Original research | Baseline 2026-08-01

What the first united states AI evidence baseline shows.

A transparent snapshot of the comparison dataset, built to show where public evidence is strong and where a buyer should slow down.

Why read

Use this baseline to see evidence gaps before you trust a shortlist.

The short answer: the dataset is useful for deciding what to investigate next, not for declaring a safe, effective, or universally best product.

For: enterprise united states buyers who need to explain why a product is on, or missing from, a shortlist.

What to do with the numbers

Start with a category, market, workflow, and measurable outcome. Then open the product profiles and ask whether the source supports the exact intended use. A missing independent source is a reason to verify more, not proof that a product fails.

Evidence boundary: the counts below are calculated from the public comparison records at the review date. Read the comparison method before interpreting them.

Country research purpose

What the local research is trying to find out.

US enterprise AI decisions cross federal guidance, sector rules, state obligations, procurement, employment, consumer protection, cybersecurity, and rapidly changing vendor terms. Buyers need a practical way to connect a use case to evidence and ownership rather than a generic national winner.

The research connects search demand, official policy, sector evidence, product scope, verified review signals where available, and buyer questions. It is used to decide what deserves a page, what needs more evidence, and what should remain a watchlist item.

Read the country buyer guides · See the monthly market watch

Evidence baseline

The dataset is useful because its gaps are visible.

This is a descriptive baseline, not a claim about which vendor is best. Source type, scope, independence, market, and workflow boundaries still need buyer verification.

Profiles

30

Six products in each of five enterprise buying categories.

Independent or regulatory

11

Profiles with at least one independent-evidence or regulator source in the current dataset.

Vendor-only

0

Profiles where current public evidence is vendor-provided; these are not upgraded to evidence-backed by default.

Markets

1

The country pack keeps local policy, sector, procurement, and availability questions in scope.

Category view

Averages are context, not recommendations.

Average displayed scores include only fully assessed profiles and are rounded for readability. The category pages retain the underlying rationales and evidence gaps.

Baseline counts and average assessed evidence scores as at 2026-08-01
Category	Profiles	Average assessed score	Buyer lens
Enterprise knowledge and customer operations	6	3.7 / 5	CIOs, customer executives, knowledge leaders, contact-centre owners, legal, privacy, and security teams.
Software engineering and IT operations	6	3.9 / 5	CTOs, CIOs, engineering leaders, security teams, platform teams, and service owners.
Health and life sciences	6	3.8 / 5	Health executives, life-science leaders, clinical informatics, research operations, and safety and privacy owners.
Finance, risk, and fraud	6	4.0 / 5	Chief risk officers, financial-crime leaders, model-risk teams, compliance, fraud, and banking operations.
Public sector and critical infrastructure	6	4.0 / 5	Agency and infrastructure executives, operations leaders, security teams, procurement, and accountable system owners.

What this does not show: no score establishes operational impact, safety, legal compliance, current local availability, ROI, or procurement fit. Future snapshots should add dated independent evidence, buyer interview findings, and anonymised pilot learnings when those assets are genuinely available.

Sources and further reading

NIST AI Risk Management Framework standards guidance
NIST Generative AI Profile standards guidance
FTC guidance on AI claims standards guidance
CISA AI cybersecurity guidance standards guidance
HHS AI strategy and governance standards guidance
FDA AI/ML-enabled medical devices standards guidance
OMB federal AI guidance standards guidance
EEOC artificial intelligence and employment standards guidance

A practical next step

Turn an evidence gap into a measurable workflow question.

Enterprise AI Group describes a 6–8 week path for a defined business process, with governance, policy management, enterprise security, and Microsoft-tenant deployment considered from the start.

Enterprise AI Group describes a 6-8 week path for a defined workflow. Timing and cost depend on scope, users, integrations, security, governance, and support. These research pages are published by Enterprise AI Group. The implementation links describe optional services; they are not product endorsements or a replacement for local United States diligence.

See the governed platform approach

Do not include personal, confidential, regulated, or other sensitive information in an enquiry.

Keep the useful part

Tell us what you are deciding in United States.

Send the United States workflow, market, or category you are researching. We will use it to shape the next clear buyer brief.

Useful detail: include the market, workflow, or category behind a united states AI evidence gap.

Request a buyer brief Browse the categories

Please do not send personal, confidential, regulated, or other sensitive information.