1. What motivated your organization to participate in this year's Global AI Awards?

Enterprise agentic AI deployments are failing, not because the models are wrong, but because the data layer underneath them wasn't built for autonomous systems. A leading global professional services firm put the requirement plainly in one of our early architecture conversations: "We need to run address standardization, geocoding, and risk enrichment across millions of policyholder records in a single call, and we need to be able to audit every step." Speed, reliability, and auditability at enterprise scale, that's what agentic AI demands of its data infrastructure, and it's not what most APIs were designed to deliver. We entered these awards because our agentic AI developer portal, Data Graph API, and MultiPass API represent a genuine answer to that problem, and we wanted to invite honest scrutiny from the broader community.

2. Could you give us an overview of the AI solution or breakthrough you submitted for consideration?

Our submission addresses the core infrastructure challenge of enterprise agentic AI: getting reliable, contextually enriched data into autonomous systems at the speed and precision they require.

The Data Graph API is a GraphQL API for AI agents that lets them traverse Precisely's interconnected global data assets, including addresses, property data, demographics, risk scores, and points of interest, in a single declarative query. Instead of chaining multiple point APIs, an agentic AI system expresses what it needs and Data Graph resolves the relationships, cutting latency and round-trips significantly. The MultiPass API is an AI API orchestration engine, chaining Precisely and external services into a single executable workflow with conditional logic, retry and fallback, JSON transformation, and structured output that serves enriched data attributes directly as ML-ready features for model consumption. A global enterprise customer runs it in production today across millions of records. Together, they form an agentic AI data layer built to match how autonomous systems actually operate at enterprise scale.

3. How did your team collaborate to develop and refine this AI innovation?

Both products were shaped by what agentic AI actually demands in production: speed, reliability, and auditability at enterprise scale. Those requirements were built into the architecture from the start rather than retrofitted, and it shows in how they perform at scale. The Data Graph API was co-designed with geospatial technology partners and validated for insurance-specific agentic workflows — real partner feedback shaped the schema and query model before general availability. MultiPass was driven by production requirements in a regulated industry, where deterministic AI API orchestration, observability, and compliance weren't optional.

4. What impact do you expect your AI work to have on the broader AI community or society as a whole?

As agentic AI systems move from answering questions to taking actions, pricing insurance policies, routing emergency resources, approving or declining financial products, the cost of data errors rises sharply. A flawed recommendation is recoverable. A flawed automated action, taken at speed and at scale, may not be. By providing an agentic AI data layer purpose-built for autonomous system consumption, we're working to make enterprise agentic AI more reliable and more accountable in the industries where that matters most. The MCP-first architecture of our agentic AI developer portal, connecting AI agents directly to live, authoritative data through open-source servers, also points toward a broader shift: agents acting on the real world with confidence rather than on training data with uncertainty.

5. Were there any notable challenges during the development of this AI solution, and how did you overcome them?

Three, each genuinely hard in different ways. Unifying thousands of independently evolved datasets under a single GraphQL API for AI agents required an enormous data engineering effort. The PreciselyID, our persistent location identifier, was the architectural anchor that made it possible, but mapping every dataset to it took years of work that doesn't show up in a demo. Building the MCP Playground in our agentic AI developer portal to work against live data with no mocks introduced reliability requirements a typical sandbox never faces, including query isolation, rate management, and result consistency, because developers needed to trust that playground behavior reflected production exactly. And the AI API orchestration requirements for MultiPass, driven by regulated-industry compliance, pushed us to build auditability and observability well beyond what a standard orchestration layer provides. Production deployments in highly regulated industries wouldn't have been possible without that work.

6. How does your organization nurture a culture that drives continuous AI innovation?

Partners and customers see it most clearly in how we run design partnerships. Feedback from early adopters influenced the Data Graph API schema, while enterprise production requirements helped shape our approach to agentic AI orchestration. Those aren't testimonials. They're evidence of a development process where external, real-world requirements get built in rather than accommodated after the fact.

Internally, product, data, and engineering teams are evaluated together on whether their outputs are actually usable in enterprise agentic AI pipelines, not just functionally correct in isolation. That creates a different kind of accountability than most organizations operate with, and it produces different results.

7. What advice would you offer to teams or companies aiming to make meaningful contributions in the AI space?

Start with data before building agentic AI agents. The organizations making the most durable contributions to enterprise agentic AI right now are the ones solving data access and quality problems upstream of deployment. An agent operating on inaccurate data doesn't just produce a bad output, it takes a bad action, potentially at scale and at speed. Build for auditability from day one, especially if you're targeting regulated industries. Deterministic AI API orchestration and observability are not features you can add later. And build for the ecosystem through open standards and genuine design partnerships. The Model Context Protocol is a recent example of how early adoption and open contribution can connect your agentic AI infrastructure to workflows you never anticipated.

8. What are your organization's long-term goals in AI, and how are you planning to advance the field moving forward?

Our long-term thesis is that trusted, location-anchored data will be the foundation of reliable enterprise agentic AI. Concretely: expanding the Data Graph API to cover more global geographies and data categories, extending MultiPass to support more complex agentic AI workflow patterns including human-in-the-loop steps, and growing our MCP server library so more agentic AI platforms and agent environments can connect to Precisely data natively. We're positioning our APIs to be the agentic AI data layer that autonomous systems reach for when they need to act on the real world, not a best-guess approximation of it. The success of enterprise production deployments is the benchmark we hold ourselves to as we build toward that goal.

9. Are there any emerging AI technologies or trends your team is particularly excited about right now?

The Model Context Protocol tops the list — not as a trend, but as a genuine infrastructure shift for agentic AI APIs. The ability for AI agents to call live, authoritative data sources rather than hallucinate from training data is the most consequential change in applied AI in the past two years, and it's the architectural principle behind everything we've built in our agentic AI developer portal. The other thing we watch closely is the convergence of spatial reasoning and large language models. As enterprise agentic AI systems increasingly act on location-dependent decisions — routing, risk pricing, logistics, emergency response — spatially enriched data delivered through a GraphQL API for AI agents becomes a critical input to reliable autonomous action. That convergence is still early, but the direction is clear, and we're already there.

To dive deeper into Precisely's award-winning work, visit their website at https://www.precisely.com/