For the second annual edition of Putting Generative AI to Work, Altman Solon, the largest global strategy consulting firm exclusively working in the TMT sectors, delved into generative AI tooling in the workplace. In the first half of 2024, our team surveyed over 400 senior-level executives worldwide and conducted in-depth interviews with a panel of 13 industry experts. Findings show widespread adoption of generative AI-powered enterprise tools, as well as growing industry-specific use cases.
Brad Lightcap, OpenAI's Chief Operating Officer, predicts that 2024 will be "the year of the enterprise" for generative AI. Our second edition of Putting Generative AI to Work illustrates the reality of Lightcap's hypothesis. We analyzed generative AI tool adoption across 11 business functions. Findings show high adoption rates in software development and professional and technical service delivery functions, as well as the emergence of industry-specific use cases. Despite worries over a generative AI "bubble," our conversations with executives paint a different picture: they have a growing interest in generative AI products and wish to leverage them across a wide range of business functions.
Software developers are using generative AI tools extensively, jumping to 78%, from 22% adoption in 2023. Tools like GitHub Copilot, CodeGuru, and general models like ChatGPT help with everything from code generation to translation to debugging. Data from GitHub shows that developers appreciate AI-powered assistants, with many reporting that the tools, which remove some of the drudgery from programming, increase satisfaction and productivity.
Generative AI has been widely adopted among professional service delivery professions (e.g., consulting teams, attorneys), with over three-quarters (76%) using generative AI tools. These products run the gamut from productivity software with generative AI integrations, to dedicated generative AI software for specific functions. Among this cohort, the most popular tools used are internal self-service data chatbots, used by 88% of respondents, and tools that draft or summarize documents with 77% adoption. Similarly, 71% of respondents in the technology service delivery function (IT professionals) incorporate generative AI tooling into the delivery of broader services.
“Casetext saves our attorneys two to three hours per document. This has allowed us to lower costs in delivering these types of services.”
Head of Technology Office, Law Firm
Our conversations with executives reveal many niche or industry-specific use cases, suggesting that generative AI solutions are increasingly tailored to the needs of various verticals. IT and financial services have the highest adoption rates, with 81% of respondents in the sector using generative AI tooling. Media, advertising, and telecommunications have similar rates of AI tool use.
The pharmaceutical industry has already started leveraging generative AI tools to model molecular structures and assist in the research and development of new medications. A prime example is the Dutch startup Cradle, which uses generative AI to design and engineer proteins and works with major biotech players.
“Using GitHub Copilot, two developers can achieve the workload of 10 developers.”
Director of Data Science, Pharmaceutical Company
Financial service companies have also made strides in sector-specific generative AI tools, creating models for in-depth market analysis. Top Wall Street firms are investing heavily in generative AI tooling. JPMorgan launched a proprietary LLM Suite that is said to do the work of a research analyst. Similarly, Morgan Stanley débuted "Debrief," a generative AI assistant that assist the firm's wealth advisors.
“We used general and off-the-shelf generative AI tools in many departments, but we trained and customized models of our core functions.”
Head of Technology Office, Financial Services Company
Despite the surge in enterprise adoption and the emergence of specific business use cases, the risks inherent to the technology remain. Hallucinations, where models produce false or misleading information, pose a serious concern across all industries, despite assurances from some industry experts that they will be swiftly resolved. What's more, 72% of respondents worry about data security. Many are concerned that using an off-the-shelf model could expose their sensitive data to third parties with poor intentions (sometimes including the model provider themselves for future training purposes). Additionally, there is growing concern that many enterprise applications would unintentionally capture too much data and communications, exposing companies to unnecessary legal risk. These concerns underscore the importance of thoughtful generative AI governance policies and rigorous negotiation of terms of use with model providers.