AI in Business: Harnessing the ‘Tremendous Power Now in Our Hands’

Artificial intelligence has the potential to solve the world’s problems, like climate change, plastics and even inequality, believes one of the readers of The Information. Whether it will do so depends “on the questions we ask of this tremendous power now in our hands,” says the reader.

A recent survey of 108 readers of The Information (see the methodology) sheds light on where companies are in their journey to harness the power of AI to transform their businesses. The survey reveals that enterprises are at the beginning of AI adoption, with IT, sales and marketing having made the biggest strides.

Beyond the level of AI adoption, the following article discusses what it will take to make AI a business success. It delves into the perceptions of which AI technologies are the most promising. It also uncovers the challenge of finding use cases for AI solutions that add business value, as Jon Victor highlighted in our recent virtual event, “Use Cases, ROI, and What’s Next for 2025?” It discusses the important interplay between people and technology to new ways of working and to redefining industries.

AI Adoption: State of Play and Beyond

It’s early days for AI adoption, with 18% of respondents surveyed by The Information saying their companies have mostly or fully adopted AI across all enterprise resource planning functions (see chart).

The most mature functions in terms of using AI are in information technology, with 44% of The Information’s survey respondents saying their companies have mostly or fully adopted AI tools, followed by sales and marketing (35% mostly or fully; see chart). It’s worth noting that even for the most mature functions, less than half of companies say they have mostly or fully adopted AI tools.

The high adoption rate of AI in IT should come as no surprise. As one survey respondent puts it: “AI unlocks a vast array of coding tasks that democratize code creation. Design and brainstorming are taken to a whole new level. The throughput on tasks like engineering and simulation increases by a factor of 10.”

In sales and marketing, the second most mature function in terms of AI adoption, survey respondents point to the use of generative AI to create content for marketing campaigns that provide direct return on investment through the use of marketing chatbots or automation of personalization. But one respondent warns about the limits of AI in sales and marketing, pointing to the importance of the human factor: “There is only so much of the sales development representative activity that you can automate before seeing declining results.”

Although just over a quarter (26%) of survey respondents say they are not planning to use AI, a wait-and-see approach to new technology is not uncommon in small to medium-size companies, which were 53% of the respondents to the survey. The highest number of respondents not intending to use AI are in supply chain management (31%) and human resources (28%).

Supply chain management is ripe for AI-driven digital transformation. For example, in the digital supply chain, the time-consuming verification of paper-based delivery notes during the goods receipts process at a factory gate affects every company with production sites. An SAP tool automates processing of documents with greater accuracy and efficiency, reducing operational costs by up to 55%.

At the same time, supply chain management presents a tough environment for the adoption of AI tools, which are based on data processing and analysis. Obtaining clean data from a single reliable source is often the biggest challenge within an organization, and the difficulty of acquiring such data from vendors or partners compounds the issue.

However, regulators are beginning to require reporting that includes third-party data. Such data are necessary for sustainability reporting, with companies’ supply chain emissions (Scope 3 emissions) significantly greater than their emissions from direct operations. While the final Securities and Exchange Commission climate disclosure rule does not specifically require Scope 3 reporting, some other regulations do.

When primary data for sustainability reporting are not available, companies can rely on AI models and averages based on emissions categories and spend to calculate Scope 3 carbon footprints. One consumer goods company, for example, uses an AI tool to extrapolate the carbon footprint for the travel of its products from a warehouse to a store.

The Technology Factor: Beyond the Low-Hanging Fruit

A new technology used in digital transformation in an enterprise has one major goal: to add business value. That value can be measured in many ways, including cost savings, productivity gains and revenue generation. While The Information’s survey finds that companies are making significant investments in AI-driven business transformation (61%), more than half (51%) of the survey respondents say companies are finding it hard to realize business gains from it.

