request_quote
A
CFO · Finance & Accounting
AP/AR automation & PO matching
Reads loan and commercial agreements, extracts key attributes, and matches them against rules and ledgers.
Enterprise AI Field Guide
Use cases, ranked by
what actually ships.
34 enterprise AI patterns across 10 domains. Each graded A+ through D on effort, ROI, token economics, and substance. Filter by role, industry, or technology to find the work worth doing now.
CFO · Finance & Accounting
AP/AR automation & PO matching
AScore breakdown
Effort
B
ROI Floor
A
ROI Ceiling
A
Token Econ
A
Substance
A+
Overall
A
What it does
Reads loan and commercial agreements, extracts key attributes, and matches them against rules and ledgers. JPMorgan's COiN interprets agreements that "formerly consumed 360,000 hours of work each year," reviewing documents "in seconds" and "less error-prone."
Vendor-motive: genuinely impactful and vendor-independent, JPMorgan built COiN in-house to solve its own contract-review bottleneck, reported via Bloomberg rather than a platform showcase, so no provider is being sold here.
Sources Independent
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
payments
B
CFO · Finance & Accounting
Cash-flow forecasting
Aggregates ERP/CRM/receivables data to project future cash position for treasury.
CFO · Finance & Accounting
Cash-flow forecasting
BScore breakdown
Effort
B
ROI Floor
B
ROI Ceiling
A
Token Econ
A
Substance
A
Overall
B
What it does
Aggregates ERP/CRM/receivables data to project future cash position for treasury. J.P. Morgan describes ML-driven treasury forecasting blending internal and market data.
Sources First-party
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
database
B
CFO · Finance & Accounting
Conversational analytics / text-to-SQL (finance)
A finance analyst asks a plain-English question in Slack and an agent writes the SQL, checks the user's permissions, runs it against curated data marts, and returns the figure in seconds.
CFO · Finance & Accounting
Conversational analytics / text-to-SQL (finance)
BScore breakdown
Effort
B
ROI Floor
B
ROI Ceiling
A
Token Econ
B
Substance
A
Overall
B
What it does
A finance analyst asks a plain-English question in Slack and an agent writes the SQL, checks the user's permissions, runs it against curated data marts, and returns the figure in seconds. Uber's Finch does exactly this for its FinTech team, replacing manual querying across Presto, Oracle EPM, and IBM Planning Analytics; Uber says answers come "within seconds" instead of "hours—or even days."
Vendor-motive: genuinely impactful and platform-independent, Finch is Uber's own internal build described on its engineering blog, not a product being sold, and the text-to-SQL pattern is provider-agnostic (it runs on Uber's own GenAI Gateway over swappable models). Note Uber publishes no hard metrics and admits Finch "is prone to hallucination," so the speed claim is qualitative.
Sources First-party
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
recommend
A
CMO · Marketing
Personalized product recommendations
Matches shoppers to products from behavioral and catalog signals.
CMO · Marketing
Personalized product recommendations
AScore breakdown
Effort
B
ROI Floor
A
ROI Ceiling
A
Token Econ
B
Substance
A
Overall
A
What it does
Matches shoppers to products from behavioral and catalog signals. Unilever accomplished this with BeautyHub PRO (Philippines & Thailand), where consumers take a short quiz and submit a selfie; computer vision reads up to 30 visual data points to suggest skincare and haircare across multiple in-house beauty brands. Reported lift vs. other channels: users are "43% more likely to complete a purchase," basket value is "39% higher." A related Dove Scalp + Hair Therapist (US) uses generative-AI Q&A trained on dermatologist content for scalp/hair recommendations with one-click Amazon checkout.
Vendor-motive: genuinely impactful and vendor-independent, this is Unilever describing its own first-party product on its own newsroom, with no platform being sold.
Sources First-party
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
mail
C
CMO · Marketing
Personalized email/copy generation at scale
Generates and optimizes message variants tuned to convert.
CMO · Marketing
Personalized email/copy generation at scale
CScore breakdown
Effort
A
ROI Floor
C
ROI Ceiling
B
Token Econ
C
Substance
C
Overall
C
What it does
Generates and optimizes message variants tuned to convert. Vanguard Institutional's AI-generated LinkedIn messaging "boosted conversion rate 15%", note the win is targeting lift, not raw volume.
Vendor-motive: reads as a genuine result reported through trade press rather than a platform pitch; the lift is the point, and it isn't tied to any one provider's stack.
