Section 2 - Generative AI in action
Generative AI to help firms rethink experiences for both employees and customers
When asked about where generative AI will provide the most positive business outcomes, 88% cited revenue, with 84% pointing to profit, 83% to cost efficiency, and 82% to business model improvement.
The rapid adoption of generative AI and the billions invested indicate that business leaders expect it to have a massive impact, perhaps even becoming a transformative technology.
Indeed, when asked whether generative AI will provide a positive or negative impact on business outcomes, 88% of respondents expected positive impacts on revenue, with 84% expecting the same for profit, 83% for cost efficiency, and 82% for business model improvement (Figure 3).
Figure 3. Generative AI business impact
Source: Infosys Knowledge Institute
Business leaders identified a surprising list of use cases where they believe generative AI will generate impact.
Generative AI is widely viewed as a content generation tool, though with varying degrees of demonstrated success. While some news outlets have embarrassed themselves by publishing AI-generated articles that contained glitchy, inaccurate, and offensive material, individuals have prompt-engineered images of gorgeous, realistic nature scenes.
However, our survey found that business leaders generally do not prioritize these capabilities as the best use of generative AI (Figure 4). Just 13% said that content and creativity were generative AI’s main applications. Instead, 42% expect user experience and personalization to have the greatest impact. Executives also said that generative AI will be used most often to improve operational efficiency and automation (26%) or streamline product development and design (20%).
Figure 4. Where companies expect generative AI to have the most impact
Leaders might be cautious due to generative AI's inconsistent performance in creative tasks, especially in public-facing written content. It may be competent enough to pass graduate school exams. But on other occasions, it might create fictional court cases and precedents for a legal brief. This unpredictability injects enough risk to make executives and PR staff cringe. Also, it can be difficult and time consuming to fact check AI output. For creative applications, generative AI tends to offer speed at the cost of quality.
Section 2 – Generative AI in action
Generative AI Radar 2023: North America
Generative AI’s expected value in user experience suggests a change in how we define this technology. Rather than an isolated writing image, music or code-generating tool, it can be seen as a personalized AI assistant with a range of customized skills synced to employee needs. Infosys has long argued that the most valuable use for AI in general is to augment humans instead of replacing them.
We see particular promise for generative AI in the software engineering cycle. For those focused on project planning and analysis, an AI assistant could help with effort estimation, risk assessment, and simulations.
Software testers can use generative AI to optimize the number and value of tests, eliminate redundancies, automate test script generation, and promote the self-healing of scripts.
User experience also applies to customers with generative AI chatbots or other tools. Amazon offers generative AI to its sellers to assist with the creation of product descriptions, titles, and listing details. User experience also applies within companies, where generative AI chatbots manage first-line IT support queries, for example. Generative AI creates actionable insights from maintenance logs and automates human workflows.
This approach not only allows companies to reimagine user experiences, but also enhances operational efficiency. This is evident in the fact that a fifth of the companies in our survey anticipate the most positive impact of generative AI will be to streamline work and automation (Figure 4).
Our research shows that most large businesses take generative AI seriously, but not all industries are equally advanced. The financial services, healthcare, and life sciences sectors — all data and tech focused — are particularly keen to use this new technology (Figure 10). However, they are traditionally cautious with new innovations due to their high level of regulatory scrutiny.
In healthcare, generative AI can be used to create synthetic data for clinical research and medical education. Synthetic data could help avoid patient privacy concerns, a significant benefit for healthcare firms. Companies such as Paige also use generative AI to accelerate cancer diagnosis and reduce detection errors. The firm is now collaborating with Microsoft to create the world’s largest AI models for digital pathology and oncology.
Financial services companies also use generative AI to create synthetic data for training purposes – although there are challenges with the use of synthetic data, which we discuss in more detail later in this report. In this case, the output data simulates fraudulent and regular financial transactions.
Machine learning models then train on both the real and synthetic data to improve their real-world accuracy. Using real data and documents, investment bank Goldman Sachs has experimented with generative AI to classify and categorize its vast content library.
It’s worth noting that while these three sectors have implemented generative AI at a higher rate than others, the high-tech sector leads the way in the number of generative AI implementations delivering business value. The financial services sector does come a close second — but healthcare and life sciences lag far behind in this regard (Figure 10).
