Upportunist

Turning emerging AI capabilities into practical outcomes.

Healthcare Insurance Member Organizations

Organizations do not need more AI pilots. They need better decisions about where AI creates measurable value – and the operational discipline to capture it. Upportunist synthesizes the evidence and frames what it means for organizations navigating this moment.

40% faster task completion with AI writing assistance Noy & Zhang, Science 2023
34% productivity gain for novice workers vs. minimal gain for experts Brynjolfsson et al., QJE 2025
46% reduction in ICD coding time with AI assistance (RCT) Chomutare et al., JMIR 2025
~30 min daily documentation time saved per physician with ambient AI scribes Lukac et al., NEJM AI 2025
84% of health insurers use AI or ML in some operational capacity NAIC Survey 2025
$1T+ estimated annual U.S. healthcare administrative cost JAMA, 2024
55.8% faster coding task completion with GitHub Copilot Peng et al., arXiv 2023
80% of healthcare data is unstructured and largely inaccessible to analytics IBM Institute for Business Value
12.2% task completion increase for BCG consultants using GPT-4 Dell'Acqua et al., Org. Science 2025
40.4% pooled EHR-related burnout prevalence across 37 studies (N=66,556) Meta-analysis, CJHT 2024
40% faster task completion with AI writing assistance Noy & Zhang, Science 2023
34% productivity gain for novice workers vs. minimal gain for experts Brynjolfsson et al., QJE 2025
46% reduction in ICD coding time with AI assistance (RCT) Chomutare et al., JMIR 2025
~30 min daily documentation time saved per physician with ambient AI scribes Lukac et al., NEJM AI 2025
84% of health insurers use AI or ML in some operational capacity NAIC Survey 2025
$1T+ estimated annual U.S. healthcare administrative cost JAMA, 2024
55.8% faster coding task completion with GitHub Copilot Peng et al., arXiv 2023
80% of healthcare data is unstructured and largely inaccessible to analytics IBM Institute for Business Value
12.2% task completion increase for BCG consultants using GPT-4 Dell'Acqua et al., Org. Science 2025
40.4% pooled EHR-related burnout prevalence across 37 studies (N=66,556) Meta-analysis, CJHT 2024
Upportunist Analysis

Featured Brief

All research & insights
Administrative Burden

Administrative work is among the strongest near-term AI use cases – and the opportunity extends well beyond documentation assistance to billing, coding, and member communications.

May 2025 · 8 min Read
Workforce

Less experienced workers consistently see the largest proportional productivity gains – with meaningful implications for healthcare contact centers, operations, and onboarding economics.

May 2025 · 7 min Read
Data & Analytics

80% of healthcare data is unstructured. AI may represent the first practical means of converting that information into operational insight at scale.

April 2025 · 8 min Read
Governance

Organizations that treat governance as a strategic capability – rather than a compliance burden – tend to adopt AI more effectively and at greater scale.

April 2025 · 9 min Read
Operationalization

The gap is no longer technology access – it is organizational capacity to move from experimentation to sustained value. Research identifies what separates organizations that scale from those that stall.

March 2025 · 10 min Read
Research Library
149 items across peer-reviewed literature, policy analysis, and original research.
From Peer-Reviewed Research

Key Evidence Across Focus Areas

Full research library
Workforce
14%

Average productivity gain for customer service agents with AI assistance – rising to 34% for the lowest-skill quartile.

Brynjolfsson, Li & Raymond · Quarterly Journal of Economics · 2025
Administrative
46%

Reduction in ICD coding time when clinicians used an AI coding assistant – in a crossover randomized controlled trial.

Chomutare et al. · Journal of Medical Internet Research · 2025
Data & Analytics
4.0%

Reserve estimation error after LLMs extracted 36 actuarial variables from unstructured claims documents, down from 6.5%.

Lieberthal et al. · arXiv · 2026
Insurance
56%

Of health insurers use AI for utilization management; 44% for claims adjudication; 37% for prior authorization.

NAIC Health Insurance AI/ML Survey · 2025 (n=93 insurers)
Focus Areas

What We Study

01

Workforce Productivity

How AI augments human performance across clinical, operational, and administrative roles – and where the evidence is strongest for less-experienced workers.

28 items in library →
02

Administrative Burden Reduction

Documenting the cost of administrative work – documentation, coding, prior auth – and examining AI's potential to reallocate time toward higher-value activities.

34 items in library →
03

Insurance Operations & Claims

AI applications in claims adjudication, fraud detection, prior authorization, utilization management, and member services across health plans and insurers.

56 items in library →
04

AI Governance & Adoption

Organizational frameworks for responsible AI deployment – oversight structures, accountability, fairness, and the conditions that enable adoption at scale.

62 items in library →
05

Unstructured Data & LLMs

How AI unlocks institutional knowledge embedded in clinical notes, claims records, communications, and documentation that structured systems cannot access.

44 items in library →
06

Clinical Documentation

Ambient AI scribes, EHR burden, and the evidence on documentation time, clinician burnout, and note quality from randomized trials and cohort studies.

38 items in library →
Recently Added to the Library

From the Peer-Reviewed Literature

2025
Clinical Documentation
Lukac et al. · NEJM AI · DOI: 10.1056/AIoa2501000
3-arm RCT (n=238 physicians, 14 specialties) found AI scribes reduced documentation by ~30 min/day per provider with clinically meaningful burnout reductions.
2025
Workforce
Dell'Acqua, Mollick et al. · Organization Science · 2025
Pre-registered RCT with 758 BCG consultants: GPT-4 raised completion 12.2%, cut time 25%, improved quality 40% – but worsened performance on tasks outside AI's capability frontier.
2025
Insurance
NAIC · 93 insurers, 16 states · 2025
84% of health insurers use AI operationally. Top barriers: skills and resources (52%), data challenges (40%). 92% aligned with NAIC governance principles.
2026
Governance
Howard et al. · npj Digital Medicine · 2026
Systematic review of 35 frameworks produced HAIRA – a five-level governance maturity model with seven critical domains, providing organizations a scalable pathway to accountable AI deployment.
Browse all 149 items
Evidence Snapshot

The Administrative Burden Case

34%

of U.S. healthcare spending goes to administration
Himmelstein et al., JAMA 2019

2–4×

hours of admin for every hour of patient care
Sinsky et al., Ann. Intern. Med. 2016

AI-Assisted Time Reductions – Selected Evidence
ICD Coding
46%
Writing Tasks
40%
Ambient Scribes
~30 min/day
Dev Coding
56%
Customer Support
25%
Per-task time reductions from peer-reviewed RCTs and field experiments. Bars show relative magnitude only; absolute savings vary by implementation context.
Research Perspective
"The challenge is no longer whether AI works. The challenge is where it works, how it is governed, and whether organizations have built the operational capacity to move from experimentation to durable value."

Upportunist Research Synthesis · 2025

About Upportunist

An Applied Research Organization

Upportunist studies how organizations identify, evaluate, and operationalize emerging technologies. Our focus is practical adoption, measurable outcomes, and responsible implementation at the intersection of technology, operations, and organizational change.

Our work is grounded in peer-reviewed research, operational evidence, and the practical realities facing healthcare, insurance, and member-driven organizations. The research library covers 149 items across 13 topic categories.

About the organization Get in touch
Research Library at a Glance
149
Total items
111
Peer-reviewed articles
13
Topic categories
6
Upportunist briefs
Browse the research library →