Almost three-quarters (74 per cent) of businesses that have implemented generative AI (Gen AI) projects reported that they had seen return on investment (RIO) on at least one of those, according to a report by Google Cloud. Just over a half (52 per cent) also reported increased overall annual revenue of between 6 and 10 per cent as a result of Gen AI initiatives.
“Gen AI can be a powerful revenue generator,” the report said. “However, to achieve the best results, you should adopt a strategic approach to integrating Gen AI into your core operations and invest in the necessary resources to support its long-term growth.”
Business benefits
Most organisations (85 per cent) reported that Gen AI had driven improved user experience and engagement, but fewer than half (45 per cent) said that the software had driven productivity. The most common productivity benefits included IT process and staff productivity (70 per cent), faster time to insight (60 per cent) and increased accuracy (59 per cent).
The strongest areas where businesses achieved RIO related to back office and business processes (43 per cent), individual productivity (40 per cent), and sales and marketing (40 per cent).
“The findings highlight the importance of moving beyond experimentation and towards a more strategic and integrated approach to gen AI adoption,” the report said.
Responsible AI
But, in a report by PWC, while 70 per cent of participants said they had invested in AI, only 11 per cent said they had been able to balance risk with reward. The overall benefits of investing in responsible AI practices (RAI) included enhanced customer experience (41 per cent), better cybersecurity and risk management (40 per cent), and facilitated innovation (39 per cent).
The report warned that while many organisations were rushing to adopt AI, many were not ensuring that it was done so responsibly. Most commonly (29 per cent), respondents complained that it was difficult to quantify risk mitigation from an RAI programme. But that could leave organisations open to a range of problems, it said.
“Privacy and consent frameworks, for example, often don’t have specific provisions for customer data entering AI systems,” the report said. “Data governance may not cover Gen AI models’ access to swaths of internal data, some of which may be sensitive. Legacy cybersecurity rarely considers risks from Gen AI’s many new users.”