Sources Independent
Minimum AI maturity required
2
Level 2
Operational
A few weeks of integration. Needs standardized data sources and a basic deployment story.
movie
C
CMO · Marketing
AI-generated marketing videos/creative
Produces ad creative and video assets from briefs/templates.
CMO · Marketing
AI-generated marketing videos/creative
CScore breakdown
Effort
B
ROI Floor
C
ROI Ceiling
B
Token Econ
C
Substance
C
Overall
C
What it does
Produces ad creative and video assets from briefs/templates. Unilever's Beauty AI Studio produces assets "up to 30% faster," "more than doubled" video-completion and click-through rates, and cut content costs "87%" (TRESemmé Thailand).
Vendor-motive: genuinely impactful and platform-independent, Unilever's own studio build, product-agnostic, with cost/speed gains it would have chased regardless of vendor.
Sources First-party
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
recommend
C
CMO · Marketing
Personalized recommendations engine (Starbucks Deep Brew)
ML personalization engine serving product recommendations and offers across mobile app and drive-thru, built on Azure ML/Databricks and launched in 2019.
CMO · Marketing
Personalized recommendations engine (Starbucks Deep Brew)
CScore breakdown
Effort
B
ROI Floor
C
ROI Ceiling
B
Token Econ
B
Substance
C
Overall
C
What it does
ML personalization engine serving product recommendations and offers across mobile app and drive-thru, built on Azure ML/Databricks and launched in 2019. Real, named deployment but no first-party performance metric exists — claims circulate via agency blog posts, and company-wide revenue growth is not attributable to the engine itself.
Sources Vendor
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
diversity_3
D
CMO · Marketing
Synthetic customer cohorts to pre-test campaigns
⚠ HYPOTHETICAL, Generates simulated personas to test messaging pre-launch.
CMO · Marketing
Synthetic customer cohorts to pre-test campaigns
DScore breakdown
Effort
C
ROI Floor
C
ROI Ceiling
B
Token Econ
C
Substance
D
Overall
D
What it does
⚠ HYPOTHETICAL, Generates simulated personas to test messaging pre-launch. Bain says "preliminary experience suggests... comparable insights in half the time and at one-third the cost," citing a Stanford/Google DeepMind study where agents "matched human survey responses with 85% accuracy", explicitly framed as emerging with limits.
Sources Independent
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
price_change
B
CRO · Sales
Dynamic pricing optimization
Resets quoted prices by segment using willingness-to-pay signals.
CRO · Sales
Dynamic pricing optimization
BScore breakdown
Effort
C
ROI Floor
B
ROI Ceiling
A
Token Econ
B
Substance
A
Overall
B
What it does
Resets quoted prices by segment using willingness-to-pay signals. McKinsey documents self-learning B2B pricing engines incorporating per-segment elasticity, with real deployments.
Sources Independent
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
contacts
A+
CCO · Customer Service
Automated post-call summarization & CRM notes
Transcribes a call and writes the wrap-up note straight into CRM.
CCO · Customer Service
Automated post-call summarization & CRM notes
A+Score breakdown
Effort
A
ROI Floor
A
ROI Ceiling
A
Token Econ
A+
Substance
A+
Overall
A+
What it does
Transcribes a call and writes the wrap-up note straight into CRM. Morgan Stanley's Debrief (GPT-4 + Whisper) auto-drafts client-meeting notes into CRM; a pilot advisor reported saving "30 minutes of work per meeting."
Vendor-motive: partly an OpenAI showcase, but the task is provider-agnostic, Morgan Stanley would have built call summarization on any capable model, so the 30-minutes-saved result reads as genuine despite the citation being a platform story.
Sources Vendor
Minimum AI maturity required
2
Level 2
Operational
A few weeks of integration. Needs standardized data sources and a basic deployment story.
smart_toy
C
CCO · Customer Service
Autonomous chatbots resolving inquiries end-to-end
Resolves routine inquiries end-to-end without human handoff.
CCO · Customer Service
Autonomous chatbots resolving inquiries end-to-end
CScore breakdown
Effort
C
ROI Floor
C
ROI Ceiling
A+
Token Econ
C
Substance
C
Overall
C
What it does
Resolves routine inquiries end-to-end without human handoff. Klarna's assistant did "the equivalent work of 700 full-time agents," handled "2.3 million conversations," cut resolution to "less than 2 mins compared to 11 mins," and drove "$40 million USD in profit improvement"; Vodafone's TOBi solves "70%... at the first time of asking." ***Caveat:*** heavily an OpenAI flagship-logo showcase, and Klarna's later partial walk-back (rehiring human agents after conceding cost-driven automation produced "lower quality") suggests the headline served the vendor's narrative more than Klarna's durable outcome; the ceiling is real, the floor is shakier than the original claim implied.