The two top challenges are finding use cases that add business value (60%), and creating business cases or proving ROI (43%). Business AI solution providers like SAP are addressing this need, with 130 use cases rolled out in 2024 and over 300 planned for 2025, as well as ROI targets for their customers, including 16% on average in 2024 and a goal of 30% in 2025. The difficulty in measuring ROI may also result from the still-evolving cost structure of AI solutions. Currently, the most prevalent payment model for AI solutions is user-based subscription (55%), while the most preferred payment model is pay-per-use pricing (36%), according to The Information’s survey.

The Information’s readers say AI offers “lots of low-hanging fruit to improve efficiency and drive productivity,” with benefits such as automation of core business functions, content creation, and business development and customer relationship management optimization.

But The Information’s readers point out that there is potential to achieve so much more. “AI should be used to create new revenue streams,” says one reader, citing one company that created a new revenue stream by converting energy production data into financial assets that can be bought and sold on open markets.

AI technology “transforms thinking into a set of deliverables, whereby the ultimate goal is all-inclusive reasoning,” says one reader who is aiming for machines to achieve humanlike results.

In terms of specific AI technologies, generative AI tops the list as the technology with the most potential for business transformation. It is followed by natural language processing, the foundation for generative AI.

AI should be used to create new revenue streams

In effect, generative AI helps solve the blank page syndrome that can stump many people at the beginning of a creative process. “Often in the designing of things the initial concepts are the most difficult phase,” says one reader. “AI helps with idea generation to get the project started.”

In terms of other AI technologies, one of The Information’s readers draws a timeline depicting when the effects of different technologies would be seen and what they would be:

• On a two- to seven-year time scale, the large language models and agent-based technologies would enable cost cutting and reengagement of the workforce in an updated way. That would require retraining and fluency in trends and forecast-based reasoning at all levels of hierarchy within an organization.
• In a ten-year time frame, three major AI-driven transformations would be: 

1. Updated definitions of the business models and strategies due to changes in key market elements like cost structures and barriers to entry.
2. Redesign of business models, goal setting and reporting for a new set of variables such as environmental impact. 
3. An updated approach to human workforce retraining and reengagement. The impact of AI on the workforce would demand a sustainable human capital approach to operating businesses and new ways to motivate employees. 

• On a seven- to fifteen-year horizon, the robotics advancements and evolution of deep learning and similar AI architectural design approaches would start redefining the industries and industry structures as we know them. 

The Human Element: Change Management Is Key to the Success of AI

The success of AI solutions will depend on people. “More than the technology, the resistance to change by the existing workforce is the cause for failure or reduced ROI on AI investments,” says one survey respondent.

More than two-thirds of The Information’s survey respondents (66%) believe the success of AI-driven transformation largely depends on change management and getting the workforce ready. In contrast, less than half (49%) say this success depends largely on the technology itself.

From the point of view of employees, the implementation of AI can sound like a judgment day of sorts. “AI makes the best workers better, brings up the average and further highlights the laggards in our firm,” says one survey respondent. “The employees that use AI most readily will separate themselves from the rest that do not,” adds another.

Some employees are embracing AI and reaping the benefits. “AI impacts me and my colleagues into becoming more independent and self-sustainable to front tasks at hand and become more efficient,” says a respondent. “AI makes a difference and adds a great value with tasks automation and productivity gains in some workflows,” says another.

The new frontier of AI in the enterprise is agentic AI, which further impacts the role of a human worker. So called because they exhibit agency, agentic AI systems can autonomously pursue goals, make decisions and learn from changing conditions without human participation. By taking on repetitive tasks, they can free up humans to focus on high-value work.

Almost half (47%) of The Information’s survey respondents believe agentic AI will prove to be a linchpin for widespread adoption of business AI. And 41% believe companies think they will benefit most by integrating generative and agentic AI into their ERP applications such as finance, human resources or procurement.

“In a few years, autonomous artificial-intelligence ‘agents’ could be performing all sorts of tasks for us, and may replace entire white-collar job functions, such as generating sales leads or writing code,” says The Wall Street Journal. Other sources point to the continued role of a human: “Ultimately, perhaps no agentic AI will be fully autonomous—humans may always need to be involved, especially when material actions are required,” according to an article from the Sutardja Center for Entrepreneurship & Technology at UC Berkeley.

