Cigna's AI Strategy: How a Top Global Health Insurer Built Conversational Care Navigation for 190M Members

TL;DR#

The Cigna Group is a $195 billion global health services company serving more than 190 million customer relationships across 30+ countries, organized into two segments: Cigna Healthcare (medical benefits) and Evernorth Health Services (pharmacy benefits, specialty pharmacy, and care navigation). In 2026, Cigna's most consequential AI investments are inside Evernorth — care navigation, behavioral health screening, and specialty pharmacy support — where the carrier is trying to move from claims-denial headlines to member-trust outcomes. The strategic problem is that the most valuable member moments are the most form-hostile: a parent quietly worried about a child's depression score, an expat trying to explain a chronic condition across languages, a member whose plan no longer fits the life event they haven't told anyone about yet. Dropdowns and PHQ-9 web forms collapse those moments; conversational AI can hold them. This is the Lemonade playbook applied to a payer five times the size and ten times the regulatory complexity — and it's where carriers like Cigna will either rebuild trust or watch members route around them.

Cigna at a glance: a $195B global health services company#

The Cigna Group reported $195.3 billion in revenue in fiscal year 2024, making it one of the five largest health insurers in the United States and one of the few US-headquartered payers with a serious international footprint. The company's most recent annual filings describe more than 190 million customer relationships worldwide across its two segments. Cigna Healthcare delivers medical, dental, behavioral, and supplemental benefits to roughly 19 million US medical customers and an additional global health benefits book covering expatriates and multinational employers in 30+ countries. Evernorth Health Services — the segment most US members don't know they're already using — includes Express Scripts (one of the three largest pharmacy benefit managers in the country), Accredo specialty pharmacy, and a fast-growing care services and care navigation business. Cigna's 2024 Annual Report frames Evernorth as the growth engine and the platform on which most of the company's AI roadmap is being built.

That structure matters for any analysis of Cigna AI strategy. Most public conversation about "Cigna AI" focuses on Cigna Healthcare — prior authorization, claims, member portals. The actual investment center is Evernorth, where pharmacy, behavioral health, and care navigation generate the highest member-touch volume and the deepest data flywheel.

Where Cigna is investing in AI: Evernorth Care, behavioral health, and specialty pharmacy#

Cigna's AI investments cluster in four areas, each tied to a member workflow where the cost of failure is high and the cost of getting it right is differentiated. The first is Evernorth Care navigation: routing members to the right care setting (virtual primary care, behavioral health, specialty pharmacy, in-person specialist) based on what the member is actually trying to solve, not the menu item they happened to click. The second is behavioral health screening and triage, which Evernorth has identified as one of the largest unmet-need categories in its book — and where conversational interfaces have a decisive advantage over forms. The third is specialty pharmacy adherence and education through Accredo, where members managing oncology, rheumatology, or HIV regimens need ongoing dialogue, not annual surveys. The fourth is prior authorization and utilization management, where Cigna — like every major payer — is under pressure from regulators, the press, and the ProPublica reporting on automated claim review to show that AI is being used to help members rather than route around them.

Across all four, the through-line is the same: the member moment is conversational, the legacy interface is a form, and the gap between the two is where trust either gets built or eroded. That gap is exactly what conversational AI in insurance is meant to close.

Why forms specifically fail for mental health intake#

Mental health intake is the single clearest case study in why conversational AI matters for a payer of Cigna's scale. The standard PHQ-9 depression screener is nine questions answered on a 0–3 scale, and on paper it takes two minutes. In practice, when surfaced as a web form to a member who is already in distress, completion rates collapse and the answers that do come back are flattened to whatever option is closest to the truth. A member who would tell a clinician "I've stopped sleeping but I don't think I'm depressed, I think it's my new job and my mother's diagnosis" will tick "several days" on item 3, ride past items 4–9, and abandon at the demographic page.

The American Psychiatric Association and clinical research repeatedly note that stigma and disclosure friction — not screener length — are the binding constraints on early identification of mental health conditions. Forms intensify both. A dropdown asks a member to label themselves before they've decided what's happening. A free-text field asks them to write a paragraph to a system that may or may not read it. A conversational interface — one that asks "what's been on your mind this week?" and follows up on the answer — does what a screener cannot: it lets the member describe the situation in their own words, then quietly maps that narrative back to the clinically validated instrument.

