AI Leader Training · Phase 1 half-day seminar for Principals · POI draft 0.7 · IIIAC · PhantomX · ISO WFX 27-3 train-up
AI Leader Training: a half day for the Principals first, the full program if the corps wants it.
Phase 1, on COL Gunter's direction of 3 September: a four-hour seminar for the corps Principals ahead of a WFX 27-3 CPX, at $15,000 plus travel, not to exceed $17,500. The Principals see the method once, on their own decisions, and set the agenda for the corps' AI Supervisor and Automating Workflow Guardrails. Phase 2 is the full program as drafted, recommended, subject to ODT concurrence and corps approval. Marks go to Dr. Mullick only; nothing here obligates the command.
- Answer the five rows in the decisions table below. Blank means the assumption stands.
- Tap Keep or Change on any block. Change opens a line for what should change.
- Press Send once. It reaches Dr. Mullick, who issues the final quotation within a day. Glowing boxes are still open; "Next open item" walks you to each.
Word and PDF copies of each phase are at the bottom. Phase 2 keeps its 2 September marks unless you change them.
Access. Code-only, and hidden from search engines. Nothing here is classified. Sends reach Dr. Mullick alone.
Phase 1 · Decisions requested from the ODT five answers finalize the half day
Each row shows what the draft assumes. Type an answer where it differs; blank accepts.
| # | Question | Assumed in this draft | Your answer |
|---|---|---|---|
| 1 | Which CPX does the seminar precede, and on what date? | A WFX 27-3 CPX; date open; 30 days' notice | |
| 2 | How many Principals will be in the room? | Up to 15 | |
| 3 | Start time and room? | 0800; Government room with projection | |
| 4 | Wi-Fi in the room? | Yes; if no, the instructor brings connectivity at no charge | |
| 5 | Strongly recommended. Will the Principals take the 15-minute pre-course exercise the week before? It is the only way the room sees cognitive offloading and surrender in its own numbers: seven staff problems, an AI advisor on each, wrong on three, scored before anyone knows which. Without it, block 1 opens on the published baseline instead of the corps' own. | Yes, by class code from the ODT |
Phase 1 · Training event concept what the ODT staffs for approval
A unit training product, not a TRADOC-validated POI. Drafted for the ODT to staff to COL Gunter.
| Event | AI Leader Training, Principals' seminar · staff training event, four hours, in person at Fort Hood. ISO the WFX 27-3 train-up. |
| Purpose | Corps Principals see, on their own decisions, what an AI advisor does to a staff under a clock, and set the agenda for the corps' AI Supervisor and Automating Workflow Guardrails. |
| Nests under | Optional, and the ODT's call. Suggested: exercise mission command; conduct the operations process (FM 6-0; ADP 5-0), ISO WFX 27-3 CPX preparation. Whatever you type here travels with your marks. |
| Audience | Corps Principals: primary and special staff who supervise AI-using sections. Up to 15, seated by warfighting function. |
| Training objectives | 1. Task: keep the check on a machine answer. Conditions: an AI advisor at every desk, wrong on some products; the order open; no indication which. Standard: the order opened before acceptance on the measured decision, each check on the record. 2. Task: state the corps' first best practices for supervising AI-using staff. Conditions: guided discussion after the AAR, the room's own record on screen. Standard: a written agenda in the Principals' words, delivered with the after-action note. |
| Resources | Government: one room for up to 15 with projection and Wi-Fi; roster; installation access for one instructor. Contractor: facilitator, scenario, simulation, pre-course exercise, after-action note. Participants use their own devices; nothing is installed. |
| Cost and funding | $15,000 firm fixed plus travel at cost, not to exceed $2,500. Total obligation not to exceed $17,500. Tailored training, by purchase order. PhantomX funds. Quotation RTA-IIIAC-2026-0903. |
| Risk | Low. Classroom event, no ranges, no vehicles, no live fire. Unclassified throughout; scenario from the public record; no CUI on any contractor system. Principal risk to the training is a schedule conflict with the CPX sequence, resolved once the ODT sets the date. |
| Output | The Principals' guardrails agenda as written in the room; cohort figures with the count shown; an after-action note in editable form within five business days. DTMS entry on request so the corps records the event. |
