Technology Landscape · Federal AI Intelligence

Technology landscape

How reported AI classifications spread across agencies, missions, lifecycle stages and sourcing. Each record carries one reported classification, so categories do not overlap. A classification describes the technique, not a specific commercial model.

Lifecycle by technology

Share of each technology's records

DeployedPilotPre-deploymentRetiredNot reported

Generative AI

Generative AI
Deployed: 20 (25.3%)Generative AI
Pilot: 13 (16.5%)Generative AI
Pre-deployment: 45 (57.0%)Generative AI
Retired: 1 (1.3%)

Classical / predictive ML

Classical / predictive ML
Deployed: 3 (33.3%)Classical / predictive ML
Pilot: 1 (11.1%)Classical / predictive ML
Pre-deployment: 5 (55.6%)

Natural language processing

Natural language processing
Deployed: 7 (29.2%)Natural language processing
Pilot: 4 (16.7%)Natural language processing
Pre-deployment: 13 (54.2%)

Computer vision

Computer vision
Pre-deployment: 2 (100.0%)

Agentic AI

Agentic AI
Deployed: 1 (12.5%)Agentic AI
Pilot: 2 (25.0%)Agentic AI
Pre-deployment: 5 (62.5%)

Reinforcement learning

Reinforcement learning
Pre-deployment: 1 (100.0%)

Other

Other
Deployed: 1 (33.3%)Other
Pilot: 1 (33.3%)Other
Pre-deployment: 1 (33.3%)

Not reported

Not reported
Deployed: 1 (16.7%)Not reported
Retired: 3 (50.0%)Not reported
Not reported: 2 (33.3%)

Sourcing by technology

Where reported

Developed in-houseContract and in-housePurchased from vendorNot reported

Generative AI

Generative AI
Developed in-house: 20 (25.3%)Generative AI
Contract and in-house: 12 (15.2%)Generative AI
Purchased from vendor: 9 (11.4%)Generative AI
Not reported: 38 (48.1%)

Classical / predictive ML

Classical / predictive ML
Developed in-house: 1 (11.1%)Classical / predictive ML
Contract and in-house: 2 (22.2%)Classical / predictive ML
Purchased from vendor: 2 (22.2%)Classical / predictive ML
Not reported: 4 (44.4%)

Natural language processing

Natural language processing
Developed in-house: 7 (29.2%)Natural language processing
Contract and in-house: 4 (16.7%)Natural language processing
Purchased from vendor: 2 (8.3%)Natural language processing
Not reported: 11 (45.8%)

Computer vision

Computer vision
Not reported: 2 (100.0%)

Agentic AI

Agentic AI
Developed in-house: 1 (12.5%)Agentic AI
Contract and in-house: 1 (12.5%)Agentic AI
Purchased from vendor: 1 (12.5%)Agentic AI
Not reported: 5 (62.5%)

Reinforcement learning

Reinforcement learning
Not reported: 1 (100.0%)

Other

Other
Developed in-house: 2 (66.7%)Other
Not reported: 1 (33.3%)

Not reported

Not reported
Not reported: 6 (100.0%)

Technology adoption by agency

Row share: each agency's records by classification · 16 largest portfolios in the filter

Agency by AI classification
	Generative AI	Classical / predictive ML	Natural language processing	Computer vision	Agentic AI	Reinforcement learning	Other	Not reported	n
HHS	50%	8%	21%		13%			8%	24
DOE	79%				5%		16%		19
VA	47%	13%	20%	13%		7%			15
TREAS	93%		7%						15
NASA	64%	9%	9%		18%				11
DHS	57%	14%	29%						7
SEC	83%				17%				6
DOI	20%	20%	40%		20%				5
GSA	75%		25%						4
DOJ	50%	25%	25%						4
NARA			100%						4
USDA	100%								3
FRB			100%						2
FDIC	50%		50%						2
DOL	100%								2
EPA			50%					50%	2

Technology by topic area

Row share within each topic

Topic by AI classification
	Generative AI	Classical / predictive ML	Natural language processing	Computer vision	Agentic AI	Reinforcement learning	Other	n
Administrative Functions	71%	6%	16%	3%	3%			31
Information Technology	79%		13%	4%	4%			24
Other	39%		39%		6%		17%	18
Science	54%	15%	8%		23%			13
Service Delivery	56%		33%			11%		9
Law Enforcement	50%	25%	25%					8
Procurement and Financial Management	86%		14%					7
Energy and the Environment	80%				20%			5
Health and Medical		40%	40%		20%			5
Cybersecurity	50%	50%						2
Human Resources	100%							2
Emergency Management	100%							1
Government Benefits Processing	100%							1

High-impact designation by technology

High-impactPresumed high-impact, determined notNot high-impactNot reported

Generative AI

Generative AI
High-impact: 1 (1.3%)Generative AI
Presumed high-impact, determined not: 1 (1.3%)Generative AI
Not high-impact: 74 (93.7%)Generative AI
Not reported: 3 (3.8%)

Classical / predictive ML

Classical / predictive ML
Not high-impact: 9 (100.0%)

Natural language processing

Natural language processing
Not high-impact: 23 (95.8%)Natural language processing
Not reported: 1 (4.2%)

Computer vision

Computer vision
Not high-impact: 2 (100.0%)

Agentic AI

Agentic AI
Presumed high-impact, determined not: 1 (12.5%)Agentic AI
Not high-impact: 7 (87.5%)

Reinforcement learning

Reinforcement learning
Not high-impact: 1 (100.0%)

Other

Other
Not high-impact: 3 (100.0%)

Not reported

Not reported
Not high-impact: 3 (50.0%)Not reported
Not reported: 3 (50.0%)

Governance disclosure by technology

High-impact records: reported in place / scope

Technology	High-impact	Testing	Impact assess.	Indep. review	Monitoring	Training
Generative AI	1	0 / 1	0 / 1	0 / 1	0 / 1	0 / 1
Classical / predictive ML	0	–	–	–	–	–
Natural language processing	0	–	–	–	–	–
Computer vision	0	–	–	–	–	–
Agentic AI	0	–	–	–	–	–
Reinforcement learning	0	–	–	–	–	–
Other	0	–	–	–	–	–
Not reported	0	–	–	–	–	–

Shows counts reported in place against all high-impact records of that technology. The remainder includes in-progress, not applicable and unanswered. Full governance view

Systems, vendors and technology are different things

29 of 132 records name a vendor (22%)

Microsoft77 records Amazon Web Services66 records OpenAI55 records Google33 records GDIT22 records Palantir22 records Accenture11 records Booz Allen Hamilton11 records Elder Research11 records ServiceNow11 records

Vendor: the supplier the agency names in a free-text field. Matched to a standard list of companies; one record may name several.

System name: the agency's name for the system (reported for 30% of records).

AI classification: the technique category the agency selected. A vendor name does not show which model or technique is used, and this platform does not infer it.