Incedo IntelliPatient AI-native adherence risk detection and personalized intervention

IntelliPatient connects the signals across a patient support program into a live, explainable view of who is at risk, why, and what action can be taken next.
Speak to an expert
See it from portfolio to patient
Problem Where it shows up Archetypes Solution Capabilities Modules Why IntelliPatient Guardrails Benefits Security Resources FAQ

Problem

Every day, patients decide whether or not to continue treatment.

The signals that reveal that decision reach the people who could support them far too late. A prior authorization taking too long, a failed shipment, engagement dropping away: the signal was there for weeks, but was sitting somewhere a case manager could not get to in time.

$5B+
spent annually on patient support programs

Source: Fierce Pharma survey, cited in Incedo whitepaper (May 2026)

3%
of patients actively use available programs

Source: Fierce Pharma survey, cited in Incedo whitepaper (May 2026)

One patient, twelve weeks. Every signal was there; the case manager saw a list at week eleven.

StartWeek 2Week 4Week 6Week 8Week 10Week 12HUBPA renewal submittedHUBRenewal stalled, no decisionSPECIALTY PHARMACYFirst refill missedENGAGEMENT PLATFORMPortal logins stopENGAGEMENT PLATFORMMessages unopenedSPECIALTY PHARMACYSecond refill missedCase manager gets the listPatient stops therapy

Where it shows up

The signal exists. Nobody sees it until the patient has decided to stop.

01

Warning signs arrive too late

By the time a case manager receives a patient list, the deterioration has already accumulated. For example, a patient misses two refills, stops opening messages, and has decided to stop treatment before anyone sees it.

02

Teams see segments, not the individual

Population visibility stops short of individual intelligence. For instance, an analytics dashboard says forty patients in a cohort are at risk, but not which one to call this morning, or why you are calling them.

03

The same patient looks different in every system

No single system holds the whole picture. Pharmacy data shows a gap, the hub shows a stalled authorization, the engagement platform shows silence, and nobody is looking at all three.

04

Access barriers surface after the patient has stopped

A prior authorization sits unanswered for three weeks; nobody notices until the patient stops therapy. These administrative gaps carry heavy clinical consequences, yet they are much simpler to resolve than medical or behavioral issues when you have operational visibility.

05

Interventions cannot be cleanly attributed to outcomes

Programs cannot attribute outcomes to interventions. Outreach, escalations and nurse calls run across thousands of patients, and the budget often gets allocated by convention rather than evidence.

Archetypes

Same high-risk label. Three completely different interventions.

Adherence risk looks the same on a dashboard and means something completely different in reality. A missed refill on a weekly injection is not a missed refill on a quarterly infusion, so patients are grouped by how they engage with therapy, not by drug.

The patient who feels well The patient before routine The family navigating diagnosis
Therapy type Lifelong, self-administered Finite, fully self-managed High-touch, caregiver-led
Engagement window Indefinite Variable, tied to response Long, often pre-diagnosis
Patient agency Moderate. Manages at home. Full. No site involvement. Low. Caregiver coordinates.
Critical gap Drop-off once they feel stable. Early, before the regimen sticks. Diagnostic delay, not adherence.
What a missed refill means Possible disengagement Possible progression Possible coordination failure

A one-size-fits-all risk score can’t tell them apart. IntelliPatient’s AI Reasoning Engine, configured to each archetype, can. Each archetype carries its own definition of risk, its own drivers and its own interventions, so a missed refill is read against that patient’s therapy, not a portfolio average.

It works from day one, with or without your own predictive model.

Solution

One solution, three altitudes.

IntelliPatient is an AI-native adherence risk detection and intervention solution for life sciences patient support programs.

  • It watches the signals programs already produce and tells you which patients are drifting, why, and what will bring them back.
  • It sits across specialty pharmacy, access services, engagement channels and case management, not replacing any of them, but seeing what none of them can see alone.

A coverage gap on its own means little. A coverage gap alongside a denied authorization and a patient who has stopped answering means something specific, and it calls for a different intervention than a patient who simply forgot.

Portfolio level

Digital Health and Patient Experience leadership

Needs to know where the patient journey is breaking down, and why.

Population level

Patient Support Program teams

Need to know which cohorts are at risk, and which need outreach versus a case manager.

Patient level

Case managers

Need to know which patient to call this morning, what the call is about, and how to assist the patient.

Capabilities

What IntelliPatient enables.

1. Understand

See the whole patient, and why they are at risk

A full patient 360 view

  • Signals from specialty pharmacy, access services, CRM and engagement channels in one place.
  • Nobody assembles the picture from multiple systems.

Patient Context Engine

  • The domain vocabulary and ontologies of patient support, built in: what each signal means and what it implies for risk and intervention.
  • Answering whether a patient is missing refills takes more than dates; it takes knowing that patient’s refill interval.

A patient risk profile you can understand and interrogate

  • Every risk assessment comes with a ranked, plain-language account of what is driving it.

A 30-day gap means nothing until you know the therapy. On a quarterly infusion it is routine; on a weekly injection it is four missed doses.

A financial barrier needs a different response from an administrative one, and both differ from a patient who has simply disengaged.

2. Act

Act on risk, one patient or one cohort at a time

Agentic omnichannel intervention matched to the risk driver

  • IntelliPatient matches the intervention to the driver, then delivers it on the channel that patient has consented to: SMS, email or portal.
  • Escalates to a case manager where digital outreach isn’t enough.
  • All of this at the patient population level, with a human in the loop.

Population-level action where appropriate

  • Cohorts grouped by why they are at risk, with approved digital outreach for the segments where it works and escalation for the ones needing a qualified case manager.

A caseload worked by urgency

  • A prioritized list of who needs attention this morning, why, what has already been tried, and what to do next.

3. Run

Run the program on evidence, inside your stack

Portfolio-level diagnosis

  • Which patient journey stage is underperforming and what is driving it, rather than a quarterly report assembled by hand.
  • So that resources can be deployed in a timely manner.

Guardrails built in, not bolted on

  • Content from the approved library only, human confirmation before anything reaches a patient, everything logged.

The systems you already run, complemented

  • IntelliPatient reads from your systems and writes confirmed actions back.
  • Where a program has its own risk model, IntelliPatient uses it; where it doesn’t, the AI Reasoning Engine assesses risk from the same signals.

A patient can be high risk and still not contactable.

Modules

Three modules over one engine, from the portfolio down to one patient.

The same context and reasoning layers serve the executive reading the portfolio, the program team acting on a population, and the case manager working one patient.

1
Patient Journey Monitor

Portfolio level
2
Patient Intelligence Hub

Portfolio level
3
Case Manager Advisor

Portfolio level

Lia, an AI Assistant   across the modules

Patient Context Engine –  the foundation beneath everything

1 · Portfolio level

Patient Journey Monitor

A portfolio view for digital health and patient services leadership. It reasons across brands and stages to show where the patient journey is breaking down, the root cause behind it, and which patient segments are most affected, whether by diagnostic delays, access barriers or disengagement on therapy.

Patient Journey Monitor

Brand: RespireX · Q2 2026 · Illustrative

HealthyWarningTroubled

Stage 1

General Wellness

% of patients deeply engaged with disease content

38%

▲ 3 pts

vs. May 2026

Stage 2

Symptom Recognition

% of patients showing care-seeking intent

29%

▼ 4 pts

vs. target 33%

Stage 3

Diagnosis & Care Coordination

% of patients with a confirmed diagnosis

74%

▲ 1 pt

vs. target 70%

Stage 4

Treatment Initiation

Treatment initiation rate

58%

▼ 11 pts

vs. Q1 2026 (69%)

Investigate

Stage 5

Ongoing Management

Non-persistence rate

67%

▲ 8 pts

vs. TA benchmark ~45%

Investigate

Root cause · AI summary

Treatment initiation fell from 69% to 58% this quarter. The drop traces to two commercial payers that added a step-edit requirement in January. Time to first dose rose from 15 to 21 days, with delays concentrated in patients on those plans.