Sources Vendor
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
event
A
CHRO · Human Resources
Interview scheduling & candidate engagement
Screens applications and runs scheduling/engagement at applicant scale.
CHRO · Human Resources
Interview scheduling & candidate engagement
AScore breakdown
Effort
A
ROI Floor
A
ROI Ceiling
B
Token Econ
A
Substance
A
Overall
A
What it does
Screens applications and runs scheduling/engagement at applicant scale. Unilever (with HireVue/Pymetrics) "saved 50,000 hours in candidate time over 18 months," "£1M annual cost savings," "90% reduction in time to hire," and "16% increase in diversity hires."
Vendor-motive: a HireVue platform sell (the case study exists to promote HireVue/Pymetrics), but the scale of Unilever's hiring problem was real and the gains are substantial enough to read as a genuine result, not just a pitch.
Sources Vendor
Minimum AI maturity required
2
Level 2
Operational
A few weeks of integration. Needs standardized data sources and a basic deployment story.
school
B
CHRO · Human Resources
Skills-gap analysis & skills-ontology mapping
Maps workforce skills to a taxonomy and surfaces gaps.
CHRO · Human Resources
Skills-gap analysis & skills-ontology mapping
BScore breakdown
Effort
C
ROI Floor
B
ROI Ceiling
A
Token Econ
B
Substance
A
Overall
B
What it does
Maps workforce skills to a taxonomy and surfaces gaps. SHRM documents skills-ontology software (e.g., Workday Skills Cloud) inventorying skills and identifying gaps.
Sources Independent
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
reviews
C
CHRO · Human Resources
Performance-review drafting
Generates first-pass reviews from feedback inputs.
CHRO · Human Resources
Performance-review drafting
CScore breakdown
Effort
A
ROI Floor
C
ROI Ceiling
B
Token Econ
B
Substance
C
Overall
C
What it does
Generates first-pass reviews from feedback inputs. HR Dive documents named deployments (Citi Performance Assist, JPMorgan) auto-drafting evaluations.
Sources Independent
Minimum AI maturity required
2
Level 2
Operational
A few weeks of integration. Needs standardized data sources and a basic deployment story.
trending_down
C
CHRO · Human Resources
Attrition / retention prediction
Predicts which employees are likely to leave.
CHRO · Human Resources
Attrition / retention prediction
CScore breakdown
Effort
C
ROI Floor
C
ROI Ceiling
B
Token Econ
B
Substance
C
Overall
C
What it does
Predicts which employees are likely to leave. A peer-reviewed Scientific Reports paper establishes ML attrition models (SVM, XGBoost, Random Forest).
Sources Independent
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
groups_2
C
CHRO · Human Resources
Workforce planning & talent-demand forecasting
Forecasts future hiring and skill needs.
CHRO · Human Resources
Workforce planning & talent-demand forecasting
CScore breakdown
Effort
C
ROI Floor
C
ROI Ceiling
B
Token Econ
C
Substance
C
Overall
C
What it does
Forecasts future hiring and skill needs. Deloitte documents AI workforce planning with named examples (Verizon, Cleveland Clinic).
Sources Independent
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
rule
A+
GC · Legal, Risk & Compliance
Contract review & clause extraction
Reads agreements and pulls key clauses/obligations for review.
GC · Legal, Risk & Compliance
Contract review & clause extraction
A+Score breakdown
Effort
B
ROI Floor
A
ROI Ceiling
A+
Token Econ
A
Substance
A+
Overall
A+
What it does
Reads agreements and pulls key clauses/obligations for review. JPMorgan's COiN interprets agreements that "formerly consumed 360,000 hours of work each year," in "seconds" and "less error-prone."
Vendor-motive: genuinely impactful and vendor-independent, JPMorgan's in-house build reported via Bloomberg; no platform is being sold.
Sources Independent
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
edit_document
A
GC · Legal, Risk & Compliance
Contract drafting & redlining
Drafts and redlines contracts and synthesizes case law with citations.