This is the exact gap Perspective AI's interviewer agent is built for in non-clinical research contexts, and it's the architectural pattern any payer doing serious behavioral health work has to adopt. The form is not the screener. The form is just one possible interface to the screener — and for sensitive intake, it's the worst one.

International members and multilingual conversational AI#

Cigna is one of the only major US-headquartered insurers with a real international book, and that footprint changes the AI math entirely. The Cigna Healthcare Global Health Benefits business covers expatriates, third-country nationals, and multinational employers in 30+ countries — a population that routinely needs to describe symptoms, navigate care, and resolve claims across language, currency, and care-system boundaries that no single benefit plan was designed to handle.

Forms scale badly across languages — every dropdown has to be translated, every conditional logic branch has to be re-tested, and every cultural assumption baked into the field labels (US ZIP codes, US insurance terminology, US date formats) breaks somewhere. Conversational AI scales differently. A voice or chat interface trained on the carrier's coverage rules can hold a conversation in Spanish, Portuguese, Mandarin, or Arabic without rebuilding the form layer for each. For a global payer that already has the operational footprint, this is one of the cleanest near-term ROI cases in the entire AI portfolio.

The parallel here is Lemonade's case study in conversational AI insurance, which built its growth on a single chatbot interface that scaled across product lines and geographies. Lemonade did it for a much smaller, much simpler book. Cigna has the harder version of the same problem — and the larger payoff.

What conversational AI unlocks beyond intake: longitudinal member research#

The intake conversation is the entry point, not the destination. The compounding value of conversational AI for a payer like Cigna is the longitudinal member research layer that sits on top of it. Every plan change, every life event, every dropped prescription, every escalation to behavioral health is a moment where a 20-second voice or chat check-in can capture context that no claims database, no NPS survey, and no annual member survey will ever see.

Three examples make this concrete. First, plan-change "why now" capture: when a member moves from a PPO to an HDHP at open enrollment, the carrier almost never learns whether the driver was premium, network, life event, or employer-mandate. A conversational check-in at the moment of change does. Second, specialty pharmacy non-fill follow-up: when an Accredo patient doesn't pick up a fill, the operational system flags it, but no one knows whether the reason is side effects, cost, transportation, or a missed call from the prescriber. A conversational reach-out gets that answer. Third, behavioral health no-show recovery: when a member misses a behavioral health appointment, a non-judgmental conversational follow-up converts a churn signal into a re-engagement opportunity. This is the same continuous-discovery pattern documented in the 2026 Continuous Discovery Report — applied to a population of 190 million.

This is why moves beyond NPS toward conversational sentiment measurement matter more for payers than for almost any other category. A health insurer's NPS score is famously low and famously useless; what the member actually thinks lives in the conversation that didn't happen.

The Lemonade lesson applied to a payer 5x the size#

Lemonade is a $1.5–2B market-cap insurer that built a category position on one thing: conversational AI as the primary member interface. Every quote, every claim, every policy change starts in a chat or voice conversation, and the company has publicly reported claim-resolution times measured in seconds for simple claims because the interface is conversational from the start. Lemonade did this with a roughly 2 million customer book and a single LOB focus (renters, homeowners, pet, term life).

Cigna is 5x Lemonade's size by customers in just the US medical book, 100x its size by revenue, and operates across medical, pharmacy, behavioral health, dental, supplemental, and global benefits. The complexity multiplier is real. But the architectural lesson is the same: the member experience is set by the entry-point interface, and forms set a ceiling on it that conversational AI does not. The carriers that figure this out first — and the early indications are that Evernorth Care is one of them — will build a member-trust moat that is very hard to copy from the back side.

The same pattern is showing up across the carrier landscape. State Farm's AI roadmap, USAA's AI customer service, Geico's AI chatbot strategy, Progressive's Snapshot and the conversational AI frontier, and UnitedHealth Group's AI strategy all share the same forward direction even though the products and member bases differ. The form is being demoted from primary interface to fallback, and the conversation is being promoted in its place.

How Perspective AI fits: conversational member research for payers#

Perspective AI is the research layer of this stack — not the clinical screener, not the prior auth engine, not the member portal, but the conversational interview tool that sits behind every one of them and answers the question payers cannot answer with claims data alone: why. Why did the member switch plans, why did they not fill the script, why did they no-show, why did they choose to call instead of message, why did they stop responding to outreach.

