Dan Liljenquist, Intermountain AI Appeal Agents

Dan Liljenquist, Intermountain AI Appeal Agents •

Dan Liljenquist, Senior Vice President and Chief Strategy Officer at Intermountain Health, has highlighted AI-powered insurance appeal letters as one of Intermountain’s practical, near-term applications of artificial intelligence.

1. Who is Dan Liljenquist?

Dan Liljenquist is Intermountain Health’s chief strategy officer and has held the role since 2018. His responsibilities include helping establish and execute the health system’s strategic priorities.

His AI strategy is part of a broader effort at Intermountain to use technology to:

  • Reduce administrative workload
  • Address workforce shortages
  • Lower healthcare costs
  • Improve operational efficiency
  • Give clinicians and staff more time for patient-focused work
  • Scale useful AI applications across the organization

By late 2025, Liljenquist said Intermountain had about 300 AI projects underway.

2. What problem are the AI appeal letters solving?

Insurance companies sometimes deny claims for medical services. When that happens, healthcare organizations can submit an appeal explaining why the treatment was medically appropriate and why the payer should reconsider the denial.

Preparing these appeals can be highly labor-intensive.

Traditionally, staff may have to:

  1. Open the patient’s medical record.
  2. Review clinical notes.
  3. Locate relevant diagnoses.
  4. Find laboratory and imaging results.
  5. Identify previous treatments.
  6. Determine why the treatment was medically necessary.
  7. Gather supporting dates and documentation.
  8. Organize the information.
  9. Draft an appeal letter.
  10. Review the letter before submitting it to the payer.

The difficulty is that the relevant information can be spread across a large electronic medical record.

3. How Intermountain is using AI

Liljenquist described an AI agent that can work through medical-record information and identify the information needed to prepare payer appeals.

The key capability is essentially information retrieval and synthesis: rather than having an employee manually search through the record, AI helps scrape through the medical record and find the relevant information needed for the appeal. Liljenquist said this application saves approximately 30 minutes per appeal letter.

This is important because the AI isn’t being presented primarily as a diagnostic tool. Instead, it is being used for a high-volume administrative workflow.

Simplified workflow

Insurance denial
↓
An appeal is required
↓
AI searches the medical record
↓
Relevant clinical information identified
↓
Information organized for the appeal
↓
Appeal letter prepared
↓
Human staff review/submission

The exact technical architecture and level of autonomous decision-making have not been publicly detailed in the sources I found, so it would be inaccurate to characterize the system as independently making medical-necessity decisions.

4. The 30-minute time saving

The most notable figure associated with the project is about 30 minutes saved per payer appeal letter.

Liljenquist specifically described this as one example among “dozens and dozens” of AI applications being used to simplify work at Intermountain.

The significance becomes much larger when multiplied across a large healthcare system.

For example:

  • 10 appeals/day × 30 minutes = 5 hours saved/day
  • 100 appeals/day × 30 minutes = 50 hours saved/day
  • 1,000 appeals/day × 30 minutes = 500 hours saved/day

These are illustrative calculations—not reported Intermountain volumes.

The important point is that even a relatively small saving per transaction can become substantial when applied to a high-volume administrative process.

5. Why appeal letters are a good AI use case

Appeal preparation has several characteristics that make it suitable for AI assistance:

High volume

Healthcare organizations process large numbers of claims and denials.

Information-heavy

The evidence needed for an appeal can be distributed throughout a patient’s record.

Repetitive

Employees repeatedly perform similar searches and documentation tasks.

Rules and documentation matter

Appeals generally require specific supporting information rather than completely open-ended creative work.

Human review remains valuable

The AI can help find and organize information while people can review the resulting documentation for accuracy and appropriateness.

6. AI is much broader than appeal letters at Intermountain

The appeal-letter application is only one example.

In a 2025 discussion, Liljenquist described Intermountain as having approximately 300 AI projects, covering numerous areas of the organization.