Suggested next step: re-verify benefits for affected patients before their first fill.

A population layer for manufacturer-side patient support teams. It groups patients by risk level and the driver behind it, with a root-cause explanation and recommended intervention per segment. Teams deploy approved outreach or escalate to case managers with the brief already assembled: agentic orchestration across channels, with a human confirming anything that reaches a patient.

2 · Population level

Patient Intelligence Hub

Signals Across the Program
Specialty pharmacy
Hub services
Engagement platform
CRM and case notes
One enrolled panel. 200 patients.
Patient Intelligence
Hub
Patient 360
Context Engine
AI Reasoning Engine
Agentic orchestration
Risk level · Driver
High · 12 Patients
Escalate to a person
Administrative case manager or nurse educator
Moderate · 61 Patients
Approved digital outreach
On the channel each patient consented to
Early · 34 Patients
Lightest possible touch
One message, one consented channel
Twelve patients need a person. Ninety-five need something else.
Case managers are the scarcest resource a program has. Illustrative panel.

3 · Patient level

Case Manager Advisor

An individual layer for case managers. A prioritized worklist, each patient’s risk explained in plain language, the next action with alternatives, and the history of what has already been tried. Structured outcome capture closes the loop.

What arrives with each escalation The signal that fired, the evidence behind it, and everything already tried.

Today's work

9 patients still need action
RespireX Asthma · direct-to-patient
Critical 4 patients need action now
71
Critical
SC
Sarah Chen
Prior authorisation unanswered for 9 days
Appeal · 5d
60
Critical
CF
Christopher Alexander Fitzgerald III Start Here
Risk rising. Adherence slipping and automated outreach unanswered
52
Critical
AP
Angela Park
Returned from suppression with risk unimproved. Three nudges unanswered
Non-responder
1 more critical patient

Across the modules

Lia, an AI Assistant

The conversational assistant embedded across Patient Intelligence Hub, Case Manager Advisor, and eventually Patient Journey Monitor. Ask why a risk segment surfaced, which patients within it are at highest risk, or why an individual’s risk changed. Answers are grounded in the program’s own data and definitions, with figures traceable to source. Where the data, definitions or permissions do not support an answer, Lia says so.

Ask Lia
your caseload
Illustrative
Who is most at risk right now?
Answer
Sarah Chen needs attention first — Critical, risk 71, 9 days since last touch. Recommended: Escalate prior authorisation (needs your confirmation). After that: Angela Park, Maria del Carmen Rodríguez-Villanueva, Christopher Alexander Fitzgerald III.
Ask Lia
your caseload
Illustrative
Why is the copay segment growing?
Answer
Eleven patients moved in this week. Nine are on the same payer, and a step-edit took effect on the 1st. The other two had copay cards lapse on renewal.
Ask Lia
your caseload
Illustrative
What is the best way to reach patients who have gone silent?
Answer
For patients silent across multiple automated nudges — Sarah Chen (9d), Angela Park (14d), Lisa Ramirez (7d) — a direct personal call or SMS from you outperforms another automated nudge. Each is yours to confirm.

The foundation

Patient Context Engine

The foundation that powers everything else. It holds the domain vocabulary and ontologies of patient support: what each signal means, how signals relate, and what they imply for risk and intervention. Without it, something like a coverage gap is just a number of days; with it, it is a risk signal that requires a specific intervention. Every figure, insight and recommendation across the three modules resolves against it.

Domain

Patient support knowledge, built in

Enterprise

Your program’s own setup

Session

Who’s asking, and what they can see

Every figure, insight and recommendation traces back to its underlying source records.