GC · Legal, Risk & Compliance
Contract drafting & redlining
AScore breakdown
Effort
A
ROI Floor
A
ROI Ceiling
A
Token Econ
A
Substance
A
Overall
A
What it does
Drafts and redlines contracts and synthesizes case law with citations. At Allen & Overy, partner David Wakeling said Harvey can save lawyers "a couple hours a week-plus," deployed to ~3,500 lawyers generating ~40,000 queries across 250 practice areas.
Vendor-motive: somewhat a Harvey platform sell, but the source is Reuters/ABA journalism (not Harvey's own channel) and A&O's adoption scale reads as a genuine practice decision rather than a staged endorsement.
Sources Independent
Minimum AI maturity required
2
Level 2
Operational
A few weeks of integration. Needs standardized data sources and a basic deployment story.
gavel
C
GC · Legal, Risk & Compliance
Predictive compliance risk assessment
⚠ HYPOTHETICAL, Scores the likelihood of compliance risk/breach.
GC · Legal, Risk & Compliance
Predictive compliance risk assessment
CScore breakdown
Effort
C
ROI Floor
C
ROI Ceiling
B
Token Econ
C
Substance
C
Overall
C
What it does
⚠ HYPOTHETICAL, Scores the likelihood of compliance risk/breach. Deloitte frames AI/ML compliance-risk scoring as an advisory capability with little scaled, validated production evidence and disclosed accuracy.
Sources Vendor
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
code
A+
CTO · Engineering & IT
Code generation & autocompletion
Suggests and completes code inline as developers type.
CTO · Engineering & IT
Code generation & autocompletion
A+Score breakdown
Effort
A
ROI Floor
A
ROI Ceiling
A+
Token Econ
A
Substance
A+
Overall
A+
What it does
Suggests and completes code inline as developers type. GitHub reports Copilot users finished a task in "1 hour and 11 minutes" vs "2 hours and 41 minutes" (55% faster) and that Copilot now generates "46% of code."
Vendor-motive: a GitHub/Microsoft Copilot sell, but corroborated by independent MIT Sloan/Vanguard and McKinsey figures, the productivity gain is one of the best-evidenced on the board, so the vendor framing is less concerning here.
Sources First-party
Minimum AI maturity required
2
Level 2
Operational
A few weeks of integration. Needs standardized data sources and a basic deployment story.
manage_search
A
CTO · Engineering & IT
Enterprise knowledge search assistant
Answers employee questions over a large internal document corpus.
CTO · Engineering & IT
Enterprise knowledge search assistant
AScore breakdown
Effort
B
ROI Floor
A
ROI Ceiling
A
Token Econ
A
Substance
A
Overall
A
What it does
Answers employee questions over a large internal document corpus. Morgan Stanley reports "over 98% of advisor teams actively use" its assistant, with document access "from 20% to 80%."
Vendor-motive: an OpenAI showcase, but the adoption metric is Morgan Stanley's own operational result and the RAG-over-documents pattern is provider-agnostic, reads as a genuine win.
Sources Vendor
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
database
B
CTO · Engineering & IT
Self-serve text-to-SQL for engineers/analysts
A user describes the data they want in plain English and a multi-agent pipeline picks the business domain, selects tables, prunes the schema to fit the context window, and generates the SQL.
CTO · Engineering & IT
Self-serve text-to-SQL for engineers/analysts
BScore breakdown
Effort
B
ROI Floor
B
ROI Ceiling
A
Token Econ
B
Substance
A
Overall
B
What it does
A user describes the data they want in plain English and a multi-agent pipeline picks the business domain, selects tables, prunes the schema to fit the context window, and generates the SQL. Uber's QueryGPT does this company-wide; Uber reports it cuts query authoring "from ~10 minutes to ~3 minutes" (~70% faster).
Vendor-motive: genuinely impactful and platform-independent, QueryGPT is Uber's own internal build on its GenAI Gateway (provider-agnostic), described on its engineering blog, not a product being sold. Honest caveat: Uber calls it a *limited release* averaging "about 300 daily active users," with 78% saying it saved time, so the ~70% figure is real but not yet company-wide, and hallucinated tables/columns remain a known issue.
Sources First-party
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
insights
A
COO · Operations & Supply Chain
Demand forecasting & demand sensing
Predicts demand across the network.
COO · Operations & Supply Chain
Demand forecasting & demand sensing
AScore breakdown
Effort
C
ROI Floor
A
ROI Ceiling
A+
Token Econ
A
Substance
A+
Overall
A
What it does
Predicts demand across the network. Walmart SVP Parvez Musani says AI "has transformed our approach to demand forecasting, inventory flow, and cost optimization," with projects "that once took months" now done "in weeks," using an in-house multi-horizon RNN.
