Before ARES arrives.
The Air Force has chosen its next system of record for aircrew scheduling, training management, and standardization and evaluation. The migration starts next year. This note is about the twelve months a flying wing controls before then, and about the difference between a system of record and the decision layer that sits above it.
One record for an Airman's whole flying career.
In July 2026 the Air Force Life Cycle Management Center, working through Kessel Run for Air Education and Training Command, awarded a $100 million other transaction agreement for the Aerospace Readiness Enterprise System, ARES. The program consolidates six platforms used today for scheduling, training management, and stan/eval into one system, for roughly 149,000 Airmen across the major commands. The public plan is a localized pilot next year, rollout across major commands in 2027, and full capability in AETC by 2028. Air & Space Forces Magazine reported that the system will eventually include an AI-assisted mission engine to cross-reference qualifications, currency requirements, and rest periods and flag conflicts as they arise.
That is a good decision. Anyone who has reconciled a flying schedule against qualification records and crew rest by hand knows why a single record matters. It is also a decision with a clock attached, and the clock is what this note is about.
A system of record answers "what is true." A decision layer answers "what breaks next."
The two are judged by different standards, and they are built by different people. A system of record is measured by completeness, authority, and audit: is every qualification, currency window, sortie, and rest period captured once, correctly, with a clear owner. A decision layer is measured by whether the commander's morning question gets an answer with evidence attached: which class falls behind this week, which resource binds first, what the least disruptive recovery is, and who approves it.
Optimize is a decision layer. Today it sits above a training organization's existing systems, reconciles them into one readiness picture, and stages recovery options for a human to approve. When a single system of record exists, the reconciliation work shrinks and the decision work does not. The forecast, the simulation, the ranked course of action, the evidence trail, and the approval gate are still the layer above the record, whoever holds the record. A wing that builds its decision layer now inherits a working one when the record changes underneath it. A wing that waits gets a new record and a blank page.
Seven things that are yours regardless of vendor.
None of this requires the new system, a contract with us, or permission from a program office. All of it makes the migration faster and the decision layer better, and every step is a Cleared FDE Standard practice you can run yourself.
One sentence: who decides, from which systems, how often, with what consequence. If the sentence has three owners, that is the first finding.
For each platform being consolidated: the owner, the refresh cadence, which fields the decision actually uses, and which fields nobody trusts.
Qualification codes, currency windows, sortie and event types, cancellation reason codes. The record will inherit whatever is ambiguous today.
Hours spent reconciling, the latency between a slip and the decision, the number of recovery plans that were never written down. This is the number the migration will be judged against.
Who asked, who approved, what changed, on every schedule intervention. An evidence ledger is a discipline before it is a schema; the public schema at federal.ai/evidence-ledger is free to use.
What the software may draft, what it may monitor, what only a human decides, and what it may execute after approval. Write it down before any vendor proposes it for you.
One decision, one squadron, ninety days, kill criteria in writing. The migration then inherits a decision layer with a baseline, an evidence trail, and trained operators.
The migration is a scope, not a restart.
Optimize is SBIR-derived technology and has been delivered under SBIR Phase III authority since 2022. Work that extends it, including integration with a new system of record, can be awarded under that authority: sole source, sized to the need, under 15 U.S.C. 638(r). The acquisition detail is on how to buy. The engineering detail is a thirty-minute conversation.
Everything above applies whether or not that conversation happens. The clock is the same for everyone.
Bring the decision, not the program.
Thirty minutes with a technical lead who has built a decision layer above training systems that were not designed to talk to each other.