As organizations scale generative AI (gen AI) proof-of-concept projects to full-scale deployment, industry-specific applications are driving maximum business value compared to the initial flight of generic, horizontal use cases.
That was a central finding from the Amazon Web Services (AWS) Generative AI Customer Research Study. The research, which canvassed 1,000 respondents from 10 countries across Europe and North America, unearthed key shifts as organizations move beyond the rush of ad hoc gen AI experimentation to building and deploying enterprise-grade use cases.
Maturity levels are rising as companies scale and embrace more formalized practices in pursuit of optimal business value from gen AI. According to the research, 53% of respondents are deploying or have fully deployed gen AI use cases compared to 47% in 2023.
Also, 45% of IT decision-makers are now prioritizing gen AI tools over other IT spending categories, including security, as part of budget planning, according to the AWS Generative AI Adoption Index.
Companies that several years ago kicked off their AI journey are now starting to productionize applications in earnest. Domain-specific models and industry-focused use cases are having the greatest impact reimaging workflows, addressing long-standing bottlenecks, and unlocking measurable business value.
Gen AI: A specialty problem solver
IT and business leaders are creating adoption blueprints around sector-specific use cases rather than generic, horizontal use cases. For example, domain-specific models help address the unique challenges of a particular industry sector whether it be healthcare, finance, retail, entertainment, or industrial manufacturing. This new class of models leans into specialized datasets and architectures, resulting in superior accuracy and performance for specialized tasks over general-purpose models, according to research from Boston Consulting Group.
Some examples of domain-specific models include:
- AlphaFold, which predicts 3D protein structure for biochemistry and drug discovery
- BloombergGPT, for summarizing reports and generating insights from financial market data
- CoCounsel, which leverages domain-specific legal data to generate legal documents and summaries
By taking an industry approach and leveraging domain-specific models, users can achieve:
- Enhanced task accuracy, thanks to optimization for industry-specific nuances
- Reduced redundancy via streamlined architectures that eliminate unnecessary computations, thus cutting inferencing costs
- Faster implementation by reducing the need for fine-tuning
- Improved explainability, which is essential for building trust
Consider the retail industry, where automated on-model fashion image generation is resulting in a 1.5x bump in retailer conversion rates. In the healthcare and pharmaceutical space, a company was able to identify a new drug candidate for idiopathic pulmonary fibrosis treatment in 21 days—a process that normally takes years using traditional methods. Customer service workflows are ripe for improvement; for example, platforms in the insurance technology industry are leveraging gen AI for use cases that cut support costs by nearly 30%.
Other key takeaways include:
- Cost considerations are now less critical for initial experimentation compared to the 2023 findings, yet they remain the primary factor when scaling workloads. ROI has gained importance as the technology matures.
- Cybersecurity is a big factor determining tool and partner choice. Strong cybersecurity is a “non-negotiable” factor in partner selection, even more so than customer service and best-in-class models.
Unleashing gen AI on industry-specific problems and pain points is the direct route to rescripting and revitalizing how companies do business. AWS and its partner ecosystem deliver the broadest spectrum of technologies and domain expertise to help IT leaders achieve the business outcomes that shareholders, executives, and customers expect.
Click here to learn how AI technologies—from generative and agentic AI to intelligent automation—are transforming enterprise operations across industries.