Whatever the future scale of human involvement, AI will affect people’s roles, scope of involvement with technology and work processes. One of The Information’s survey respondents sees the human factor as requiring the most time to get right for the success of AI: “The longest time frame in these AI-driven transformation processes will not be brought by the ‘in the lab’ tech advancements. Rather it’ll be in change management for the rollout and at-scale application of these technologies.”

The good news is that human metrics figure prominently in how companies measure the benefits of AI. After savings of time and cost, the next three metrics are directly linked to employees: number of people needed to perform tasks, employee adoption and employee satisfaction with using AI tools (see chart). Almost half of companies measure employee adoption (49%) and satisfaction (46%). But there is much room for improvement in terms of employee uptake of AI, with just over a third of survey respondents saying they see benefits in these areas (36% for employee adoption and 35% for employee satisfaction).

The Information’s readers see the human aspect of working with AI as simultaneously the greatest challenge and the greatest opportunity. “AI has incredible potential but really needs to be adopted and managed through a change management framework to be effective and yield results faster,” notes one of the respondents to The Information’s survey. Another adds that the opportunity of AI can be realized via an optimistic scenario of harnessing the AI-driven training to actively reengage humans in new ways of working.

CONCLUSION

Harnessing the power of AI in an enterprise comes down to deriving business value from AI solutions. The AI-driven business success can be accomplished by following these guidelines:

Fuel AI with clean and relevant data. Obtaining clean data from a single reliable source is often the biggest challenge within an organization, and this challenge becomes tougher to overcome when considering AI solutions that encompass third-party data from vendors or partners.But clean and relevant data is necessary for AI to yield trustworthy solutions. Without trust, AI will not take hold.

Create relevant use cases with ROI targets. Finding use cases for AI solutions that add measurable business value is the top challenge for enterprises, according to The Information’s readers. SAP has rolled out 130 use cases in 2024 and plans 300 for 2025. It also sets ROI targets for its customers, including 16% on average for 2024, and a goal of 30% for 2025.

Go beyond the first wave of use cases. Benefits such as automation of core business functions like invoice processing, personalized marketing, and customer service chatbots are just the first wave of what’s possible with AI. Aim for creating revenue streams through AI-driven data monetization and the redefining and redesigning of business models.

Embrace agentic AI. The new frontier of AI in the enterprise is agentic AI, so called because agentic AI systems act on our behalf to make decisions and continuously learn. By taking on repetitive tasks, agentic AI can free up humans to focus on high-value work. It’s especially powerful when integrated within ERP systems.

Focus on people. The human aspect of working with AI is simultaneously its greatest challenge and greatest opportunity. The Information’s readers believe the success of AI-driven transformation largely depends on change management and getting the workforce ready. Achieving success with business AI requires a strategic approach to human workforce retraining and reengagement.

METHODOLOGY

This article is based on a survey of 108 readers of The Information conducted in December 2024. The survey respondents came from companies of different sizes and multiple industries, functional areas and corporate ranks.

Company size: More than half (53%) had revenues under $10 million, 17% had revenues between $10 million and $100 million, 5% had revenues of $100 million to $500 million, and another 5% had revenues of $500 million to $1 billion, while the remaining 21% had revenues of $1 billion or more.

Industry: The biggest group of respondents (37%) came from technology, media and communications, followed by professional services (13%) and the financial sector (12%). The remainder represented healthcare and life sciences, government and the public sector, retail and consumer products, and other fields.

Functional area: The biggest group of respondents (36%) came from general management, followed by research and development (14%), marketing and communications (12%), sales (11%) and IT (10%). The remainder represented finance, human resources, legal and manufacturing, and operations.

Rank: The biggest group of respondents was directors (27%), followed by CEOs and owners (17%), owners (11%), managers (10%), senior vice presidents and vice presidents (10%), CEOs (9%), chief information officers and chief technology officers (9%) and other roles (7%).