2 months ago
62
Learn astir an AI-powered acquisition launched arsenic a minimum viable merchandise to stitchery existent idiosyncratic feedback and iterate rapidly.
When we speech astir agentic AI, it’s casual to default to abstract conversations astir models, prompts, and orchestration. But the astir compelling stories I spot are the ones wherever AI unlocks thing profoundly human—creativity, intuition, and expertise—at wholly caller velocity and scale.
That’s wherefore I was excited to big Color Meets Code: Pantone’s Agentic AI Journey connected Azure, a webinar featuring 2 Pantone leaders, Kristijan Risteski, solutions architect, and Rohani Jotshi, elder manager of engineering. During the session, Kris and Rohani shared however they’re applying agentic AI to 1 of the astir foundational elements of originative work: color—and however an AI-ready database, Azure Cosmos DB, plays a cardinal relation successful making that possible.
The challenge: Scaling colour expertise successful a real-time, interactive world
Pantone is wide recognized arsenic a planetary authorization connected color. For decades, their teams person combined quality expertise, colour science, and inclination forecasting to assistance designers and brands define, communicate, and power colour crossed industries—from manner and merchandise plan to packaging and integer experiences.
But arsenic Pantone explained successful the webinar, translating that extent of expertise into a modern, conversational AI acquisition came with existent challenges. Creating colour palettes is some clip consuming and captious to the plan process. Designers often stitchery inspiration by navigating betwixt tools, colour pickers, and inclination reports earlier they ever onshore connected a usable palette.
Pantone saw an accidental to rethink that workflow entirely: What if designers could interact with decades of Pantone research, inclination data, and colour science done a chat-based interface—and make curated palettes instantly?
Introducing the Palette Generator: An agentic AI experience
The effect is Pantone’s Palette Generator, an AI-powered acquisition launched arsenic a minimum viable merchandise to stitchery existent idiosyncratic feedback and iterate rapidly. Rather than offering static recommendations, the Palette Generator uses multiagent architecture to respond dynamically to idiosyncratic intent, conversational context, and humanities interactions.
In the webinar, the Pantone squad described however they designed the strategy to see specialized agents—such arsenic a “chief colour scientist” cause and a palette procreation agent—each liable for antithetic aspects of reasoning, discourse retrieval, and effect generation. These agents enactment unneurotic to present curated colour palettes that bespeak Pantone’s proprietary information and expertise.
What stood retired to maine was not conscionable the sophistication of the AI, but the architectural subject down it. Agentic AI isn’t conscionable astir models—it’s besides astir data.
Why Azure Cosmos DB was foundational
At the bosom of Pantone’s Palette Generator is Azure Cosmos DB, serving arsenic the system’s real-time information layer. Azure Cosmos DB is utilized to store and negociate chat history, punctual data, connection collections, and idiosyncratic enactment insights—all of which are indispensable for responsive, fast, context-aware agents.
As we did our probe to find the champion persistence storage, we explored antithetic databases. What we recovered for Azure Cosmos DB was however casual it was to integrate it into our systems. We were capable to marque our archetypal impervious of conception with a fewer lines of codification and retrieve each the information very, precise fast, similar successful a fewer milliseconds.
Kristijan RisteskiAzure Cosmos DB was besides chosen due to the fact that of its scale, allowing Pantone to service users each implicit the satellite with accelerated information retrieval.
This is simply a captious point. As applications displacement from “doing” to “understanding,” databases indispensable enactment acold much than elemental transactions. They request to grip monolithic volumes of operational data, accommodate arsenic AI workflows evolve, and enactment precocious scenarios similar conversational memory, analytics, and vector-based search.
Pantone’s architecture demonstrates what it means to beryllium “AI-ready.” Azure Cosmos DB provides the scalability and flexibility needed to way idiosyncratic prompts and conversations crossed sessions, on with analytics that assistance Pantone recognize however customers interact with the Palette Generator implicit time.
From substance to vectors—and what comes next
Another penetration Pantone shared during the webinar was however their architecture is evolving to amended relevance, accuracy, and contextual understanding. While the existent strategy already supports affluent conversational experiences, the squad outlined adjacent steps that impact moving from accepted substance retention to vector-based workflows. This includes embedding idiosyncratic prompts and contextual data, allowing for vector search, and enriching responses with deeper semantic understanding.
Azure Cosmos DB plays a relation present arsenic well, supporting vectorized data, integrating with cause orchestration, and embedding models deployed done Microsoft Foundry. This allows Pantone to iterate without rearchitecting the full system—an indispensable capableness erstwhile moving successful a fast-moving AI landscape.
Real-world results from agentic architecture
Pantone didn’t conscionable speech astir vision—they shared factual results from existent usage of the Palette Generator. According to the webinar data, users crossed much than 140 countries engaged with the tool, generating thousands of unsocial chats wrong the archetypal period of merchandise and interacting successful dozens of languages. The strategy observed aggregate queries per idiosyncratic session, indicating that designers were actively experimenting, refining prompts, and exploring ideas conversationally.
Just arsenic importantly, Pantone emphasized however rapidly they’ve been capable to larn and adapt. Prompt sensitivity, idiosyncratic behavior, and architectural tradeoffs astir speed, cost, and reliability each informed ongoing refinements. Azure Cosmos DB’s flexibility made it imaginable to seizure these insights and germinate the acquisition without slowing innovation.
Lessons for anyone gathering agentic AI
Pantone’s travel reinforces respective lessons I spot repeated crossed customers gathering AI agents connected Azure:
- Agentic AI is inherently information driven. Without a real-time, scalable database layer, adjacent the astir precocious models conflict to present consistent, context-aware experiences.
- Feedback loops matter. By capturing prompts, responses, and idiosyncratic interactions successful Azure Cosmos DB, Pantone tin continuously amended some the AI and the merchandise acquisition itself.
- Flexibility is nonnegotiable. AI architectures germinate quickly—from orchestration patterns to embedding strategies—and databases indispensable germinate with them.
What Pantone has built with the Palette Generator is much than a feature; it’s a blueprint for however organizations tin construe heavy domain expertise into intelligent, agent-driven applications. By combining Microsoft Foundry, Azure AI services, and an AI-optimized database similar Azure Cosmos DB, Pantone is showing however creativity and exertion tin determination guardant together.
As much organizations clasp agentic AI, the question won’t beryllium whether they tin deploy models—but whether their information foundations are acceptable to enactment real-time understanding, memory, and scale. Pantone’s travel makes that reply clear: AI-ready applications commencement with AI-ready data.