Why IntelliPatient

Built for how patient support actually works.

It looks completely different across therapies and patient archetypes.

A patient on a lifelong biologic who has started feeling well and has completely disengaged, a patient three months into an oral therapy before the routine has formed, a family navigating a rare disease diagnosis: the same high-risk label, three different problems. A single model across a portfolio cannot tell them apart.

IntelliPatient runs one engine that produces a different model for each therapy and archetype. Risk is defined for that therapy, the drivers are specific to it, and the interventions come from what works there. Configured to each rather than rebuilt, and working from day one with or without a predictive model of your own.

It reads from the systems a program already runs: specialty pharmacy, patient access services, CRM, case management, and patient channels.

It doesn’t replace any of them. Access platforms still submit the authorizations, engagement platforms still send the messages, analytics still report on the quarter. IntelliPatient is what notices that authorization has been sitting for three weeks while the patient stopped answering texts.

Which signals matter, where the thresholds sit, what each risk driver leads to, and which interventions are available: all configured per program and per therapy.

Whether risk comes from your own model or from IntelliPatient’s AI Reasoning Engine, the same solution serves a lifelong biologic and a three-month oral therapy without being rebuilt for either.

Any system reasoning about adherence must know, for instance, what a coverage gap means, and why a copay lapse differs from a PA denial.

A horizontal AI layer can be configured toward this, but doesn’t come with this knowledge baked in, and every program would have to build it from scratch. IntelliPatient ships with this patient support domain already encoded. You simply configure the solution to your models, business rules, intervention library, and approved content.

Guardrails

Trust by design: every action and every answer stays inside defined boundaries

Human confirmation and consent

  • Every patient-facing action is confirmed by a qualified patient support professional.
  • Content comes only from the program’s MLR-approved templates, with approval checked at dispatch.
  • Opt-outs and consent apply regardless of risk: a high-risk patient may still not be contactable.

Scoped access, full audit trail

  • Case managers see their panel, program leads their program, leadership the full portfolio.
  • Out-of-scope requests get an explicit refusal, not an empty result.
  • Every recommendation, approval, override and intervention is logged and attributable, refusals included.

Grounded answers, or none

  • Every figure resolves to a source record.
  • Where the data, definitions or permissions don’t support an answer, IntelliPatient says so.
  • For programs that must explain an intervention to compliance, this is the precondition, not a feature.

Benefits

Keep more patients on therapy.

IntelliPatient helps patient support programs act on risk earlier and direct effort where it still makes a difference.

01

Better therapy adherence

More patients stay on therapy because risk surfaces while there’s still time to act on it.

Signals are read as they arrive, not at the next list refresh.

Prior authorization delays and copay lapses are caught as leading indicators, not after the patient has gone.

02

Better patient outcomes

Patients get the intervention that fits their barrier, not a generic nudge.

A financial barrier gets administrative action, an anxious patient gets a nurse educator, and a forgotten refill gets a reminder.

03

Patient support process effectiveness

Scarce case manager capacity goes where intervention still works, not to whoever appears on a list.

Interventions are captured against the driver they addressed, so a program can see which ones work and fund accordingly.

04

Portfolio visibility

Where the journey breaks down is visible in weeks, not in a quarterly report another team must assemble.

Security

Incedo’s security and compliance posture.

ISO 27001:2022 Certified

The globally recognized benchmark for information security management, validated through independent third-party onsite audits conducted annually.

SOC 2 Type 2 In progress

Building on the ISO 27001:2022 foundation, subjecting operational controls to independent audit validation across the Trust Service Criteria.

Resources

Go deeper.

thumbnail-beyond-the-portal
  • Whitepaper

Beyond the Portal

the next era of patient centric pharma is powered by AI thumb
  • Whitepaper

The next era of Patient-Centric pharma is powered by AI

Built with our technology partners

AWS

OpenAI

Databricks

Google Cloud

Salesforce.com logoA cloud computing company based in San Francisco, California, United States

Salesforce

FAQ

Frequently asked questions.

