Cybatar Enterprise AI Evidence

What is actually working in enterprise AI.

Structured evidence of AI deployments, operating outcomes and implementation patterns across industries. Each deployment retains its source and provenance so reported results remain distinguishable from Cybatar-measured evidence.

76tracked deployments
75organisations
158reported outcome measures
14sectors represented
Evidence standard. Provider-reported outcomes are preserved as provider-reported. Independent external evidence and Cybatar-measured results are labelled separately when available. Inclusion is not an endorsement, and organisations shown are not implied to be CRG clients.
Reset
provider reported

NHS England

HealthcareUnited KingdomMicrosoft

NHS England announced a large-scale Microsoft 365 Copilot rollout following a 30,000-worker trial and is adding agent-building capabilities.

505,000 peopleplanned users
30,000 peopletrial users
43 minutesadministrative time saved per user daily
Source: Microsoft · verified 23 Aug 2026
provider reported

Norges Bank Investment Management

Financial ServicesNorwayAnthropic

NBIM uses Claude to augment analytical work across its investment organization and portfolio research.

20 percentproductivity gain
213,000 hourshours saved
9,000 companiesportfolio companies covered
Source: Anthropic · verified 23 Aug 2026
provider reported

Novo Nordisk

Life SciencesDenmarkAnthropic

Novo Nordisk uses Claude in clinical documentation workflows to compress document production and review cycles.

99.9 percentclinical documentation time reduction
50 percentreview cycle reduction
Source: Anthropic · verified 23 Aug 2026
provider reported

NVIDIA

Technology & SoftwareUnited StatesOpenAI

NVIDIA engineers and researchers use Codex for complex engineering tasks and end-to-end machine-learning experiments.

10 xresearch workflow speed improvement
40,000 peopleemployees with Codex access
Source: OpenAI · verified 23 Aug 2026
provider reported

OneUnited Bank

Financial ServicesUnited StatesGoogle

OneUnited uses Gemini in integrated workflows including fraud-pattern analysis on customer disputes.

Source: Google Cloud · verified 23 Aug 2026
provider reported

Orion Health

HealthcareNew Zealand / GlobalAmazon / AWS

Orion Health built Oribot on Amazon Bedrock to retrieve answers from more than half a million internal records.

50 hoursstaff hours reclaimed daily
10 xrelative cost efficiency
2 monthsprototype delivery time
1 minuteanswer retrieval time
Source: AWS · verified 23 Aug 2026
provider reported

Philips

HealthcareGlobalOpenAI

Philips is scaling AI literacy and hands-on use across its global workforce as part of responsible AI adoption.

70,000 peopleemployees in AI literacy programme
Source: OpenAI · verified 23 Aug 2026
provider reported

Public Service Commission of Canada

GovernmentCanadaMicrosoft

PSC built an in-house Microsoft 365 Copilot agent for bilingual travel-policy support and travel approval queries.

90 percenttravel mailbox email reduction
5 dayscomplex query time before
30 secondscomplex query time after
120,000 CADannual salary cost avoidance
Source: Microsoft · verified 23 Aug 2026
provider reported

Quadrivia

HealthcareGlobalGoogle

Quadrivia uses Gemini to automate routine clinical tasks with low-latency multimodal patient interactions and safety controls.

10 xtime-to-first-token improvement
98.4 percentno-harm clinical safety
82.3 percentcare delivery success
1.8 secondsaverage response time
Source: Google Cloud · verified 23 Aug 2026
provider reported

Rakuten

Technology & SoftwareJapanAnthropic

Rakuten uses Claude Code to accelerate feature development.

79 percentfeature development time reduction
Source: Anthropic · verified 23 Aug 2026
provider reported

Ramp

Financial ServicesUnited StatesAnthropic

Ramp uses Claude Code across engineering and incident investigation workflows.

1,000,000 linesAI-suggested code implemented in 30 days
50 percentweekly active engineering usage
80 percentincident investigation time reduction
Source: Anthropic · verified 23 Aug 2026
provider reported

Reveleer

HealthcareUnited StatesGoogle

Reveleer uses Gemini agents to interpret clinical charts and support prospective risk adjustment with traceable evidence.

