.webp)
At Re:imagine 2025, Gartner’s Danielle Casey delivered a clear roadmap for product and technology leaders navigating the generative AI curve: not all use cases are created equal—and not all will succeed. Drawing from hundreds of case studies across industries, the session broke down where GenAI is already delivering value, where it’s just beginning to show promise, and where adoption may never scale due to complexity, risk, or lack of ROI. For product leaders, the takeaway was simple: If you’re not being deliberate about use-case strategy, you’re already falling behind.
Where generative AI works today
The majority of current enterprise deployments fall into a tight band of feasibility:
- Low complexity
- Moderate value
- Quick to implement
Think content generation, summarization, retrieval, and surface-level customer interactions.
Gartner spotlighted:
- A Fortune 50 automaker using GenAI to generate campaign visuals at scale
- A healthcare provider moving from basic note summarization to discharge prediction and risk modeling
- A global travel company building a GenAI-based booking agent that increased valid bookings tenfold
The lesson? Start with simple use cases, but plan for scale.
Where GenAI is headed: three technologies to watch
Gartner identified three forces accelerating the next wave of enterprise AI:
1. Domain-specialized language models (DSLMs)
Forget general-purpose LLMs. DSLMs are:
- Trained on industry, function, or task-specific data
- More accurate, more efficient, and faster to deploy
- Better suited for vertical workflows and privacy-sensitive environments
Example: A document LLM designed to understand complex financial documents by reading both the text and the document layout. It outperforms general AI models in tasks like contract analysis and compliance, helping teams work faster and more accurately. DSLMs enable smaller, cost-effective models tailored for real-world business logic over general knowledge.
2. Multimodal interfaces
Gartner projects that by 2030, nearly every enterprise system will support multimodal interaction. That includes:
- Text
- Voice
- Charts
- Tables
- Maps and visual data
One example: a Canadian wealth management firm using GenAI to process and generate reports across text, tables, and charts. Cutting report time by 80%. It expands automation potential by up to 50%, unlocking tasks that weren’t previously AI-compatible.
3. Agentic AI
This is where automation becomes intelligent.
Gartner defines agentic AI by six traits: goal-setting, planning, autonomy, collaboration, reasoning, and adaptability. It’s a shift from “responding to inputs” to executing toward outcomes.
Example: an Australian water utility using three autonomous agents managing water levels, optimizing energy usage, and scheduling pump maintenance—all operating with interdependent goals.
Where gen AI might not work (yet)
Gartner called out barriers that are slowing or stalling adoption:
Market
Interoperability suffers when AI agents don’t speak the same language. Without common protocols, collaboration between specialized and general systems is difficult.
Business
Organizations still struggle to tie GenAI to measurable outcomes. Many pilot programs look impressive but fall short of proving sustained value or ROI.
Technology
Not every task fits a GenAI-first approach. For use cases requiring ultra-high accuracy (e.g., prediction, simulation, digital twins), hybrid models—rules-based, classical ML, and neuro-symbolic AI—are still essential.
What enterprises should do next
Gartner offered three actions to focus on now:
1. Audit your current GenAI use cases.
Look beyond volume. Are they delivering ROI or just outputs?
2. Prioritize trust and control.
Adopt platforms that balance automation with governance, observability, and model flexibility.
3. Invest in the enablers of scale:
- Domain-specialized models
- Multimodal UX
- Agentic architectures that grow with you
Conclusion: Kore.ai’s take
The message is clear: success in AI won’t come from isolated use cases; it will come from how intelligently and intentionally organizations build. At Kore.ai, we’re aligned with Gartner’s vision and proud to support enterprise teams in deploying systems that are not just generative but orchestrated, agentic, and ready for real-world complexity. If you missed the keynote, now’s your chance to catch up.