| Recommendation | Recommend approval to execute on the date at row 1 of the decisions table, resourced by PhantomX, and inclusion on the corps training calendar. |
Phase 1 · Half-day seminar for the Principals program of instruction · mirrors quotation RTA-IIIAC-2026-0903
| Event title | AI Leader Training, Principals' seminar (working title; ODT sets the final). Phase 1, ISO the WFX 27-3 train-up. A unit training product built on the TR 350-70 form; not a TRADOC-validated POI. |
| Requiring activity | IIIAC, PhantomX. Event owner: Operational Data Team, LTC Adam Dials. Sponsor: LTC William Sitze, PhantomX Deputy. Direction: COL Gunter, 3 Sep 2026. |
| Purpose | The Principals see, on their own decisions, what an AI advisor does to a staff under a clock, and set the agenda for the corps' AI Supervisor and Automating Workflow Guardrails. |
| Audience | The corps Principals: primary and special staff who supervise AI-using staff sections. Up to 15. Seated by warfighting function. |
| Length and place | One four-hour block: 3.5 hours of instruction and practical exercise, two breaks, 10 minutes of administration. In person at Fort Hood, one Government room. Date set with the ODT; 30 days' notice is sufficient. |
| Prerequisite | The 15-minute pre-course exercise, taken the week before by class code from the ODT. Read-ahead under 15 minutes. |
| Delivery | One facilitator, Dr. Mullick. Adversary moves pre-scripted; no Red Cell travels. Participants on their own devices in a browser. Unclassified; no CUI on any contractor system. |
| Price | $15,000 firm fixed, published half-day rate for up to 15; travel at cost under the FTR, not to exceed $2,500. Tailored training, by purchase order. Quotation RTA-IIIAC-2026-0903, 3 Sep 2026. |
Phase 1 · Learning objectives action, condition, standard (TR 350-70 form) · enabling objectives optional
TLO 1 · the standard
Keep the check on a machine answer, on the record.
TLO 2 · the standard
State the corps' first best practices for supervising AI-using staff.
TLO 3, as time allows. Place AI in the operations process, Army Design Methodology through the MDMP.
Phase 1 · Run of the event 0800 start assumed · mark any block
Questions are taken inside each block. If time is lost, block 1 shortens first, then a vignette in block 2 drops; the simulation, the AAR and the guardrails block hold.
| Time | Block | Min | Method | TLO |
|---|---|---|---|---|
| 0800 | Administration and purpose. | 10 | Conference | — |
| 0810 | 1 · Design before MDMP; the Principals' own record. Where an AI advisor helps and where it has no part (ADP 5-0; FM 5-0). Results of the pre-course exercise. | 30 | Conference | 3, 1 |
| 0840 | Break | 10 | ||
| 0850 | 2 · Decision-making models and checking a machine answer. Where staff decisions go wrong under time pressure: anchoring, confirmation bias, groupthink, the fast-versus-slow trade. Then the check in Army terms: treat the AI answer as an assumption until validated (FM 5-0 ch. 5); rate the source A to F and the information 1 to 6 (ATP 2-33.4); test it against the order. Two short vignettes. | 35 | Conference · PE | 1, 3 |
| 0925 | 3 · The simulation, one pulse. Desks by warfighting function, an AI advisor at each, wrong on purpose on some products: data poisoning in the advisor's inputs. A target inside a no-fire area, recommendation already drafted. Each decision on the record: sign, return, or rework, and whether the order was opened first. | 65 | Practical exercise | 1 |
| 1030 | Break | 10 | ||
| 1040 | 4 · AAR on the record. What was signed, what was checked, what the advisor got wrong and who caught it. | 45 | AAR | 1 |
| 1125 | 5 · Guardrails agenda and way ahead. The Principals set what the corps memorandum must cover. Shape of Phase 2, if the corps wants it. | 35 | Guided discussion | 2 |
| 1200 | End. 3.5 hours of instruction and practical exercise, two breaks, 10 minutes of administration. | 240 |
Phase 1 · What the ODT receives, and what it furnishes
| Item | Format | Due | Rights |
|---|---|---|---|
| Read-ahead and class codes for the pre-course exercise | PDF; email to the ODT | 7 days before | Licensed |
| The seminar | In person, Fort Hood | Agreed date | — |
| After-action note: cohort figures with the count shown; the Principals' guardrails agenda as written in the room; options for Phase 2 | Editable Word file, unmarked | 5 business days after | Unlimited rights |
Government-furnished. One room for up to 15 with projection and Wi-Fi, confirmed 21 days out; roster; installation access for one instructor; class codes distributed. Participants use their own devices.
