How Forerunner Approaches Responsible AI in Government

At the 2026 Forerunner Summit, Jackie Teran, Forerunner’s VP of Engineering, addressed a challenge facing public agencies as they explore AI: how can they benefit from the technology without compromising accountability? AI can make information easier to access and reduce time spent reviewing records, but inaccurate outputs can have consequences for compliance, service delivery, and the residents agencies serve.

For government teams, adopting AI responsibly means more than making work faster. It requires clear safeguards that protect sensitive information, make AI activity traceable, and ensure people remain responsible for the decisions they make.

Three principles of responsible AI

Teran outlined three principles that guide Forerunner’s approach, each designed to help agencies use AI without losing control over their information or decisions.

• Human oversight: AI can draft letters, populate fields, and summarize records, but staff remain responsible for reviewing and approving its outputs. The goal is to support professional judgment, not replace it.
• Verifiable and traceable activity: AI requests, responses, and subsequent activity are logged, creating a record that supports audits, public records requests, and ongoing quality checks.
• Data integrity: Agencies retain ownership of their data, and each organization’s information remains separate from that of other customers.

Together, these principles establish the foundation for using AI in ways that support public sector work while preserving accountability.

Putting safeguards into practice

These principles also need to be backed by technical controls. Forerunner encrypts data at rest and in transit, uses role-based access controls, and supports single sign-on integrations. The company holds SOC 2 Type II and TX-RAMP Level 2 certifications, which provide independent assurance around its security controls.

Teran also explained that AI features undergo security reviews before release, with ongoing audits to monitor performance. Forerunner is pursuing FedRAMP authorization, with a target of early 2027.

These measures help protect agency information, but safeguards are only part of the equation. Agencies also need to understand what AI can do, where its limits lie, and when human judgment must take over.

Making information easier to find, not making decisions for staff

Forerunner’s AI is designed to help staff find and interpret information in existing records rather than make official determinations. Teran illustrated this distinction with a practical example: a staff member could ask which properties within a flood zone are missing elevation certificates and receive a consolidated list with citations linking to the supporting records.

Rather than searching through files one by one, staff can use that time to verify the information and determine what action is needed. The AI helps surface relevant records, but it does not independently determine flood zone designations or replace the professional judgment required for official decisions.

This distinction is central to Forerunner’s approach. AI can reduce the effort required to find information, while agency staff retain responsibility for interpreting it and deciding what happens next.

Building trust through transparency

Teran closed by reinforcing Forerunner’s commitments to transparency, human oversight, and ongoing testing for accuracy and fairness. AI activity remains identifiable and traceable, allowing teams to investigate issues when they arise.

Forerunner also commits to not selling agency data, sharing it with third parties, or using it to train AI models. Agencies retain ownership of their information throughout the process.

Ultimately, Teran’s message was that responsible AI is not defined by a single feature or certification. It requires an ongoing commitment to security, transparency, and human accountability. For public agencies, success means making information more accessible and reducing repetitive work without sacrificing the safeguards and professional judgment that residents depend on.

Watch the session here: