It’s no secret that generative artificial intelligence (AI) has enormous economic potential—McKinsey & Company projected that the benefits could total over $4.4 trillion in value for the global economy. But the dust is still settling after a frenzy of companies investing heavily in generative AI, and many are finding that ROI remains elusive.
Unsurprisingly, technology teams have been eager to experiment with various proof of concepts (POCs). But as Ben Schreiner, Head of AI and Modern Data Strategy Business Development at AWS, notes “The real sticking point has been getting POCs into production. This is partly because large language models (LLMs) are the Swiss army knives of technology—a tool capable of so many different tasks.” With such a huge scope of applications across product development, customer experiences, and much more, what makes generative AI so exciting is ironically what has stopped companies from getting the most value out of it.
“Many have struggled to focus on working backwards from a real business problem. Even once they’ve uncovered a well-defined business use case, failure to build internal alignment stunts ROI. To make investments pay off, a paradigm shift is needed,” Schreiner adds. His recommendation for driving generative AI’s revenue potential? “Companies of all sizes should pursue a purposeful strategy encompassing employees, business processes, and customers.”
Value Driver #1: Augment Your People
As the saying goes, time is money. From writing emails faster to automating menial tasks, the value lift of improving employee productivity is undeniable. But as Schreiner points out, “You can’t reach new levels of efficiency if your people feel uncomfortable about generative AI. Any barrier to adoption will be a barrier to business outcomes.” According to a recent Prosper Insights & Analytics survey, employees’ top three concerns about AI were that it needs human insight (32 percent), it can provide incorrect information (29 percent), and that it will cause job losses (27 percent).
Ongoing training and enablement will help alleviate these worries, while also empowering teams to reap the benefits of tools. As Schreiner says, “While certain functions will be automated, most people will be augmented by generative AI, not replaced by it. To get the best results, we need to analyze its impact on employees, prepare them for the evolution of work, and bring them on the journey.”
He continues, “Ultimately, AI systems are predictive, not determinist, which makes a human-in-the-loop approach critical. By upskilling people to verify outcomes, you can both prevent hallucinations and mitigate worries.” The manufacturer Georgia-Pacific is just one business seeing the rewards of this strategy. By leaning into employee expertise when creating a tool for routine maintenance, they’ve empowered workers to become more effective and efficient. Their LLM works in harmony with subject matter experts and machinery data to give operators answers to maintenance queries quickly and improve their experiences. Georgia-Pacific now estimates millions in potential savings from generative AI.
Value Driver #2: Rethink Business Processes
Workflows inevitably become more profitable with greater throughput and improved quality. Data is every business’s ally here, identifying where AI can create efficiencies. But you also need to pinpoint bottlenecks. As Schreiner says, “You don’t want to automate a bad process. To optimize workflows, companies should work backwards from the customer experience and the employee experience. By proactively seeking ways to free up employees’ time, they can focus on value-added tasks.”
From Schreiner’s experience, aligning people across the organization in this effort is key again—business and technology teams need to be in it together. As he explains, “IT can’t work in silos. Only business executives know the questions that need to be answered, and therefore the data that’s needed to answer those questions. Then you can properly determine the right models, algorithms, and technology to deploy.”
“From the get-go, there should be a solid understanding of where data is coming from, what it’s being used for, and how it should be protected,” Schreiner advises. The findings from Prosper Insights & Analytics’ survey also show that 86 percent of employees are concerned about their privacy being violated from AI’s use of data, and 85 percent of small business owners share the same concern. To overcome these barriers, Schreiner recommends building enterprise-grade security from the beginning: “This will protect sensitive data to avoid fines from breaches, while also fostering trust in AI-driven outputs”
With regulations brewing, Schreiner explains the importance of proving you’re in control of generative AI systems by narrowly defining the measurement of success and putting guardrails in place to focus the solution on the task at hand. “Business applications shouldn’t be able to give you grandma’s chocolate chip cookie recipe. Being really purposeful in what you want AI to do and what you don’t want it to do will prevent Pandora’s box being opened and ensure that you only get contextually appropriate responses,” he explains.
Value Driver #3: Master the Art of Reinvention
Lastly, Schreiner emphasizes that driving business value from generative AI means seizing the opportunities for creation with both hands. That’s because not only does reinventing products, experiences, and business models accelerate revenue—it also opens completely new revenue streams. “There’s no shortage of ideas for how to use generative AI but figuring out how to enhance its competitive edge with AI is where companies can really win by unlocking increased sales, value-per-customer, and value per average sale price,” he adds.
It stands to reason that those who don’t continually enhance offerings are putting their competitive position at stake. As such, Schreiner points out: “Retaining current customers and acquiring new ones means analyzing and responding to evolving needs, especially when AI is changing how users interact with technology and get value from it.”
He continues: “For software companies, transforming customer experiences could mean building new features and capabilities. But for traditional companies, including those in the SMB space, leadership should consider how to evolve customer engagement. One such use case is using intelligent customer service agents. Rather than taking time to manually look up answers, AI agents can feed answers to human agents while on calls to customers—streamlining resolutions and removing friction for both the employee and customer.”
Manage Risks, Maximize Rewards
Schreiner’s overarching advice? “For the best chance of realizing business value, companies should choose feasible projects where they have the skills to execute tangible impact or find a partner to help. Quantifying the benefits of solving the problem for a given use case is critical to having a strong ROI and avoiding wasting time. For instance, with the cloud you can just pay for what you use, helping to ensure that the benefits offset the costs. Technology partners can prove hugely valuable here, enabling you to find the right opportunities for transforming processes, services, and how your people work.”
It’s clear that with all the possibilities AI presents, prioritization becomes extremely important. To put it into perspective, McKinsey assessed over 2,100 work scenarios to analyze generative AI’s economic potential—and there are many more to be explored. As Schreiner summarizes, “You could have the best AI tools, but without strategic prioritization, business alignment, and change management, you’re unlikely to achieve the results you hoped for.”