Closing the AI gap: How next-generation knowledge access unlocks mission outcomes for government

Next-generation knowledge access represents an advanced approach to finding, understanding, and using information, moving beyond traditional search methods and static knowledge repositories.

Within the public sector, this translates to AI systems and operators gaining access to the right knowledge at the right time and with the right context, eliminating the need to manually search across disconnected sources.

Fiscal strain is intensifying across many regions. Across the 193 UN member states assessed, investment in resilient infrastructure and advanced technologies continues to grow.² Leaders are being asked to deliver better outcomes with fewer resources, and they're turning to AI to close that gap.

The strategic implication for public sector leaders is clear. AI isn't a discretionary innovation initiative anymore. It's increasingly becoming a policy and service delivery priority, and agencies that can't operationalize it will face growing pressure from oversight bodies, elected officials, and the public they serve.

The 72% figure deserves closer examination. IDC's research identifies the primary barriers to scaling AI: skill gaps, security and compliance risks, and data readiness.¹

Traditional data warehouses were built for periodic analysis of curated, structured data. Agentic AI depends on fast, governed access to a much broader knowledge surface that could include:

• Policies and procedures

• Case notes and correspondence

• Call transcripts, forms, and operational logs

• Real-time cybersecurity threat signals

Most of this content is unstructured and sits in siloed systems, making it difficult to correlate and analyze. It also may not be discoverable or usable by AI agents in its current state.

The challenge often emerges at the retrieval layer. Without governed access to authoritative knowledge, AI applications may produce less reliable outputs, making it harder for agencies to build and maintain public trust.

Regulations are beginning to define the requirements for deploying AI in production. The EU AI Act and Interoperable Europe Act reflect a broader global trend toward formal standards for trustworthy, interoperable digital public services. At the same time, digital and AI sovereignty are increasingly influencing procurement decisions and operational requirements.

According to IDC research, public sector organizations scaling generative and agentic AI prioritizing:

• Data residency guarantees (46%)

• Safety guardrails, including retrieval-based grounding and policy filters (43%)

• Auditability of prompts and actions (40%)

• Human-in-the-loop approvals (39%)

• Model provenance and software bill of materials (38%)¹

These requirements need to be considered from the start, not added after deployment. They shape architecture, tool selection, and system design. As IDC notes from conversations with senior officials, policymakers want AI infrastructure they can "control in time of crisis," reflecting a broader shift from data location requirements to strategic autonomy over AI roadmaps.¹

For decision-makers, this means evaluating governance and sovereignty requirements early in the selection process. Model performance matters only if the technology can operate within the agency’s compliance requirements.

Three converging trends are shaping this evolution:

• Hybrid retrieval as a standard pattern: Hybrid search blends keyword (lexical) and semantic search to support RAG and agentic workflows, balancing exact matches with contextual understanding.

• Open integration protocols: Standards, such as the Model Context Protocol (MCP) — an open specification for connecting large language model (LLM) applications to external data sources and tools — accelerate integration while raising governance requirements around what's connected and how it's logged.

• Security and risk controls as safety mechanisms: Permission filtering, redaction, audit logs, and provenance tracking align retrieval architecture with frameworks, such as the National Institute of Standards and Technology (NIST) AI Risk Management Framework.³

Together, these capabilities transform search into an AI-ready retrieval service that can keep data in compliant locations while enabling controlled, low-latency access across distributed systems.

When retrieval is governed and consistent, the benefits extend across the organization. For IT teams and developers, standardized connectors, indexing approaches, and retrieval APIs reduce one-off integration work and accelerate delivery across projects. For security and governance teams, they can limit unnecessary data copying and make information access easier to audit.¹

Agencies can expect:

• A unified operational picture that breaks down data silos, improves visibility, and gives teams the context they need to make faster, more informed decisions

• Efficiency gains from consolidating fragmented search and retrieval tooling

• Reduced compliance exposure through built-in permission filtering, redaction, and audit logs

• Faster time to production for AI use cases across government cybersecurity, case management, and citizen services

• Strategic autonomy over AI roadmaps, with the ability to deploy in sovereign or on-premises environments as policy requires

These outcomes matter because oversight bodies and elected officials will ultimately look for evidence of governance, auditability, and mission impact. An AI investment that can’t demonstrate these outcomes may be at risk.

The research points to a clear strategic direction. Scaling AI isn't primarily a model selection problem. It's an architecture problem, and the retrieval layer plays a critical role.

For public sector leaders, three priorities emerge from the IDC findings:

• Treat data readiness as a prerequisite, not a parallel workstream. Before expanding AI pilots, invest in the governed retrieval foundation that will support them at scale.

• Design for sovereignty from the start. Data residency, auditability, and human oversight aren't compliance checkboxes. They're architectural requirements that shape platform selection.

• Choose platforms built on open standards. Interoperability protects against lock-in and preserves strategic autonomy as policies and geopolitical conditions evolve.

Agencies that act on these priorities now will be the ones delivering measurable mission outcomes with AI over the next two to three years.

The full IDC Spotlight report, cosponsored by Elastic and AWS, provides the complete survey data, global policy analysis, and architectural guidance referenced throughout this piece. Download the IDC Spotlight report: Closing the AI gap in government

• Explore public sector AI use cases from Elastic.

• See how public sector organizations use Elasticsearch.

• Read the article on GenAI and semantic search in the public sector.