Done with the half day? Send from here. Nothing in Phase 2 is needed to issue the final quotation.
What the ODT and PhantomX have already told us 2 and 3 Sep · carried into both phases
3 Sep, LTC Sitze on COL Gunter's direction. "He wants to plan and get a half-day session on the books ahead of/leading into one of the CPXs for Principals." "He feels this is the pressing need but is interested in assessing feasibility of some follow-on sessions to round us out a bit more." Phase 1 below is that half day; Phase 2 is the follow-on for his approval.
Guardrails "to be adopted by the Corps." Now TLO 4, lesson 15, and a deliverable.
Design alongside MDMP. Lesson 1.
How-We-Fight products (CUI) and a new doctrine note. Briefed orally to the facilitator in the room; the scenario stays unclassified.
"AI requires iterative interaction not task relegation." TLO 3, in those words.
Schedule (LTC Dials, by phone). Three consecutive days of O-4s and above is hard to give during a train-up; spread it over a few weeks in three-hour seminars. Schedule A → shows the same 19.5 hours as eight three-hour seminars, two a week for four weeks; the three-day block is Schedule B.
Phase 2
The full program, if the corps wants it.
Recommended follow-on to the half day · subject to ODT concurrence and corps approval · 19.5 academic hours · eight three-hour seminars over four weeks, or three consecutive days
One question for the ODT
Should Phase 2 be planned now, alongside the half day, or wait until the Principals have seen the method?
Sections 1 to 9 below each carry a Keep or Change box.
The full program of instruction as drafted on 2 September, kept for the ODT's planning. Priced from the published card, with the one-time first-engagement reduction reserved for it. Sections 1 to 9 keep their 2 September marks.
1. Course administrative data
| Course title | AI Leader Training (working title; ODT sets the final). ISO the WFX 27-3 train-up. |
| Requiring activity | IIIAC, PhantomX. POI owner: Operational Data Team, LTC Adam Dials, Command Chief Data & Analytics Officer. Sponsor: LTC William Sitze, PhantomX Deputy. |
| Provider | Red Teaming Associates LLC · SDVOSB · CAGE 6B7Q4 · UEI H54BTD3BJXJ7 · NAICS 611430 · PSC U008. Instructor: Dr. Haider Mullick. |
| Purpose | Leaders keep the check on a machine answer and catch a wrong one. The corps sets its rules for when to direct a tool, automate a workflow, or coordinate a decision across warfighting functions. |
| Audience | Leaders who supervise AI-using staff: primary and special staff, O-5 to O-6, senior warrants and NCOs, seated by warfighting function (FM 6-0; FM 3-0). Class size up to 25; the desk plan scales to the number in the room. ODT to confirm who attends and which sections. |
| Guardrails deliverable | AI Supervisor and Automating Workflow Guardrails: a draft corps policy memorandum in AR 25-50 form, written by the participating leaders in the final session from the exercise record, delivered editable with the after-action package for staffing to the Commanding General. This is a large ask. Guardrails are best practices for a staff's use of AI, and no published Army or DoD guidance yet says what a staff officer must check before signing machine-drafted product. The corps would be writing a first. The course supplies the frame, the doctrine references and the record of what its own leaders did; the rules are the leaders' to write and the command's to adopt. |