For payer research teams, member experience leaders, and Evernorth-style care services orgs, this is the layer that turns conversational AI from a customer service play into a strategic intelligence asset. Run a Jobs-to-be-Done interview on members who switched away at open enrollment. Run a churn interview on members whose specialty pharmacy fills lapsed. Run a win/loss interview on members who switched away at open enrollment. These are the conversations that won't happen on a form and won't show up in claims — and they are exactly the conversations a payer of Cigna's scale needs running continuously, not annually.

If you lead member experience, behavioral health product, or care navigation at a payer, the practical first step is small: pick one segment, run one continuous conversational study, and compare what you learn to what your current survey instrument reports. The gap is usually the whole point.

Frequently Asked Questions#

What is Cigna's AI strategy in 2026?#

Cigna's AI strategy in 2026 centers on Evernorth Health Services — particularly care navigation, behavioral health screening, specialty pharmacy adherence through Accredo, and utilization management — rather than on Cigna Healthcare's claims layer alone. The strategic bet is that conversational AI interfaces can replace form-based intake for sensitive member moments (mental health screening, plan changes, non-fill follow-up) and produce both better clinical outcomes and a longitudinal member research signal that claims data cannot generate on its own.

What is Evernorth Care AI?#

Evernorth Care AI refers to the suite of AI-driven care navigation, behavioral health, and specialty pharmacy capabilities housed inside Cigna's Evernorth Health Services segment. It includes member-facing conversational tools that route to the right care setting, adherence support for high-touch specialty drugs through Accredo, and behavioral health triage. Evernorth is the growth engine of The Cigna Group and the platform on which most of the company's member-facing AI is being built.

How many members does Cigna have globally?#

The Cigna Group reports more than 190 million customer relationships worldwide across its Cigna Healthcare and Evernorth Health Services segments. Inside that total, Cigna Healthcare serves roughly 19 million US medical customers and an additional global health benefits book covering expatriates and multinational employers in 30+ countries. Evernorth's customer count is larger and overlaps significantly with employer-sponsored medical members and external PBM clients.

Why is conversational AI a better fit for mental health intake than forms?#

Conversational AI is a better fit for mental health intake because the binding constraint on early identification is stigma and disclosure friction, not screener length. Forms ask members to label themselves before they've decided what's happening, flatten "it depends" answers into dropdowns, and abandon members at the demographic page. A conversational interface lets the member describe the situation in their own words and then quietly maps the narrative back to a clinically validated instrument like the PHQ-9, capturing both the score and the context.

How is Cigna using AI for international members?#

Cigna uses AI to support its Global Health Benefits population — expatriates, third-country nationals, and multinational employers in 30+ countries — by deploying conversational interfaces that work across languages and care systems. Voice and chat interfaces trained on the carrier's coverage rules can hold conversations in multiple languages without rebuilding form layers for each market, which is one of the cleanest near-term ROI cases in the company's AI portfolio.

What can payers learn from Lemonade's conversational AI playbook?#

Payers can learn from Lemonade that the member experience is set by the entry-point interface and that forms set a ceiling on it that conversational AI does not. Lemonade built its category position on a chat-and-voice-first interface for quoting and claims, achieving resolution times measured in seconds for simple claims. The architectural lesson applies even at 5–10x Lemonade's complexity — the carriers that move first to conversational intake build a member-trust moat that is very hard to copy from the back side.

Conclusion: conversational AI is how Cigna rebuilds member trust#

The Cigna AI strategy that will matter in 2026 isn't the prior auth headline or the claims-engine debate. It's the Evernorth Care wager that conversational AI — applied to mental health intake, care navigation, specialty pharmacy adherence, and international member support — produces both better clinical outcomes and a member-trust signal that 190 million people will feel before they ever see it in an NPS score. Forms cannot do that work. Dropdowns cannot hold "it depends." The carriers building the next decade of health insurance AI are the ones letting members speak in their own words first and structuring the data second.

Perspective AI is the research layer behind that shift. If you're building member research, behavioral health intake, care navigation, or onboarding programs at a payer — and you want to see what your members are actually saying instead of what your survey field forced them to choose — start a research study, browse use cases, or compare alternatives to see how conversational member research fits into a modern health insurance AI stack.