50 percentnoise reduction versus legacy NLP
Source: Google Cloud · verified 23 Aug 2026
provider reported

Rising Academies

EducationSub-Saharan AfricaAnthropic

Rising Academies uses Claude in WhatsApp-based student tutoring and teacher support tools across multiple African countries.

150,000 studentsstudents reached
0.3 standard deviationslearning outcome effect size
Source: Anthropic · verified 23 Aug 2026
provider reported

Rogo

Financial ServicesUnited StatesGoogle

Rogo uses Gemini and Vertex AI to deliver agentic financial research for investment bankers and analysts.

6,000 usersinvestment bankers and analysts served
34.1 percenthallucination rate before
3.9 percenthallucination rate after
Source: Google Cloud · verified 23 Aug 2026
provider reported

SK Telecom

TelecommunicationsSouth KoreaAnthropic

SK Telecom uses Claude in Amazon Bedrock for in-call agent assistance and post-call processing in customer service.

34 percenthuman-agent quality rating increase
68 percentlow-quality response reduction
89 percentpost-call processing quality versus human agents
Source: Anthropic · verified 23 Aug 2026
provider reported

Starling Bank

Financial ServicesUnited KingdomGoogle

Starling uses Gemini multimodal capabilities in Scam Intelligence to assess marketplace ads and help customers identify scam signals.

Source: Google Cloud · verified 23 Aug 2026
provider reported

Steno

LegalUnited StatesAnthropic

Steno uses Claude to analyze legal deposition transcripts and provide citation-linked insights to attorneys.

1,200 firmslaw firms served
1 weekproof-of-concept build time
Source: Anthropic · verified 23 Aug 2026
provider reported

Sun Life

Financial ServicesGlobalAmazon / AWS

Sun Life moved from generative-AI experiments to a secure internal chatbot used for employee knowledge access.

10,000 queriesqueries answered weekly
40 experimentsgenerative AI experiments
Source: AWS · verified 23 Aug 2026
provider reported

Taisei Corporation

Construction & InfrastructureJapanOpenAI

Taisei rolled ChatGPT Enterprise into talent development and company-wide knowledge work.

3,300 GPTscustom GPTs created
90 percentweekly active usage
5.5 hourstime saved per employee weekly
Source: OpenAI · verified 23 Aug 2026
provider reported

TELUS

TelecommunicationsCanadaAnthropic

TELUS has scaled Claude across its workforce and custom internal AI solutions, including software engineering.

57,000 peopleteam members using Claude
13,000 solutionscustom AI solutions created
30 percentengineering shipping speed increase
500,000 hourshours saved
Source: Anthropic · verified 23 Aug 2026
provider reported

The Clinician

HealthcareGlobalGoogle

The Clinician uses Gemini to automate configuration and launch of healthcare outcomes projects while enforcing compliance controls.

42 daysproject authoring time before
3.1 hoursproject authoring time after
47 projectsmonthly project launches
2.1 million USDannualized external cost savings
Source: Google Cloud · verified 23 Aug 2026
provider reported

Travelers

InsuranceUnited StatesOpenAI

Travelers deployed an autonomous voice Claim Assistant for first notice of loss and expanded it countrywide after an initial eight-state launch.

90 percentcustomer completion through AI Claim Assistant
Source: OpenAI · verified 23 Aug 2026
provider reported

Ufonia

HealthcareUnited KingdomGoogle

Ufonia uses Gemini and Google Cloud to automate routine clinical conversations with continuous model evaluation and safety testing.

110,000 consultationspatient consultations completed
30 percentmonthly cloud spend reduction
10 xactivity increase
Source: Google Cloud · verified 23 Aug 2026
provider reported

Virgin Atlantic

Transportation & AviationUnited KingdomOpenAI

Virgin Atlantic used Codex to refactor legacy code, improve test coverage and ship customer-facing software.

80 percentlegacy codebase size reduction
100 percentunit test coverage on new app
30 minuteslegacy refactor time
Source: OpenAI · verified 23 Aug 2026
From evidence to action

Explore the use-case taxonomy behind the deployments.

Cybatar maps enterprise evidence to structured use cases so leaders can move from anecdotes to repeatable implementation patterns.

Explore AI use cases