| Prerequisite | The 15-minute pre-course experiment (class code from the ODT). Read-ahead under 30 minutes. |
| Length | 19.5 academic hours. Schedule A: eight sessions of three hours or less, two a week for four weeks (LTC Dials, 2 Sep). Schedule B: 3 consecutive days, 0830 to 1630 (conference 3.5 · practical exercise 12.5 · discussion 1.5 · AAR 2). One hour for lunch and two 15-minute breaks each day, outside the count. |
| Delivery | Resident, in person. One facilitator, a Red Cell of 2 to 3, participants on their own devices. No online version. |
| Security | Unclassified throughout. Scenario from the public record. No CUI on any contractor system. |
| Location | Fort Hood, to confirm. Two Government rooms with projection and Wi-Fi. |
| Scheduling | Dates fixed 60 days before each delivery as the standard, with flexibility for the corps' calendar; scenario review 14 days before. |
| Version | Draft 0.7, 3 Sep 2026: Phase 1 half-day seminar for Principals added on COL Gunter's direction (LTC Sitze, 3 Sep), with a training event concept, a five-answer decision table and a mark on every block; page put behind an access code; Word and PDF copies of each phase; full POI kept as the recommended Phase 2, priced from the same card. Draft 0.6, 2 Sep 2026: acquisition wording tightened (tailored-training route, labor-standards determination, rights in deliverables); four-week schedule added on LTC Dials’s call; instruction content detailed; revised on LTC Sitze’s mark-up. |
2. Terminal learning objectives action · condition · standard
TLO 1
Locate the decision failure at the step where it occurs.
- Army Design Methodology before MDMP: frame the operational environment, frame the problem, develop the operational approach, reframe (ADP 5-0 para 2-1 series; ATP 5-0.1 ch. 1 to 5; FM 5-0 ch. 3). Design entered Army doctrine in FM 5-0, March 2010, ch. 3.
- Eight models, each at the MDMP step where it fails.
TLO 2
Keep the check on a machine answer.
- Offloading (keep the check) against surrender (drop it).
- Treat a machine answer as an assumption until validated (FM 5-0 ch. 5). Four checks: source rated A to F and information 1 to 6 (ATP 2-33.4, table 2-1); the order; independent confirmation; time available.
- Spot an adversary attacking the data: poisoned intake, broken baseline, spoofed provenance, stale data at full confidence.
TLO 3
Work the advisor iteratively.
TLO 4
AI Supervisor and Automating Workflow Guardrails: the corps' best practices for AI use.
- What a staff officer checks before signing machine product, against the DoD AI principles the corps is bound by: responsible, equitable, traceable, reliable, governable (DoD 2020; CDAO RAI Toolkit 2023).
- Sort the exercise's decisions: direct, automate, coordinate.
3. Lesson sequence proposed · 19.5 academic hours · two schedules for the same content
Schedule A, four weeks: eight sessions of three hours or less, two a week. Suggested by LTC Dials on 2 Sep so that no leader gives up a full day. The instructor is resident at Fort Hood for the four weeks; the Red Cell files its order and the facilitator adjudicates between the two simulation sessions, so the clock and the adversary are held across the week. Once a week stretches the same eight sessions to eight weeks; twice a week holds better.
