brianletort.ai

AI Transformation Library

See how AI changes the work.

102 real-world cases. What changed, what improved, and where people still decide.

CardsTable

102 cases

49–72

CaseAI valueOpen
An editorial scene for Commonwealth of Pennsylvania contrasts researches, drafts, summarizes, or troubleshoots using conventional tools and manual synthesis. with uses the approved enterprise assistant to brainstorm, draft, simplify, summarize, research, or support it work. in the employee-centered chatgpt enterprise pilot with labor participation workflow.
Commonwealth of Pennsylvania		Participants reported an average 95 minutes saved per day of use.
An editorial scene for Rio Tinto contrasts drives heavy haul truck along mine roads and intersections. with navigates trucks on predefined gps courses using fleet position, speed, and direction data. in the autonomous haulage system for iron-ore operations workflow.
Rio Tinto		In 2018 each autonomous truck was estimated to operate 700 more hours on average with 15% lower costs.
An editorial scene for United Kingdom Serious Fraud Office contrasts reads documents to identify material potentially covered by legal professional privilege. with screen privilege content, categorize files, identify patterns, and remove duplicates. in the ai legal-privilege screening and investigative document analysis workflow.
United Kingdom Serious Fraud Office		The pilot processed up to 600,000 documents per day, described as 2,000 times faster.
An editorial scene for MIT randomized professional-writing study contrasts drafts a press release, report, analysis plan, or sensitive email from an occupation-specific prompt. with prompts the model for a draft or revision during the same occupation-specific task. in the chatgpt-assisted drafting with human submission workflow.
MIT randomized professional-writing study		Average time decreased 40%; output quality increased 18%
An editorial scene for Michigan Medicine / Epic Sepsis Model v1 contrasts monitor admitted patients and recognize sepsis through contemporary clinical practice. with calculates a proprietary risk score every 15 minutes and generates alerts at the selected threshold. in the real-time proprietary sepsis alerts workflow.
Michigan Medicine / Epic Sepsis Model v1		Missed 67% of sepsis while alerting on 18% of hospitalizations; AUROC .63
An editorial scene for New York City Fire Department contrasts select and schedule inspectable buildings from policy requirements and local information. with scores and ranks buildings from fire history, characteristics, violations, complaints, and other risk factors. in the firecast risk-ranked inspection lists through rbis workflow.
New York City Fire Department		Average violations increased 19% in the first 30 days and remained 10% higher after 60 days
An editorial scene for Google and Google DeepMind contrasts estimate uncertain output without the ml forecast and avoid firm hourly commitments. with predicts wind output 36 hours ahead from weather forecasts and turbine history. in the ml-recommended hourly wind-energy delivery commitments workflow.
Google and Google DeepMind		Roughly 20% higher value
An editorial scene for Maccabi Healthcare Services contrasts perform primary diagnosis on prostate biopsies with perform primary diagnosis on prostate biopsies in the prostate pathology quality control workflow.
Maccabi Healthcare Services		100%
An editorial scene for MODEC contrasts monitors equipment and addresses maintenance needs without a machine-learning early-warning layer. with analyze vessel operating data to detect deteriorating equipment performance and identify processing-plant problems. in the sensor, digital-twin, and machine-learning predictive maintenance on fpso mv29 workflow.
MODEC		MODEC reported a 65% reduction in downtime
An editorial scene for Pacific Gas and Electric Company contrasts detect faults from alarms, outages, or reports after conditions escalate. with scans sensor and smart-meter patterns for abnormal precursors. in the continuous monitoring center machine-learning risk triage workflow.
Pacific Gas and Electric Company		PG&E reported 17 potential ignitions intercepted, 12 million unplanned customer-outage minutes avoided, 2,620 emergency response hours reduced, and about $6 million saved
An editorial scene for Vodafone Germany contrasts handle simple navigation questions and requests that followed fixed decision processes. with interpret complex or vague requests and generate answers from vodafone web and internal product information. in the tobi to supertobi customer-service triage workflow.
Vodafone Germany		Vodafone reported that SuperTOBi resolved 25% more issues at first contact and independently resolved 74% of all inquiries in under ten seconds.
An editorial scene for Zurich Commercial Insurance (Zurich Insurance Group) contrasts manually read, review, and interpret every incoming risk submission (unstructured, multilingual documents) with converts multilingual submissions into structured risks aligned with zurich's underwriting framework. in the cytora headless risk digitization feeding z-horizon+ underwriting workbench workflow.
