Generative AI adoption in numbers, it’s growing…

The adoption of generative AI (GenAI) is set to be significantly influenced by domain-specific knowledge management and the rise of agentic AI assistants over the next three years. I think this evolution requires product leaders to focus on developing multimodal capabilities and agentic workflows, integrating domain models to achieve autonomous decision-making in AI solutions and enhanced return on investment (ROI).

Let’s look at some of our key trends and strategic planning assumptions:

Domain-specific GenAI models will dominate; by 2027, over 50% of GenAI models used by enterprises will be domain-specific, a dramatic increase from just 1% in 2024. This shift indicates a move towards more tailored and effective AI applications.

Conversational AI and search engines will become more GenAI-driven; it's expected that by 2028, more than 60% of search engines will be GenAI-driven and conversational, compared to less than 20% today. Furthermore, 75% of conversational AI offerings will support bidirectional multimodal user interfaces beyond text and voice, a substantial increase from less than 10% in 2024.

Virtual assistants (VAs) will become more pervasive: By 2028, GenAI-enabled VAs are predicted to support 80% of knowledge workers' tasks, up from 30% in 2024.

Simulation data will underpin strategic decisions: By 2030, 20% of strategic business decisions will be underpinned by simulation data, a significant rise from 1% in 2024. This reflects a growing reliance on synthetic data for scenario exploration.

Multimodal capabilities will be crucial for content creation and discovery, enabling industries to leverage diverse media formats for enhanced engagement and analytics. This will facilitate comprehensive analysis and streamlined data extraction across various sectors. The future of content creation and discovery is tied to the use of more than one modality in model input, output or both, with the market expected to see a rise in solutions that accommodate multiple modalities at both stages.

Agentic AI is an emerging capability poised to revolutionise business process automation. By integrating agentic workflows into knowledge management, virtual assistants and software development, organisations can boost productivity, enable autonomous decision-making and foster innovation. Agentic AI workflows enable action-oriented solutions and provide full automation and decision-making capabilities at scale.

Synthetic data generation and scenario exploration in simulation use cases are set to transform industries by improving data privacy and facilitating advanced predictive analytics and AI model training. This will allow for enhanced R&D and decision making for complex challenges more efficiently and cost-effectively.

I want to highlight that GenAI adoption is currently in a second wave, with knowledge management and search now the leading use case. This signifies a shift from content generation to content discovery, effectively leveraging existing knowledge assets. The technology has also seen rapid implementation across industries, moving quickly from proof of concept to production.

GenAI is transforming software development, with a strong emphasis on code generation, completion and test case creation. The comprehensive integration of GenAI across the entire software development lifecycle (SDLC) is anticipated, including automation of specifications, documentation, coding assistance, testing, debugging and maintenance.

Furthermore, GenAI adopters are seeing tangible business value including:

Operational improvements, such as increased employee productivity and faster resolution times.

Enhanced competitive differentiation, through improved customer experience and engagement.

Improved go-to-market strategies, by growing sales and reducing time to market. Specifically, employee experience is being improved through GenAI-enabled VAs which provide better and faster access to information, while customer experience is enhanced through automated support and reduced response times.

Product leaders should focus on domain-specialised models to offer improved accuracy and efficiency at lower costs. They must also communicate the business value of their GenAI offerings, emphasising successful use cases with quantified value. It is essential to invest in systems that support diverse data types as this is crucial for maintaining competitiveness.

Organisations are adopting a phased approach to GenAI, starting with limited rollouts and iterating to address issues. Once a successful GenAI business value story is established, adopters are eager to expand its use. To facilitate adoption, comprehensive strategies including user training, data management and cost-effective technologies are necessary to overcome challenges.

In summary, to remain competitive product leaders should focus on smaller, domain-specific models, multimodal capabilities, and advanced features like agentic AI workflows. Additionally, they should explore opportunities in simulation, synthetic data generation, and scenario exploration. Long-term, maintaining cost optimisation, security, privacy, and compliance is critical.

This analysis is based on a six-month Gartner case-based research project, involving over 100 interviews and analysis of nearly 190 adopter use-case studies.