I used ChatGPT, a report from my Bi-Pap and the requirements for 'pre-authorization' to help me qualify for some medication. It speeded up the approve process significantly by building out a report with the requirement and how I met the requirement w/ the data to back it up.
I have an AI client who has leveraged the new IRS Rules incentivizing employers up to $640 per employee and employees up to $1200 per year to adopt preventive health on top of their health plan. If you want to learn more message me!
Great idea, Sabrina, to make such a practical use case! As someone who works in this industry and sometimes serves as a content manager, it is useful. Some healthcare companies claim they don't use AI (even though everyone does, whether their company knows it or not)... One reason I'm trying to help is to support those who have a policy but don't know how to communicate about it.
Airlines republish schedules by the hour. A rebooking assistant that reads last Tuesday's PDF will offer you a flight that no longer exists with the same confidence as one that does. Provider directories and formularies are the same shape of data. They are supposed to stay current, and they are the first thing to go stale. Your 'flag anything requiring verification' rule is the only thing standing between a confident answer and a wrong network.
Girlfriend, this is a great piece of information that older people need. The question is if they’re opposed to learning or trying AI they could be missing out on a lot of money.
Healthcare payer coverage is not simple. Medicare, commercial payers, authorization rules for inpatient and outpatient services, it's a constantly moving target. Medicare alone changes multiple times a year as updates and the annual changes in coverage. Commercial payers either follow Medicare's rules or write their own. Sometimes both. It's a web, and navigating it takes clinical knowledge, not just AI automation. Something the bean counters won’t see.
Here's what a typical day looks like for me. I cover several hospitals in a large health system. We review outpatient surgical cases two days out. We meet with patient financial services, pre-service, in a daily pre-surgical huddle where UM, case management OR leaders. All of us touching the same patient list, asking the same question: is this case staying outpatient, or is it going to flip to inpatient? That question matters because the answer determines whether we get paid, get denied, or get audited.
And that question requires clinical judgment. When I'm reviewing a pre-service case, I'm looking at the surgery, the patient's comorbidities, the physician's request, what the payer actually approved, and whether post-op monitoring is going to be needed. No AI does that the way a trained nurse does, at least not yet, and honestly, maybe not ever.
That said, 2027 is something we need to be talking about now. Medicare's inpatient-only list is going away. Those procedures are inpatient for a reason and they were never meant to be done in an outpatient setting. Once that list disappears, payers will have room to push cases outpatient, and we're going to see a wave of denials we aren't fully prepared for.
AI has a role in all of this. I'm not anti-AI. It's useful for pulling data, identifying patterns, and supporting medical necessity reviews using tools like InterQual, MCG, and Milliman. But AI also pulls wrong information just as confidently as it pulls right information. That's not a small problem in healthcare. Clinical validation isn't optional, it's the safety net.
I'm a Revenue Integrity Nurse Auditor. I've worked the payer side, the hospital side, third-party vendors, and offshore staffing. I've seen this industry from every angle. My honest concern is that AI is going to eliminate roles that genuinely need to stay human and we're going to feel that loss before we fully understand it. This work is complex, it changes constantly, and it requires someone who can look at a patient holistically and make a call.
I watch the same discipline fail in Indian insurance. Clients sign endowment and ULIP policies they never read, over-insure through products they don't understand, and stay under-covered on plain term.
The renewal notice lands, the rider changes, nobody translates it.
AI surfacing the questions to ask a human advisor is the exact value here — the decision still belongs to a person who has read the document.
Sabrina's own prompts already build that in. The instruction to write 'not found' instead of guessing keeps this from turning into a bad recommendation. In a market like the one you're describing, without one central place like Medicare.gov to check an answer against, that same default probably needs to be even stricter, since there's no single place left to catch a wrong guess.
Love this approach. As a pharmacist, this is exactly where I think AI can be incredibly useful in healthcare — helping people understand complicated information and work out the right questions to ask.
Use AI to become better informed, then verify with the right healthcare professional. That’s the sweet spot.
This is great timing to post this Sabrina, just in time for open enrolment. I am in the benefits space and what i hear about every day in healthcare (other than rising costs) is about denials. I've created a tool to help appeal a denial at healthcaregenius.com . The sad fact is that way too many claims are denied, and way too few of the denials are appealed. The good news is that if you appeal a denial, you have a good chance of having it overturned.
Brilliant!
I used ChatGPT, a report from my Bi-Pap and the requirements for 'pre-authorization' to help me qualify for some medication. It speeded up the approve process significantly by building out a report with the requirement and how I met the requirement w/ the data to back it up.