Vendor-motive: genuinely impactful and platform-independent, Walmart built this in-house on its own models to solve a real forecasting problem; no vendor is being sold.
Sources Independent
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
build
A
COO · Operations & Supply Chain
Predictive maintenance of equipment
Flags equipment faults before they cause downtime.
COO · Operations & Supply Chain
Predictive maintenance of equipment
AScore breakdown
Effort
C
ROI Floor
A
ROI Ceiling
A+
Token Econ
A
Substance
A+
Overall
A
What it does
Flags equipment faults before they cause downtime. Siemens Senseye saved BlueScope "~2,000 hours of unplanned downtime across three years"; Sachsenmilch saved "in the low six figures" from early fault detection; Siemens cites "up to 50%" downtime reduction.
Vendor-motive: a Siemens Senseye platform sell, these are Siemens customer cases and the figures assume sensors/assets feed Senseye; the downtime savings are concrete but tied to the product. Kept on the strength of named-customer attribution and the volume of asset-side integration work the buyer still owns.
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
local_shipping
A
COO · Operations & Supply Chain
Logistics & delivery-route optimization
Continuously re-routes deliveries against live conditions.
COO · Operations & Supply Chain
Logistics & delivery-route optimization
AScore breakdown
Effort
C
ROI Floor
A
ROI Ceiling
A
Token Econ
A
Substance
A
Overall
A
What it does
Continuously re-routes deliveries against live conditions. Walmart describes a "multi-agent architecture" making "micro-decisions about routing, driver availability and assignment," recalibrating to "weather, traffic and demand."
Vendor-motive: genuinely impactful and vendor-independent, Walmart's own engineering blog describing a system it built itself, no platform pitch.
Sources First-party
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
warehouse
B
COO · Operations & Supply Chain
Warehouse computer-vision inspection
Inspects line/warehouse output visually for defects.
COO · Operations & Supply Chain
Warehouse computer-vision inspection
BScore breakdown
Effort
C
ROI Floor
B
ROI Ceiling
A
Token Econ
B
Substance
A
Overall
B
What it does
Inspects line/warehouse output visually for defects. Unilever's Hefei factory saw "8% increase in overall equipment effectiveness and 20% wastage reduction."
Vendor-motive: genuinely impactful and platform-independent, Unilever's first-party factory result, product-agnostic.
Sources First-party
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
lightbulb
C
COO · Operations & Supply Chain
New-product-development trade-off optimization
⚠ HYPOTHETICAL, Generates and evaluates design candidates with in-silico testing.
COO · Operations & Supply Chain
New-product-development trade-off optimization
CScore breakdown
Effort
D
ROI Floor
C
ROI Ceiling
B
Token Econ
C
Substance
C
Overall
C
What it does
⚠ HYPOTHETICAL, Generates and evaluates design candidates with in-silico testing. McKinsey describes AI generative design, but its 2025 data indicates most organizations are not yet using AI agents in product development.
Sources Independent
Minimum AI maturity required
4
Level 4
Transformational
Year-long program. Needs deep system integration, custom models, and executive sponsorship.
library_books
A
R&D · Research & Development
R&D literature-review research assistants
Synthesizes prior research and proposes/screens design candidates computationally.
R&D · Research & Development
R&D literature-review research assistants
AScore breakdown
Effort
A
ROI Floor
A
ROI Ceiling
A
Token Econ
A
Substance
A+
Overall
A
What it does
Synthesizes prior research and proposes/screens design candidates computationally. Unilever says it can "compress decades of lab work into days."
Vendor-motive: Unilever's framing is first-party and genuine; the research-assistant value holds regardless of provider.
Sources First-party
Minimum AI maturity required
2
Level 2
Operational
A few weeks of integration. Needs standardized data sources and a basic deployment story.
diversity_3
B
R&D · Research & Development
Synthetic data generation for trial design
⚠ HYPOTHETICAL, Generates synthetic control/patient data to simulate and optimize trial design.
R&D · Research & Development
Synthetic data generation for trial design
BScore breakdown
Effort
C
ROI Floor
B
ROI Ceiling
A
Token Econ
B
Substance
A
Overall
B
What it does
⚠ HYPOTHETICAL, Generates synthetic control/patient data to simulate and optimize trial design. Deloitte uses conditional language; FDA has issued only draft AI guidance, and synthetic data is treated as simulation, not confirmatory evidence.
