s AI strategies. A range of seniorities are represented, from C-level executives to senior individual contributors. Additionally, the research includes responses from across the globe, including the United States (45%), Canada (8%), the UK (8%), France (8%), Germany (8%), Australia and New Zealand (8%), Japan (8%), and South Korea (8%).

The margin of error for this sample size is +/- 2 percentage points at the 95% confidence level. The totals presented in figures and tables throughout this report may not add up to 100% due to rounding.