| Wk | Session | Hours | Method | TLO |
|---|---|---|---|---|
| Week 1 · Instruction · 5.0 | ||||
| 1 | S1 · Design before MDMP. Frame the environment, frame the problem, develop the approach, reframe; the Eightfold Path beside it; the AI point. | 2.5 | Conference | 1 |
| 1 | S2 · Eight decision models at the MDMP step where each fails; each leader’s own record from the pre-course experiment. | 2.5 | Conference · PE | 1, 2 |
| Week 2 · Instruction · 4.5 | ||||
| 2 | S3 · The assumption check. Source, order, confirmation, time. Four data-corruption vignettes, each AAR’d. | 2.5 | Practical exercise | 2, 3 |
| 2 | S4 · Open discussion. What the staff does with AI now, where it has gone wrong, what the guardrails must cover. | 2.0 | Discussion | 3, 4 |
| Week 3 · Simulation · 6.0 | ||||
| 3 | S5 · Simulation A. Receipt of mission; Pulse 1, mission analysis to planning guidance; Pulse 2, one COA wargamed against the Red Cell’s order. State saved; Red files between sessions. | 3.0 | Practical exercise | 1, 2, 3 |
| 3 | S6 · Simulation B. Pulse 5, the constraint test; Pulse 6, the reframe test; ENDEX. | 3.0 | Practical exercise | 1, 2, 3 |
| Week 4 · Review and guardrails · 4.0 | ||||
| 4 | S7 · AAR on the record. Individual records back; cohort pattern on screen; Red Cell mirror memo. | 2.0 | AAR | 1, 2, 3 |
| 4 | S8 · Guardrails session. Leaders draft the IIIAC memorandum. Next steps. | 2.0 | Guided PE | 4 |
| Total. Against the three-day block: Pulse 3 (FRAGORD skeleton) and Pulse 4 (execution opens) are cut; four measured decisions per desk instead of six; the models and the own-record block share a session; the guardrails session grows to two hours. | 19.5 | Conf 5 · PE 10.5 · Disc 2 · AAR 2 | ||
Schedule B, three consecutive days, 0830 to 1630. The original shape, kept as the alternate for a staff that can clear the days.
| No. | Lesson | Hours | Method | TLO |
|---|---|---|---|---|
| Day 1 · Instruction · 6.5 academic hours · 0830–1630 | ||||
| 1 | Design before MDMP. Army Design Methodology: frame the operational environment, frame the problem, develop the operational approach, and reframe when execution shows the problem has moved (ADP 5-0; ATP 5-0.1; FM 5-0 ch. 3). Design is conceptual planning and is a command decision based on time and complexity; MDMP is the detailed planning that turns the approach into an order (FM 5-0 ch. 5; FM 6-0). The AI point: an advisor is fast at detailed planning and has no part in framing. A tool tasked before the problem is framed solves the wrong problem well. | 1.5 | Conference | 1 |
| 2 | Eight decision models, each at the MDMP step where it fails. | 1.5 | Conference | 1 |
| 3 | Your own record. The pre-course experiment, offloading against surrender. | 1.0 | Practical exercise | 2 |
| 4 | The assumption check. Source, order, confirmation, time. Four data-corruption vignettes, each AAR’d. | 1.5 | Practical exercise | 2, 3 |
| 5 | Open discussion. What the staff does with AI now, where it has gone wrong, what the guardrails must cover. | 1.0 | Discussion | 3, 4 |
| Day 2 · Simulation · planning · 6.5 · 0830–1630 | ||||
| 6 | Receipt of mission. Order on screen, advisor open. | 0.5 | Conference | 1 |
| 7 | Pulse 1. Mission analysis to planning guidance. | 2.0 | Practical exercise | 1, 2, 3 |
| 8 | Pulse 2. One COA, wargamed against the adversary’s order. | 2.0 | Practical exercise | 1, 2, 3 |
| 9 | Pulse 3. Decision and FRAGORD skeleton. | 1.5 | Practical exercise | 1, 3 |
| 10 | Day 2 review. Questions on the record so far. | 0.5 | Discussion | 2, 3 |
| Day 3 · Simulation · execution and review · 6.5 · 0830–1630 | ||||
| 11 | Pulse 4. H-hour. Update brief, targeting working group. | 1.0 | Practical exercise | 2, 3 |
| 12 | Pulse 5. The constraint test. A target inside a no-fire area, recommendation already written. | 1.0 | Practical exercise | 2 |
| 13 | Pulse 6. The reframe test. Reserve decision on a degraded network. The record shows whether a desk re-tasked the advisor against the old frame or named that the problem had changed first (ATP 5-0.1 ch. 5, reframing). ENDEX. | 1.0 | Practical exercise | 1, 2, 3 |
| 14 | AAR on the record. Individual records back; cohort pattern on screen; Red Cell mirror memo. | 2.0 | AAR | 1, 2, 3 |
| 15 | Guardrails session. Leaders draft the IIIAC rules. Next steps. | 1.5 | Guided PE | 4 |
| Total. Each day: one hour for lunch, two 15-minute breaks, and slack for questions inside every block. | 19.5 | Conf 3.5 · PE 12.5 · Disc 1.5 · AAR 2 | ||
3a. What is taught
| Block | Content | Where it lands in the operations process |
|---|---|---|
| Design before MDMP | Frame the environment, frame the problem, develop the approach, reframe. A tool tasked before the problem is framed solves the wrong problem well. | ADM; receipt of mission; mission analysis |
| Eight decision models | Rational actor, organizational process, bureaucratic politics, bounded rationality, dual-process, prospect theory, heuristics and bias, groupthink. Each placed at the MDMP step where it is most often seen. | All seven MDMP steps; the decision brief |
| Biases and debiasing | Fast and slow thinking under a clock. Anchoring, confirmation, availability, sunk cost, speed asymmetry. Debiasing a staff can run: red team and devil’s advocate, premortem, probabilistic language in estimates, decision quality separated from outcome. | Facts and assumptions; the wargame; the decision brief |
| The AI advisor | Offloading (keep the check) against surrender (drop the check). Each leader’s own record from the pre-course experiment against the published baseline. The assumption check in Army terms: source A to F, information 1 to 6, the order, confirmation, time. | Any step where a machine drafts a product; the pre-signature check |
| The adversary in the data | Poisoned intake, a taught baseline broken on purpose, spoofed provenance, stale data at full confidence. Four vignettes, each AAR’d. | IPB; running estimates; the targeting working group |
| The staff as an organization | What a staff produces because of how it is organized, what the sections fight over, what the commander sees. Partner-nation caveats. How the order, the screen and the advisor’s defaults shape what a desk decides. | Task organization; the coalition seat; the FRAGORD |
| Briefing the decision | The central question, the options with trade-offs, the recommendation that anticipates its critics, under a clock. | Mission analysis brief; COA decision brief |
| Guardrails | Direct, automate, coordinate. What a staff officer checks before signing machine product. The corps writes its own best practices from the exercise record, where doctrine is silent. | Commander’s guidance; the unit SOP |
4. The simulation Breakpoint: the Kesland crisis · working sample
A multinational corps under a JTF, a partner division OPCON, a degraded mission partner network, a live Red Cell, an advisor on every desk. It runs today. What changes for IIIAC:
| IIIAC said | Proposed | Status |
|---|---|---|
| Leaders supervising people | Leaders play leaders. Their own action officers work the advisor; where a leader comes alone, the advisor drafts as an action officer would. The measured decision: sign, return, or rework, and whether the order was opened first. | Proposed; depends on who is in the room |
| Guardrails for the Corps | Each filed decision tagged direct, automate or coordinate, so lesson 15 starts from the staff's own calls and the table in the draft memorandum is pre-filled from the decision record. | Proposed; ODT decides |
| Design | Pulse 6 is the reframe test. | In the sample |
| Vignettes; the JRTC use cases ("Yes - love the use cases") | The four data-corruption cases from the JRTC rotation account stay: day-1 vignettes, then the same four as injects on days 2 and 3. | In the sample |
| CUI products | Briefed orally to the facilitator. Injects tuned where the unclassified scenario allows. | Depends on the ODT |
| Tool-agnostic | Advisor runs on a commercial model in our environment. No corps tool touched or named. | In the sample |
Not promised. This is not the WFX 27-3 scenario. Corps-specific injects are written after detailed planning and reviewed with the ODT 14 days before delivery.
5. Evaluation
Why measure this · Wharton, 2026 · n = 1,372
Shaw & Nave, Wharton, Jan 2026, Study 1. A wrong advisor was followed 73.2% of the time. Baseline is ordinary adults, not staff officers; the corps' own pre-course number replaces it.
Baseline. Pre-course experiment: seven staff problems, advisor wrong on three. Cohort view to the ODT by class code.
During. The exercise record per leader and desk: each consultation, check, decision. Returned in lesson 14.
After. Cohort findings and an individual assessment per leader against the four TLOs, within five days of ENDEX. Short reaction survey at the close.