What is IntelliPatient?

IntelliPatient is an AI-native adherence risk detection and intervention solution for life sciences patient support programs. It connects signals from dispensing, patient access and benefits, engagement, case management, and other available program data into a connected view of which patients may be at risk of coming off therapy, why they may be at risk, and what action can be taken. It works across three levels: portfolio, population, and individual patient.

Does IntelliPatient replace our existing patient support platform?

No. IntelliPatient works with the systems already running the patient support program. It is not a patient call center, specialty pharmacy platform, or CRM, and it does not require organizations to replace those systems.

We already work with a hub services provider. How does IntelliPatient fit alongside them?

IntelliPatient integrates with the existing hub ecosystem rather than competing with it. The hub continues to manage activities such as enrollment, benefits investigation, prior authorization, and case management. IntelliPatient connects the signals generated by those activities with specialty pharmacy, engagement, and other available data to provide risk intelligence and recommended actions.

Do we need our own predictive model to use IntelliPatient?

No. Where a program already has a predictive risk model, IntelliPatient can consume its output. Where one does not exist, the solution can derive risk levels from available signals with its AI Reasoning Engine. Programs can also evolve their risk models over time.

Does anything reach a patient automatically?

No. Every patient-facing action requires confirmation by a qualified patient support professional before it is sent. Patient-facing content is sourced from the MLR-approved library, and actions outside approved parameters are not permitted.

How does IntelliPatient handle MLR compliance for patient outreach?

Patient-facing content is drawn from the program’s approved content library, with substitution permitted only at approved points. Access is scoped by role and enforced before data reaches the intelligence layer. Recommendations, approvals, overrides, and outreach actions are logged to an audit trail.

What data does IntelliPatient need?

At minimum, the platform can work with dispensing or shipment data with days supply, patient access and benefits status, and engagement activity. Claims data can further strengthen the picture where available. IntelliPatient is designed to work with the data a program has rather than requiring a complete data set.

How long does implementation take?

Implementation depends on the number of source systems involved, the condition of the underlying data, and the scope of the use case. IntelliPatient is configured around the program’s own definitions, including how adherence and risk are measured and which interventions are approved, rather than imposing a standard model. Implementation scope is determined during the initial engagement.

How is IntelliPatient different from the adherence analytics we already have?

Adherence analytics identify at-risk cohorts and report on what already happened. They name the segment, not the patient.

IntelliPatient reasons across the same signals and goes further: it identifies the individual patient, explains what is driving their risk, recommends the intervention that fits that driver, and routes it to the person who can act while there is still time. Analytics tell a program what happened. IntelliPatient tells it who to help, why, and how.

What therapy areas does IntelliPatient support?

Adherence analytics identify at-risk cohorts and report on what already happened. They name the segment, not the patient.

IntelliPatient reasons across the same signals and goes further: it identifies the individual patient, explains what is driving their risk, recommends the intervention that fits that driver, and routes it to the person who can act while there is still time. Analytics tell a program what happened. IntelliPatient tells it who to help, why, and how.

How does IntelliPatient handle patient data?

IntelliPatient handles patient data within applicable HIPAA and PHI-sharing constraints, with role-based access controls and an audit trail across the platform. Incedo is prepared to enter into appropriate contractual arrangements, including Data Processing Agreements (DPAs) and, where applicable, Business Associate Agreements (BAAs), subject to review and agreement on the specific scope of services, data processing activities, and contractual terms.

Are your patients coming off therapy for reasons nobody saw in time?

See it against your own program. Talk to an expert about what IntelliPatient would surface.

Speak to an expert

Talk to Our Experts Now

Incedo Page – Download Now

Incedo Case Study – Download Now