The Social Security Forum

Dispatches from Camp: SSA’s AI Buildout—Reading the Tea Leaves in Four New Procurement Signals

September 25, 2026

David Camp, NOSSCR Senior Policy Advisor

A procurement notice posted by SSA this summer, read alongside the agency’s broader pattern of activity, is among the clearest signals yet that AI-assisted case processing (including medical evidence review and claim triage) is coming to the disability system. Here’s what each one says, and what it means for how your clients’ files may soon be handled.

If SSA’s pattern holds, that means AI touching medical evidence review, screening for consistency or fraud indicators, and eventually the triage decisions that determine how a claim moves through the system. Here are the four notices that point that way.

The Four Signals

First, SSA issued a Request for Information on an Enterprise AI Strategy: the framework that will decide how AI tools eventually reach adjudicators’ desks and claimants’ files, rather than staying siloed in pilot programs (Solicitation No. 28321326RI0000041, posted August 14, 2026 by SSA’s Office of Acquisition and Grants, with responses due September 28). The notice has drawn attention from trade press including FedScoop, ExecutiveGov, and HSToday, and covers governance, data readiness, platform layers, and training for agentic AI tools.

Second, and issued six days later, SSA put out an RFI labeled simply “Request for Information–Ontology,” a term that in SSA’s context likely points toward building a structured data model, the kind of system that could let software reason over unstructured case content rather than merely search it as text (Solicitation No. 28321326RI0000043, posted August 20, 2026, responses due September 10). The public notice itself is a single line, so this reading is my own inference, not something the notice states outright: teaching a computer system how concepts relate to one another so it can reason over a disability file full of medical records, function reports, and vocational evidence, rather than just search it as text. If that inference holds, it would be a necessary step before SSA could meaningfully automate review of medical evidence or flag files for particular handling.

Third, SSA published an RFI on mobile application shielding, the security layer that has to harden before SSA can expand any AI-enabled tool onto mobile-facing platforms claimants and representatives might use (Solicitation No. 28321326RI0000042, posted August 20, 2026, with responses due September 3). This one is less about adjudication directly and more about the infrastructure hardening that has to accompany a larger digital footprint.

Fourth, SSA awarded a contract for a Coder Enterprise Platform, procured through reseller Flywheel Data LLC (PIID 28321326FDX030137, signed August 25, 2026, $117,300, covering premium support for 100 users, with a period of performance running September 1, 2026 through August 31, 2027). This is a modest purchase in dollar terms, and I don’t read much into the size of the award. The signal here is the tool choice itself: a platform for the agency’s own developers and contractors to build, test, and deploy code faster and more uniformly.

What the Pattern Suggests

None of these four items is a rule, and none of them is final. But taken together, an enterprise AI strategy, a possible data model for reasoning over case content, a security posture built for a larger mobile and application surface, and a development platform to build on, describe an agency methodically laying the foundation for AI-assisted case processing at scale.

This is not a new interest for SSA, and it is not one NOSSCR is encountering for the first time. In our May 2025 letter to Commissioner Bisignano on the Disability Determination Ready Claim Process, we proposed that SSA consider a pilot using AI to flag certain case types for reconsideration, precisely because targeted, transparent use of these tools can reduce burden if it is designed with due process in mind. I have also written publicly, including on LinkedIn, about SSA’s IMAGEN tool for analyzing medical evidence as one option worth watching on the claims-processing efficiency front. At our Spring 2026 National Conference in Baltimore, Commissioner Bisignano reported progress on AI applications to accelerate disability casefile review, and we built out conference programming specifically on AI in legal work because the direction of travel was already clear.

What I Am Watching For

The open question is not whether SSA will use AI more heavily in the years ahead. It will. The open question is whether that buildout comes with the transparency, testing, and error-correction pathways that due process requires. A data model that reasons over medical evidence is only as good as the medical and vocational judgment it encodes, and a screening tool that flags cases for extra scrutiny needs an equally rigorous process for flagging cases that deserve less. We have pushed SSA before on making its data-sharing and evidence-gathering tools work for claimants rather than merely for throughput, and we will bring the same lens to this enterprise AI strategy as it develops.

For now, there is nothing in these four notices that changes how you handle a pending claim. But representatives are often the first to notice when an agency’s internal process shifts, whether it is a change in how quickly file reviews come back or a new kind of request buried in a denial notice. If you start seeing anything in your practice that looks like automated triage, unusual patterns in reconsideration timing, or evidence requests that read like they were generated rather than drafted, I encourage you to post it on the NOSSCR Martin Listserv so we can see whether it is isolated or part of a broader rollout. We will keep tracking these procurement notices as they mature into actual system changes, and we will report back when there is something more concrete to act on.


For daily updates on the latest Social Security happenings, follow me: