Military Crisis Line: Dial 988, then press 1. Free, confidential, 24/7.
Proposed readiness tool. Concept in development.

Know your
baseline.

DatumForce.ai is a proposed readiness tool for the phone you already carry. It would establish your baseline during the 120 days before deployment and show changes during the 120 days after you return. Raw data and individual modeling would stay on your phone.

For Soldiers For Leadership
No wearables No routine surveys No location data Your data stays yours
09:41Illustrative
Your scoreOnly you
78of 100
vs. your baseline
+4
7-day trend
Steady
Sleep est.
7.1h
Steps
9.4k
Platoon aggregate
74one level up
09:41Illustrative
Your trend120 days
Baseline vs. now
Day 1Day 120
Sleep est.
Up
Activity
Steady
Insight

Your sleep estimate has been above your baseline for 9 days.

09:41Illustrative
ProgressSeparate from score
Days logged streak
42days
Level 61,240 / 2,000 XP
Badge
First 7
Badge
Crew up
Badge
Locked
09:41Illustrative
Your crewPrivate
Buddies connected
3 of 5
What they can get

A general check-in request. Never your score. Never details.

Need to talk now?
Dial 988, press 1
Scroll or tap to flip screens
For Soldiers

No wearables. No routine surveys. No individual data for command.

The proposed tool would use your own phone, with no wearable or routine self-report survey. Your raw inputs and individual model would stay on the device.

01

Nothing to charge

No watch. No ring. No extra gear to lose. DatumForce.ai is designed to run on the phone in your pocket using the sensors it already has.

02

Nothing to fill out

No daily check-in forms. DatumForce.ai would read passive signals only: steps, sleep estimates, screen time, and Bluetooth proximity patterns. That is the whole list.

03

Your raw data stays local

The design keeps raw inputs and individual modeling on your phone. Protected aggregate contributions and authorized buddy requests may leave it. Command would receive aggregates only, never your individual score. Geographic location is excluded.

How it works

Three steps. Zero homework.

DatumForce.ai is proposed for two windows: 120 days before deployment and 120 days after you return. Everything is measured against your own normal, not anyone else's.

1

Set your baseline

Under the proposed approach, setup would establish permissions before sensing begins. Across the 120 days before deployment, DatumForce.ai would establish your personal baseline from passive phone inputs.

2

Watch your trend

See your own score and trend. Only you can see it. For context, you also get a one-level-up unit aggregate, like your platoon's number if you are in a squad. Never peer scores.

3

Pick up where you left off

For 120 days after you return, DatumForce.ai would compare readiness-related trends with your own pre-deployment baseline.

Observation windows DatumForce.ai is not designed to run during deployment

The full 120-day pre and post windows span at least 240 observation days plus the deployment period. DatumForce.ai would operate only where permitted and does not assume use during deployment.

Try the score Conceptual

Try a sample baseline.

Explore an illustrative score using sleep estimates, steps, and a sample routine indicator. The labeled values are a made-up baseline, not recommended targets. Move a slider to see how this demonstration responds.

Sleep estimate7.0 h
Your baseline: 7.0 h
Activity9,000 steps
Your baseline: 9,000 steps
Routine regularitySteady
Your baseline: steady day-to-day timing
Illustrative only. This toy model lowers its number for changes in either direction from an invented baseline. Routine regularity is an illustrative derived indicator, not a fifth collected input. This is not a validated readiness rule or guidance about sleep or activity.
09:41Illustrative
Your scoreOnly you
80of 100
vs. baseline
0
Trend
Steady
Illustrative 14-day trend
Insight

These inputs match the illustrative baseline.

Privacy

Your data stays yours.

The proposed design would compute your baseline and score on your phone. Raw inputs and individual modeling would remain local; protected contributions to your unit's aggregate and the buddy requests you authorize may leave the phone.

  • Computed on your phoneRaw inputs and your individual model never leave the device by design.
  • No location. Period.DatumForce.ai does not collect geographic location data.
  • Aggregates only, one level upYou see your platoon's number, not your squad's and not any individual's. Leaders see aggregates only.
  • You hold the switchesYou choose which buddies can be contacted. You can revoke at any time.
Who sees what

Pick a seat. See the screen.

You. The only person who sees your score.

Your score, your trend, your progress, and a one-level-up unit aggregate for context.

  • Your own score and trend
  • Your streaks, XP, and badges
  • Your parent unit's aggregate (for example, platoon if you are in a squad)
  • Your private "X of 5 buddies connected" count
  • Anyone else's score
  • Your own small-unit breakdown
09:41Illustrative
Your scoreOnly you
78
Trend
Steady
vs. baseline
+4
Platoon aggregate
74
Buddies connected
3 of 5
Squad breakdown
Not shown. By design.
09:41Illustrative
FormationConceptual
Formation, 30 days
74 steady
Aggregate by week
W1W2W3Suppressed

Week 4 is suppressed: too few contributors. Small-group data is withheld to protect privacy.

Individual Soldier data is not available in this view.

09:41Illustrative
Check-inRequest
From DatumForce.ai

Hey. A buddy who added you could use a check-in this week. A quick text or call goes a long way.

I'll reach out
Not now

No score. No trend. No reason given. That is the whole message.

Level up Conceptual

Streaks. XP. Badges. Zero effect on your score.

The proposed progress system would recognize participation through streaks, XP, and badges. These illustrative rewards would stay separate from your readiness score.

Progress is separate from readiness. XP, streaks, and badges track participation only. They never feed your readiness score, and more buddies do not earn bonus points. Features shown are proposed.
Days logged Illustrative
84of 120
Current streak Illustrative
42days

Longest: 42 days

LVL 61,240 / 2,000 XP
Badges Illustrative2 of 4 unlocked
First 77-day streak
Crew upFirst buddy connected
Baseline set120 days logged
Back inPost-return check-in
Your crew

Pick up to five people who have your back.

Choose up to five buddies. A friend, a sibling, a battle buddy, even your commander. The proposed tool could send a general check-in request to buddies you authorize. The trigger rules still require review; buddies would never receive your score or risk details.

  • You choose themYou decide who is eligible. You can remove anyone at any time.
  • They never see your scoreBuddies receive a general check-in request. Never scores, trends, or risk details.
  • Same message for everyoneIf you add your commander, they get the same generic request as any other buddy. No special access.
  • Your count is private"X of 5 buddies connected" counts accepted connections only, and only you can see it. It earns no points and does not affect your score.
Private to you Illustrative
3 of 5
buddies connected
Illustrative
DatumForce.ai
To a buddy you chose
Hey. Someone who added you as a buddy could use a check-in this week. A quick text or call goes a long way.
No score. No details. No reason given.
Need to talk right now?

Dial 988, then press 1.

The Military Crisis Line is free, confidential, and open 24/7. DatumForce.ai would put it one tap away at all times. It is always your call, and it never depends on your score.

Call 988, press 1
For Leaders

Formation-level readiness insight, without individual data.

DatumForce.ai is a proposed readiness tool from NGZ AI. It is designed to help the Army assess changes in readiness before and after deployment without requiring wearables, routine self-report, or command access to individual behavioral data. The concept is at TRL 2 (provisional). It is not built, not validated, and not a medical device.

Aggregate reporting

Command personnel would receive aggregates only. A configurable parent-unit hierarchy selects the comparison group. It is not an individual or small-unit drill-down. Results would be suppressed when the parent group has too few contributors.

  • Minimum contributor counts and suppression
  • Limits on repeated queries
  • Soldiers compare against one level up only (a squad member sees a platoon aggregate, never squad or peer scores)
Army›Corps›Division›Regiment / Brigade›Squadron›Troop›Platoon›Squad

Example: a Soldier in a squad (lime) sees the platoon aggregate (olive). The hierarchy is configurable.

Low burden

No wearables, no operational tests, no routine surveys, and no new hardware to issue or charge. DatumForce.ai would use passive signals from Soldiers' own phones: steps, sleep-related estimates, screen time, and Bluetooth proximity patterns.

  • Deployment-aligned baselines: 120 days before and 120 days after
  • Complements, rather than replaces, specialized alertness or heat-risk tools
  • Participant-controlled support: up to five chosen buddies and participant-initiated 988 access
120 days before deployment  |  120 days after return

Soldier's own smartphone

Passive inputs. No wearables or surveys.
StepsSleep estimatesScreen timeBluetooth proximity patterns

Local baseline + readiness modeling

  • Own score, trend, gamified progress
  • Private: "X of 5 buddies connected"
  • Raw inputs remain local by design

Protected aggregate reporting*

Soldier: parent unit. Command: aggregates.

No peer scores or small-unit drill-down

Participant-authorized buddy requests

Up to 5 chosen buddies. Generic check-in only.

Never scores or risk details

Participant-initiated 988 connection

Available regardless of score or buddy response.

Dial 988, then press 1
Figure 1. Proposed architecture, not measured results. *Aggregation protocol and minimum reporting cohort require testing; moving one level up alone does not guarantee anonymity. Adapted from the NGZ AI xTech Search 10 concept white paper.
Take the concept with you

Download the concept brief

A one-page PDF for staff review that matches this site: problem, approach, privacy controls, proposed architecture, and status. Every statement is drawn from the xTech Search 10 white paper. v0.2 draft.

Download PDF

Concept status

DatumForce.ai is a design concept. Nothing on this page represents a built, tested, or validated capability. Table 1 of the white paper defines feasibility tests, not achieved results.

  • TRL 2, provisional. A distinct NGZ AI readiness concept. The proposed integration is not built or validated.
  • Phone-only sleep estimation, social-proximity interpretation, and score validity remain open R&D questions.
  • Full 120-day pre and post validation requires at least 240 observation days plus deployment, in a later, separately resourced study.
  • Azure Government is the planned backend. This is a plan, not an authorization claim. No compliance certification is claimed.
  • NIST AI RMF functions would govern permissions, map misuse, measure validity and privacy, and manage model changes.

What DatumForce.ai does not do

  • Does not make deployability determinations, automated or otherwise.
  • Does not make health care decisions. It is a readiness tool, not a medical device.
  • Does not give leaders access to any individual Soldier's data, score, or trend.
  • Does not collect geographic location data.
  • Does not assume use during deployment. It operates only where permitted.
  • Does not escalate automatically to command. Trigger thresholds and alert sequencing require safety review.
  • Does not count buddy connections or game engagement as a measure of readiness.

Proposed success measures: adoption, data coverage, useful change identification, trust, battery burden, and unwanted disclosures.

A concept from NGZ AI

Privacy-first by design. Focused on readiness.

DatumForce.ai is a distinct readiness concept from NGZ AI, Inc. Its proposed approach combines personal-phone baselines, on-device modeling, and protected aggregate insights for the periods before deployment and after return.

NGZ AIEthical, trustworthy AI for Soldier and Veteran health. Offices in Alexandria, Cambridge, and Williamsburg.

Proposed research collaborators span the University of Pennsylvania, William & Mary, and Stop Soldier Suicide. Roles and commitments for this new scope are pending confirmation.