If wanted. Re-run the pre-course instrument before each CPX for a trend. A class of 16 to 25 is reported with the count beside every figure.
6. Resources
Room 1 · Blue · corps main
Six desks by warfighting function, up to 25 leaders
Room 2 · Red and White
Adversary and control
| Government-furnished | Contractor-furnished |
|---|---|
| Two rooms for three days, a projector in each, Wi-Fi. Wi-Fi confirmed yes or no at the 60-day point: if the rooms cannot provide it, the contractor brings its own connectivity (cellular hotspots or a satellite terminal), which needs that notice and installation approval to operate. Roster and desk assignments. Installation access for one facilitator. Class codes distributed. Oral brief of CUI products to the facilitator. | Facilitator and Red Cell. Scenario, injects, adversary orders, advisor, all materials and instruments. After-action package: cohort findings, individual assessments, failure modes in the corps' own products, editable debrief slides, the draft guardrails. Scenario review with the ODT 14 days out. Nothing installed on a Government network. Scenario content and instructional materials remain the contractor’s commercial technical data, licensed for Government use; the after-action package, the individual assessments and the participant-drafted guardrails are delivered without restriction. |
Either switch adds a not-to-exceed ceiling to the figure in section 7 and to the quotation. Billed at actual cost with receipts; the unused part of a ceiling is never invoiced. Decide at the 60-day point.
7. Fit to the WFX 27-3 train-up and what each shape costs
The last three Warfighters ran in spring 2022, 2023 and 2025, each after CPX I in the fall, CPX II in February, CPX III in March. The ODT confirms the 27-3 dates. The half day is quoted separately (RTA-IIIAC-2026-0903). Two shapes remain for Phase 2; the decision sits with the corps.
Published, first engagement
$72,900
Published rate $81,000. The 10% is a one-time commercial discount on one delivery; the card price is the ordinary selling price.
Total obligation with at-cost ceilings
$81,000
Program price plus travel NTE.
Half, two and three days are published rates for 16 to 25. Other lengths are indicative.
The builder prices one delivery length at a time. It opens on the first delivery; the others are quoted from the published card as separate lines.
8. How it is bought
What it is
Education and training services
NAICS 611430 · PSC U008. Tailored training: contracting office action at any dollar value, not a purchase-card route. No software, no license, no government data, no ATO.
Price
Firm fixed, per delivery
Published card through 30 Sep 2027. The calculator in section 7 says whether the aggregate, with travel, venue and connectivity ceilings, sits under the $350,000 simplified acquisition threshold.
Route
SDVOSB sole source, or a purchase order
FY26 funds by 30 Sep; delivery in FY27. Applicability of Service Contract Labor Standards is the contracting officer’s determination; the price is firm fixed either way.
Paperwork
Ready
Requirements record, PWS, QASP, quotation, sole-source facts. One Word file from the builder.
9. Open items what we need from the ODT
| Question | Decides |
|---|---|
| WFX 27-3 dates; CPX I, II, III windows. | Delivery window and shape. |
| How many leaders, and does each bring an action officer? Which sections? | Desk plan, class band, and whether the advisor plays the action officer. |
| What the staff uses AI for now, in general terms, and where it has gone wrong. | Day-1 vignettes and the guardrails agenda. |
| Who briefs the How-We-Fight products, and when. | Inject tuning before the 14-day review. |
| Contracting officer and fund cite. | Who receives the paperwork. |
| Rooms, projection, installation access. Wi-Fi: yes or no. | No venue line; whether the contractor brings connectivity. |
| Target dates for each delivery, fixed 60 days out. | Scenario build, travel, and connectivity planning. |
| Run the pre-course experiment before the next meeting? | The corps' own baseline before design is final. |
| What data the ODT wants from the exercise, and who holds it. | Whether an anonymized export joins the after-action package. |
| Would the corps want later sessions priced now as options the KO can exercise, or bought one at a time? | Shape of the quotation: base plus option lines, or a single delivery. Every option counts toward the $350,000 threshold. |
Send your marks and answers nothing obligates the command
Answers, marks and the line below go in one Send. Unmarked means no objection. Nothing leaves this browser until you press Send.