He has emphasized that AI can help address two major healthcare problems simultaneously:

  • Increasing administrative complexity
  • A tightening healthcare labor market

Intermountain sees AI as a way to make existing employees more productive rather than simply adding more labor to increasingly complex processes.

7. The importance of Intermountain’s data

One reason Liljenquist believes established health systems have an opportunity in AI is that they already possess large quantities of healthcare data.

He has noted that much of the data needed to train and operate AI systems is already inside health-system environments.

For an appeal-letter application, that is particularly relevant because the information needed to support an appeal is already contained in the patient’s electronic medical record.

The challenge is not necessarily collecting new data; it is finding the right information quickly and using it appropriately.

8. Cloud infrastructure also matters

Liljenquist has connected Intermountain’s AI progress with its broader technology infrastructure.

He explained that moving data into the cloud and building on existing platforms has helped the organization integrate AI and other tools more quickly.

The appeal-letter example is therefore not an isolated experiment. It illustrates how infrastructure can enable practical AI applications across an integrated health system.

9. AI governance is critical

Healthcare AI has obvious risks because it operates around highly sensitive clinical information.

Liljenquist has discussed AI governance and oversight as part of Intermountain’s broader AI strategy.

For appeal letter automation, important safeguards would include:

  • Protecting patient privacy
  • Restricting access to authorized users
  • Ensuring retrieved clinical information is accurate
  • Avoiding fabricated information
  • Maintaining appropriate human review
  • Checking that the final appeal accurately reflects the medical record
  • Maintaining auditability of AI-assisted work
  • Following applicable healthcare and payer requirements

The public sources do not provide enough detail to claim exactly which safeguards are implemented in this particular appeal-letter workflow.

10. Why this matters strategically

The biggest lesson from Liljenquist’s example is that healthcare AI does not have to begin with futuristic clinical applications.

A health system can generate meaningful value by automating administrative friction.

The appeal-letter example demonstrates a straightforward equation:

Existing clinical data + AI retrieval/synthesis + human oversight = less administrative work

And the result is measurable: roughly 30 minutes saved per appeal letter, according to Liljenquist.

That matters because administrative work consumes substantial resources across healthcare. AI that removes repetitive work can potentially allow employees to spend more time on higher-value activities.

11. What makes Intermountain’s approach different?

Intermountain

Liljenquist has described Intermountain’s approach as pragmatic. Rather than treating AI as a single massive transformation project, the organization is pursuing many applications that can produce practical benefits.

The appeal-letter use case illustrates this philosophy particularly well:

AreaTraditional approachAI-assisted approach
Record reviewManual searchingAI-assisted retrieval
Information gatheringStaff locate informationAI identifies relevant information
Appeal preparationLabor-intensiveAI assists drafting/preparation
Staff timeHigherLower
ScaleLimited by available staffGreater potential scalability
Final responsibilityHumanHuman oversight remains important

12. The bigger healthcare lesson

Liljenquist’s message is essentially that AI should solve real operational problems.

Intermountain isn’t using AI simply because it is an emerging technology. It is looking for areas where AI can make healthcare work:

  • Faster
  • Less expensive
  • Less repetitive
  • Easier for employees
  • More scalable
  • More sustainable as labor pressures increase

The payer-appeal example is particularly compelling because the benefit can be quantified immediately: about 30 minutes of administrative work eliminated from each appeal letter.

Bottom line

Dan Liljenquist’s Intermountain Health AI appeal-letter initiative is a practical example of agentic/generative AI being applied to healthcare administration rather than directly to clinical diagnosis. The AI helps search and synthesize information from medical records so staff can prepare insurance appeal letters more efficiently. Liljenquist has said the technology cuts approximately 30 minutes from each payer appeal letter, and he presents it as one of many AI applications being developed across Intermountain.

The broader strategy is significant: use AI to remove repetitive administrative work, leverage the health system’s existing data, and free human workers to focus on higher-value healthcare activities.

Leave a Reply

Your email address will not be published. Required fields are marked *