"Across our service footprint, we reach one in three Americans, handling everything from public health insurance enrollment to unemployment benefits to national emergency support services." — Mike Raker, Chief Technology Officer, Maximus
Conversation as the new front door
For more than two decades, federal websites served as the digital portal for citizens seeking services and information. That model—measured by navigation, load times, mobile responsiveness and Section 508 compliance—served a browser-centric era. According to Mike Raker, that era is shifting: conversation, driven by AI, chatbots and voice assistants, is becoming "as common a path as the web page." The result is a move from a channel strategy to an omnichannel experience strategy that centers on conversational interfaces, a medium the source says humans prefer.
How conversation works in practice
Raker frames conversation as the natural resolution mechanism for higher‑stakes interactions: people pick up a phone or invoke a voice interface when questions are complex. Across examples cited, end users now ask AI assistants how to check a benefit status, find a clinic, or apply for unemployment and receive direct answers. The shift narrows the distinction between human‑mediated and AI‑mediated interactions—not only on speed and efficiency but on the core attributes that define good service.
- Understanding context
- Delivering accurate answers
- Showing empathy
- Adapting to the unexpected
As those attributes improve in AI, the technology simply changes the mechanism; the human preference for dialogue remains constant.
Federal API modernization, identity, and auditability
Raker argues that the rise of conversational delivery requires different technical foundations than public web pages. Federal API modernization is described as foundational: secure, machine‑readable service access across channels and AI assistants. Government data interoperability should extend that foundation to include authentication and identity frameworks built for delegated access on a person’s behalf, along with logging and audit trails that make AI‑mediated transactions traceable to the same standard as human‑mediated ones.
Accessibility obligations, Raker notes, extend into this conversational and API layer as well: people using screen readers, voice assistants, translation tools, caregivers, or personal AI agents are entitled to equal access, reliable outputs, auditable decisions, and a clear path to challenge errors.
Human‑in‑the‑loop government and contact center modernization
The piece positions human authority and frontline oversight as central to AI governance in government. On the front line, AI is presented as a tool that surfaces case history and evidence faster: a contact‑center representative may see relevant information appear before a caller finishes saying hello; an eligibility examiner may see specific evidence to match a rating criterion. That routine work being handled by AI frees human workers to focus on complex judgment, escalation, and strategy.
Crucially, human operators retain override authority: supervisors who can pull a representative off a queue, escalation paths that override default protocol, and decision logs that survive audit. Raker characterizes these controls as what makes AI "governable"—human teams set the rules models run against, monitor production performance, spot failures, and pause or adjust flows when something stops looking right. In this framing, human‑in‑the‑loop governance is not a limitation but the mechanism that allows safe scaling.
What this means for frontline teams, agencies, and citizens
Frontline teams: Operators and eligibility examiners are repositioned to handle novel, complex cases rather than routine triage and documentation, with the ability to override AI when needed.
Agencies: Building AI governance capability—clear accountability, auditable controls, and the discipline inherited from contact‑center escalation patterns—becomes part of modernization, alongside API and identity work that make conversational channels traceable.
Citizens: The shift promises omnichannel access—on a call, in chat, or through an AI assistant—held to the same standards of accuracy, auditability, and accountability, and subject to accessibility obligations and clear paths to challenge errors.
Raker sums the change as structural: conversation is moving to the center of gravity for how people access public services, and the architecture, governance and frontline practices must follow. "This is the promise of conversational service delivery, right at the government’s new front door."
