NationGraph maps the US public sector through Perplexity’s API

The Problem

For businesses trying to sell to US government agencies, getting data on their targets is difficult. While US federal data is often centralised, the state, local and education sectors are an administrative labyrinth. With more than 100,000 independent entities currently across US governments, including 12,000 school districts, there’s no standardised data structure.

NationGraph’s customers span education, public safety, transport, cybersecurity and IT, and the finance and administration industries. Its platform is designed for revenue leaders, account executives and sales development representatives who are doing research, looking to displace competitors and/or validating budgets.

Before adopting Perplexity’s API, NationGraph faced three key challenges:

• Fragmentation: Each municipality hosts its own transcripts, budgets and meeting minutes on distinct web structures.
• Research: Identifying simple details, such as the ERP software a city uses or the size of a local vehicle fleet, required hiring dozens of contractors to perform manual click-and-inspect research.
• Limited tools: There’s no real infrastructure to analyse a county or a school district with the same granularity that Wall Street analysts use to evaluate public companies.

NationGraph needed a research engine that could search the internet at scale, verify facts through citations and normalise unstructured data into actionable workflows.

The Solution

Perplexity helps NationGraph build data pipelines into the publicly accessible data that government entities publish. Specifically, NationGraph uses the Perplexity Search API and Deep Research API to offer a research engine for its customers, making Perplexity a core pillar of NationGraph’s research infrastructure. Unlike standard LLMs that might hallucinate, lose context or provide outdated information, Perplexity works with NationGraph’s mapped universe of government entities, fuelling NationGraph’s customer growth.

Each NationGraph user can thus run hundreds of parallel Perplexity-powered queries to find the specific information they need.

Key technical advantages:

• Deep Research: NationGraph used Perplexity’s research capabilities to balance speed and cost. While other, more expensive models might take 20 minutes to produce a lengthier report, Perplexity delivered high-fidelity results 10x faster, in two to three minutes. That made it commercially viable for NationGraph to integrate Perplexity so that its users can use Perplexity directly within its product.
• Source verification: Perplexity’s ability to cite specific government sources such as filings, documents and meeting minutes was key. This allowed NationGraph to maintain a feedback loop to validate its results and keep them accurate.
• Handling complexity: Perplexity excelled at navigating paginated government websites and unstructured documents, extracting the niche details that NationGraph’s customers are trying to understand. For example, Perplexity makes it possible for a NationGraph customer to find the addresses, meal schedules and mail regulations of all municipal correctional facilities.

These features gave NationGraph the confidence to embed Perplexity directly into its Workspace product. Users can now build a spreadsheet of 300 institutions and run a Smart Column query such as “Who is the current Director of IT and what are their current CMS (content management system) priorities?” across the entire list.

One NationGraph customer wanted to identify school districts in need of help transporting their pupils to school. That meant issues with maintaining fleets, routing and bell times, compliance and incident response, parent communications, recruiting and retaining qualified drivers, or even overspending with existing vendors. Perplexity helped NationGraph perform the deep research needed to produce a high-quality dataset. “This was just a problem that without tools like Perplexity, we just don’t even know how we would search,” says Eden Ding, CTO of NationGraph. “We would have had to manually comb through everything.”

The Result

By offloading the heavy lifting of web-scale research to Perplexity, NationGraph is able to offer public sector intelligence more quickly and cheaply.

By the numbers

• 500,000+ queries: NationGraph has run over half a million Perplexity-backed research queries to date.
• 350,000+ Smart Columns: These queries power the automated research layers within the NationGraph Workspace.
• 2 months saved: Integrating Perplexity’s search infrastructure saved the engineering team approximately two months of development time that building a custom web agent would otherwise have required.
• Days to minutes: For a NationGraph end user, researching 300 institutions used to take days of manual clicking; it now takes less than two minutes.

Strategic impact

Thanks to the Perplexity integration, NationGraph can focus more on its product. Instead of maintaining research infrastructure, its engineers focus on building product features that help vendors, agencies and the public interact with a more transparent system. NationGraph and Perplexity have turned what were once “impossible” data projects into one-shot research runs.

“When our customers perform a search, NationGraph brings its data and Perplexity augments the public side of each query,” Eric says. “The end result is that customers can save tens of thousands of dollars and months of time by asking just a handful of questions.”