I have an AI client who has leveraged the new IRS Rules incentivizing employers up to $640 per employee and employees up to $1200 per year to adopt preventive health on top of their health plan. If you want to learn more message me!
I want to learn more re incentive for preventive health
Great find some time next week: schedule.gordonjones.org
Info please
email me at gordonjonesdha@gmail.com
Great idea, Sabrina, to make such a practical use case! As someone who works in this industry and sometimes serves as a content manager, it is useful. Some healthcare companies claim they don't use AI (even though everyone does, whether their company knows it or not)... One reason I'm trying to help is to support those who have a policy but don't know how to communicate about it.
u in here already? https://www.skool.com/women-building-ai-2577
Yes since you started but limited time but appreciate your info. I will make time…
This is great advice. Thank you
This seems like a natural app play.🤷🏿♂️ Painful problem; large TAM; reccuring; compex process; quick win.
Airlines republish schedules by the hour. A rebooking assistant that reads last Tuesday's PDF will offer you a flight that no longer exists with the same confidence as one that does. Provider directories and formularies are the same shape of data. They are supposed to stay current, and they are the first thing to go stale. Your 'flag anything requiring verification' rule is the only thing standing between a confident answer and a wrong network.
Info please
Girlfriend, this is a great piece of information that older people need. The question is if they’re opposed to learning or trying AI they could be missing out on a lot of money.
Healthcare payer coverage is not simple. Medicare, commercial payers, authorization rules for inpatient and outpatient services, it's a constantly moving target. Medicare alone changes multiple times a year as updates and the annual changes in coverage. Commercial payers either follow Medicare's rules or write their own. Sometimes both. It's a web, and navigating it takes clinical knowledge, not just AI automation. Something the bean counters won’t see.
Here's what a typical day looks like for me. I cover several hospitals in a large health system. We review outpatient surgical cases two days out. We meet with patient financial services, pre-service, in a daily pre-surgical huddle where UM, case management OR leaders. All of us touching the same patient list, asking the same question: is this case staying outpatient, or is it going to flip to inpatient? That question matters because the answer determines whether we get paid, get denied, or get audited.
And that question requires clinical judgment. When I'm reviewing a pre-service case, I'm looking at the surgery, the patient's comorbidities, the physician's request, what the payer actually approved, and whether post-op monitoring is going to be needed. No AI does that the way a trained nurse does, at least not yet, and honestly, maybe not ever.
That said, 2027 is something we need to be talking about now. Medicare's inpatient-only list is going away. Those procedures are inpatient for a reason and they were never meant to be done in an outpatient setting. Once that list disappears, payers will have room to push cases outpatient, and we're going to see a wave of denials we aren't fully prepared for.
AI has a role in all of this. I'm not anti-AI. It's useful for pulling data, identifying patterns, and supporting medical necessity reviews using tools like InterQual, MCG, and Milliman. But AI also pulls wrong information just as confidently as it pulls right information. That's not a small problem in healthcare. Clinical validation isn't optional, it's the safety net.
I'm a Revenue Integrity Nurse Auditor. I've worked the payer side, the hospital side, third-party vendors, and offshore staffing. I've seen this industry from every angle. My honest concern is that AI is going to eliminate roles that genuinely need to stay human and we're going to feel that loss before we fully understand it. This work is complex, it changes constantly, and it requires someone who can look at a patient holistically and make a call.
That's not something you automate away.
I watch the same discipline fail in Indian insurance. Clients sign endowment and ULIP policies they never read, over-insure through products they don't understand, and stay under-covered on plain term.
The renewal notice lands, the rider changes, nobody translates it.
AI surfacing the questions to ask a human advisor is the exact value here — the decision still belongs to a person who has read the document.
Sabrina's own prompts already build that in. The instruction to write 'not found' instead of guessing keeps this from turning into a bad recommendation. In a market like the one you're describing, without one central place like Medicare.gov to check an answer against, that same default probably needs to be even stricter, since there's no single place left to catch a wrong guess.
Love this approach. As a pharmacist, this is exactly where I think AI can be incredibly useful in healthcare — helping people understand complicated information and work out the right questions to ask.
Use AI to become better informed, then verify with the right healthcare professional. That’s the sweet spot.
Thank you for sharing
This is great timing to post this Sabrina, just in time for open enrolment. I am in the benefits space and what i hear about every day in healthcare (other than rising costs) is about denials. I've created a tool to help appeal a denial at healthcaregenius.com . The sad fact is that way too many claims are denied, and way too few of the denials are appealed. The good news is that if you appeal a denial, you have a good chance of having it overturned.