If the network blocks Send, the email button opens a message with your answers already typed; Copy puts the same text on the clipboard.
Word files carry your answers and marks as typed. PDF prints that phase alone.
References
- TRADOC Regulation 350-70, Army Learning Policy and Systems, 10 July 2017, ch. 3 (lesson plans; learning objectives as action, condition, standard; terminal and enabling objectives) and ch. 4 (ADDIE). TRADOC Pamphlet 350-70-14, Training and Educational Development in Support of the Institutional Domain, 2021. TRADOC Pamphlet 350-70-7, Army Educational Processes, 4 October 2018. AR 350-1, Army Training and Leader Development, 10 December 2017. AR 25-50, Preparing and Managing Correspondence, 10 October 2020 (form of the guardrails memorandum).
- ADP 5-0, The Operations Process, 31 July 2019 (with change 1): the operations process; ADM, MDMP and troop leading procedures as the three planning methodologies; commander's understand, visualize, describe, direct, lead, assess.
- ATP 5-0.1, Army Design Methodology, 1 July 2015: framing the operational environment, framing the problem, developing the operational approach, reframing; the relationship of ADM to the MDMP (ch. 2).
- FM 5-0, Planning and Orders Production, 16 May 2022: ch. 3 Army Design Methodology; ch. 5 the Military Decision-Making Process (mission analysis, critical facts and assumptions, constraints, use of available time).
- FM 6-0, Commander and Staff Organization and Operations, 16 May 2022: staff running estimates and the decision on whether to initiate ADM before the MDMP.
- ADP 6-0, Mission Command, 31 July 2019: commander's intent, disciplined initiative, the exercise of authority through the staff.
- FM 3-0, Operations, 1 October 2022: large-scale combat operations; corps as a tactical headquarters; multidomain operations.
- FM 5-0, The Operations Process, 26 March 2010, ch. 3, Design: the edition that brought Design into Army doctrine.
- ATP 5-0.2-1, Staff Reference Guide, Volume I, 7 December 2020: staff products and the integrating processes (IPB, information collection, targeting).
- ATP 2-01.3, Intelligence Preparation of the Battlefield, 1 March 2019 (with change 1, 2021): IPB inside mission analysis.
- ADP 7-0, Training, 31 July 2019, and FM 7-0, Training, 14 June 2021: unit training management; the train-up cycle a Warfighter rests on.
- DoDD 3000.09, Autonomy in Weapon Systems, 25 January 2023: cited to state what this course does not govern.
- ATP 2-33.4, Intelligence Analysis, 10 January 2020 (change 1, May 2025), table 2-1: evaluation ratings for source reliability A to F and information accuracy 1 to 6. Source of the pre-signature check.
- DoD AI Ethical Principles, 24 Feb 2020. DoD Responsible AI Strategy and Implementation Pathway, 2022. CDAO Responsible AI Toolkit, Nov 2023.
- Shaw & Nave, Thinking—Fast, Slow, and Artificial, Wharton, Jan 2026.
- Bardach & Patashnik, A Practical Guide for Policy Analysis; Allison & Zelikow, Essence of Decision; Kahneman, Thinking, Fast and Slow; Heuer, Psychology of Intelligence Analysis; Janis, Groupthink. Sources for the decision models and debiasing blocks.
- Lajeunesse, Palazini & Tulskie, "Artificial Staff, Human Command: An AI Integration Experiment," Modern War Institute at West Point, 2026. The JRTC rotation account behind the vignettes.
- IIIAC public record: WFX 22-4, 23-4, 25-4; CPX II Feb 2025; PhantomX, 16 Apr 2026. Full list on the program page.
Red Teaming Associates LLC · SDVOSB, SBA VetCert certified · CAGE 6B7Q4 · UEI H54BTD3BJXJ7 · NAICS 611430 · PSC U008
contactus@redteamingassociates.com · 202-510-0725 · Unclassified · unlisted, not indexed, access by code · draft for planning · build 2026-09-03-2255