Zurich Commercial Insurance (Zurich Insurance Group)		15 minutes (80% reduction)
An editorial scene for BMW Plant Regensburg contrasts visually inspects painted vehicle surfaces and identifies areas needing treatment. with uses reflected stripe patterns to locate, size, map in 3d, and classify paint deviations on all surfaces. in the ai surface inspection and robotic processing workflow.
BMW Plant Regensburg		All vehicle surfaces inspected in an automated standard-production process
An editorial scene for Netherlands Tax and Customs Administration / Benefits contrasts reviews applications and available records to select suspected fraud for investigation. with scores applications using dozens of indicators, including citizenship, and selects cases as suspicious. in the self-learning risk-classification model for benefit applications workflow.
Netherlands Tax and Customs Administration / Benefits		The Dutch Data Protection Authority found improper and discriminatory processing from at least March 2016 to October 2018 because nationality was used in the model.
An editorial scene for EY contrasts defines the risk population and selects transactions for testing under ey's sample-first baseline. with loads general-ledger and subledger data and runs full-population analyses across erp systems. in the ey helix full-population general-ledger analysis workflow.
EY		Approximately 480 billion general-ledger lines loaded into the US analyzer in FY2025
An editorial scene for Rio Tinto Iron Ore contrasts drives the heavy-haul train between mine and port until the 12-hour shift limit. with controls speed, braking, and responses to restrictions and alarms across the private mainline. in the autohaul autonomous heavy-haul rail workflow.
Rio Tinto Iron Ore		Road travel eliminated
An editorial scene for UPS contrasts build and adjust a route sequence from service commitments and local knowledge. with calculates an optimized route that meets service commitments from package and network data. in the orion algorithmic route sequencing for delivery drivers workflow.
UPS		UPS reported 10 million gallons of fuel saved and 100,000 metric tons of emissions reduced each year
An editorial scene for Allegheny County DHS contrasts receives an eligible general protective service referral and reviews allegations and available county records. with uses administrative data to produce a 1–20 future-placement risk score for eligible referrals. in the afst risk-assisted investigation screening workflow.
Allegheny County DHS		Improved identification and modest reduction in racial disparity in case openings
An editorial scene for Copenhagen Emergency Medical Services contrasts listen to the caller, recognize suspected cardiac arrest, and initiate the protocol. with listens to ongoing calls and flags suspected cardiac arrest. in the real-time ml listening and dispatcher cardiac-arrest alerts workflow.
Copenhagen Emergency Medical Services		93.1% with the alert, P=.15; no statistically significant improvement
An editorial scene for UK Department for Work and Pensions contrasts receives a new claimant request and applies existing eligibility and fraud controls. with scores new advance requests and refers predicted high-risk cases. in the ml risk referral with human fraud decision workflow.
UK Department for Work and Pensions		2.9 times as effective; estimated £4.4m saved
An editorial scene for European Centre for Medium-Range Weather Forecasts contrasts runs physics-based numerical weather prediction from operational analyses. with generates deterministic global forecasts four times daily from ecmwf analyses. in the aifs single operational ml forecast production workflow.
European Centre for Medium-Range Weather Forecasts		4–6% overall skill improvement; up to 12% precipitation improvement in v1.1.0
An editorial scene for Anonymous Fortune 500 software company contrasts reads the incoming chat, searches available knowledge, and drafts a response. with observes the live chat and proposes response text based on patterns from prior interactions. in the generative ai agent assistance workflow.
Anonymous Fortune 500 software company		About 14% average increase; 35% for novice/low-skill agents
An editorial scene for Klarna contrasts produce images and translate campaign collateral with generate marketing images and translate copy in the ai marketing production workflow.
Klarna		-25%
An editorial scene for University of Liverpool contrasts chooses formulations, prepares samples, operates instruments, records results, and decides the next experiment. with selects the next batch from prior measured outcomes. in the mobile robotic chemist with bayesian experiment selection workflow.
University of Liverpool		688 experiments in eight days; best mixture six times more active
Curated collections →Evidence methodologyPublication history
AI Transformation Library — Page 3 | Brian Letort