Sources Independent
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
biotech
C
R&D · Research & Development
Generative drug discovery / molecule design
Designs candidate molecules and prioritizes them for synthesis.
R&D · Research & Development
Generative drug discovery / molecule design
CScore breakdown
Effort
D
ROI Floor
C
ROI Ceiling
A+
Token Econ
D
Substance
A
Overall
C
What it does
Designs candidate molecules and prioritizes them for synthesis. Insilico Medicine reached a preclinical candidate "in under 18 months" for "one-tenth of the cost and one-third of the time" of traditional methods.
Vendor-motive: partly an NVIDIA showcase (published on NVIDIA's blog to sell GPU/BioNeMo compute), but Insilico's clinical milestone is a real, independently notable scientific result, not just a vendor demo.
Sources Vendor
Minimum AI maturity required
4
Level 4
Transformational
Year-long program. Needs deep system integration, custom models, and executive sponsorship.
article
A+
Industry · Industry Verticals
Clinical documentation / ambient note-taking
Listens to a visit and drafts the clinical note.
Industry · Industry Verticals · Healthcare
Clinical documentation / ambient note-taking
A+Score breakdown
Effort
B
ROI Floor
A
ROI Ceiling
A+
Token Econ
A
Substance
A+
Overall
A+
What it does
Listens to a visit and drafts the clinical note. DAX Copilot users report "7 minutes saved per encounter," "50% less time spent on documentation," and "70% reduction in feelings of burnout."
Vendor-motive: a Microsoft/Nuance DAX platform sell, the metrics are the product's pitch, but ambient documentation is a genuine clinician pain point and the time-saved result is widely replicated across DAX customers, so the impact reads as real.
Sources Vendor
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
request_page
A
Industry · Industry Verticals
Insurance claims intake (FNOL) & triage
Captures first notice of loss and triages/schedules next steps.
Industry · Industry Verticals · Insurance
Insurance claims intake (FNOL) & triage
AScore breakdown
Effort
B
ROI Floor
A
ROI Ceiling
A
Token Econ
A
Substance
A+
Overall
A
What it does
Captures first notice of loss and triages/schedules next steps. Via CCC, Allstate enables "over 80% of customers to book a repair appointment within one day," with cycle time "improved 14%."
Vendor-motive: a CCC platform sell, figures presuppose Allstate's claims flow through CCC's network, though the cycle-time win is concrete and operationally meaningful.
Sources Vendor
Minimum AI maturity required
2
Level 2
Operational
Scoped project. Needs production-grade data plumbing, governance, and a clear human-in-the-loop.
visibility
A
Industry · Industry Verticals
Computer-vision quality / defect inspection
Inspects output visually for defects.
Industry · Industry Verticals · Manufacturing
Computer-vision quality / defect inspection
AScore breakdown
Effort
C
ROI Floor
A
ROI Ceiling
A
Token Econ
A
Substance
A
Overall
A
What it does
Inspects output visually for defects. Unilever's Hefei factory saw "8% increase in overall equipment effectiveness and 20% wastage reduction."
Vendor-motive: genuinely impactful and platform-independent, Unilever's first-party factory results, product-agnostic.
Sources First-party
Minimum AI maturity required
3
Level 3
Advanced
Multi-quarter program. Needs MLOps maturity, evaluation harnesses, and cross-team ownership.
monitor_heart
C
Industry · Industry Verticals
Medical imaging analysis & diagnostics
Flags findings in scans and prioritizes urgent cases.
Industry · Industry Verticals · Healthcare
Medical imaging analysis & diagnostics
CScore breakdown
Effort
D
ROI Floor
B
ROI Ceiling
A+
Token Econ
C
Substance
A
Overall
C
What it does
Flags findings in scans and prioritizes urgent cases. Via Aidoc, University of Miami Health saw a "61.2% reduction in median TAT" for intracranial hemorrhage; a Sheba study showed "30% reduction in mortality" for intracerebral hemorrhage.
Vendor-motive: an Aidoc platform sell, but the mortality/turnaround figures come partly from a peer-reviewed study, which is far stronger evidence than a typical vendor case study, genuine clinical impact.
Sources Vendor
Minimum AI maturity required
4
Level 4
Transformational
Year-long program. Needs deep system integration, custom models, and executive sponsorship.