How CSMO Automated 90% of Referral Workflows End to End

Welcome and introductions

Irene: Welcome, everyone. I'm Irene from Assort's marketing team, and I'll be doing some quick housekeeping before I turn it over to our speakers. Today's conversation is how CSMO, the Center for Sports Medicine & Orthopaedics, was able to automate 90% of referral workflows end to end.

A few quick notes before we begin. We are recording the session, so we'll send it out to everyone who registered. Please feel free to submit questions to our speakers throughout the entire webinar. We'll save time at the end to answer as many as we can, and we'll follow up on anything we missed. Now I'd love to turn it over to Marie to do some intros.

Marie D'Aiuto: Thank you, Irene. Hi everybody, my name is Marie D'Aiuto. I'm the VP of Patient Access Innovation at Assort Health. I've spent most of my career leading patient access and operations for large orthopedic organizations, including Rothman Orthopaedics, Florida Orthopaedic Institute, and Orthopedic Solutions Management, an MSO.

What I love about my role today at Assort is that I can bring that operator perspective to the technology side, really helping healthcare organizations think differently about how they use AI and technology to improve access, simplify workflows, and ultimately create a better experience for both patients and staff. I'm excited to be here and looking forward to the conversation.

We were going to have Connie, the Chief Operations Officer at CSMO, join us today, but unfortunately she wasn't able to attend. So I'm going to hand it over to Jensine for an introduction.

Jensine Joseph: Just like Marie said, my name is Jensine. I manage our phone center at CSMO. I help the team with getting referrals done and with the scheduling calls that come in each day. We have a high volume of calls, so I'm really working through making sure patients are scheduled correctly and with the correct doctors.

Marie D'Aiuto: Thank you so much for being here today, Jensine. Here's where we're headed. We'll start with why referral queues are so difficult and what the process used to look like at CSMO. Then we'll show how the workflow works today, walk through a live demo, and come back to results and lessons learned from the team. Please send questions as we go; we'll leave time for them at the end.

The thread to keep in mind is simple: a referral represents demand, but it only becomes care after someone processes it, finds the right path, reaches the patient, and gets the visit scheduled.

Why referral queues are so hard

Marie D'Aiuto: Let's start at the moment a referral arrives. It may already represent a patient who needs care, but the practice still has to turn an unstructured document into a correct appointment. That work between "referral received" and "patient scheduled" is where the queue begins to build.

Before we talk about CSMO, let's look at why this is so hard across specialty care. Referrals can arrive through fax, email, EHR portals, and web forms, as most of you are probably aware. The format changes, but the work is usually the same: you open the document, interpret it, enter the information in the EMR, determine the right provider, and then your staff reach out to patients. If that document sits in a queue, the patient is waiting, and some will look elsewhere before the practice even reaches them.

There's also a visibility issue. Many practices can't quickly answer: Where did this referral come from? Has anyone contacted the patient? Where did it stall? Did it become an appointment? A lot of those data points aren't readily available to practice administrators. That is where an operational queue becomes a patient access and growth issue.

This is the compounding effect of manual work. Each referral may look manageable on its own, but the same steps repeat dozens of times a day while new referrals keep arriving. Jensine, can you bring this to life for us? What did a normal day in CSMO's referral queue look like before Assort?

Jensine Joseph: That's a good question. We had somewhere between 60 and 75 referrals coming in a day. It got to the point where we could not handle the volume, and that resulted in a backlog of patients having to wait one to two weeks to be scheduled. We wanted to be able to keep up with the pace and the volume of these referrals, and that's when we started looking into other options.

Marie D'Aiuto: In my experience, we had very similar issues. On Monday you have the call queue, which is really hard to get through, with high abandonment rates in call centers. Along with that, you have all the referrals coming in on a Monday for patients who were seen that day. You spend the rest of the week trying to resolve that Monday issue, and meanwhile those referrals are creating a backlog you can't get to urgently. That's the broader problem.

Growth and complexity at CSMO

Marie D'Aiuto: Jensine, I know this was originally meant for Connie, but maybe you can shed some light on it. As the organization started adding physicians and locations, what did that growth mean for the team managing referrals?

Jensine Joseph: Again, it meant high volume. A lot of things were happening all at once, with big moving parts. We had to make sure the templates were created correctly, that the doctors were happy with how they were being scheduled, and that their templates were being followed. Adding physicians and locations meant more work for our staff to get all of that done in a timely and efficient manner.

Marie D'Aiuto: And as you add providers, you have all these different protocols coming on board. You naturally get more call volume and more referrals. Those new doctors are probably working with your marketing team and going out to PCPs to build up their practice, so I can only imagine the influx you were getting as you continued to grow. I'm sure that's when you started looking at technology to see what repetitive work could be offloaded.

For some additional context, CSMO had 22 physicians across 12 locations in Tennessee and North Georgia. Jensine, how did that scale change the referral workflow?

Jensine Joseph: The workflow changed because we knew a lot of referrals were being missed, so we had to go back and modify the process to make sure they were coming in correctly. It also changed how we dealt with patients as a whole: making sure chart creation was done correctly and that their information, like insurance and demographics, was uploaded. A lot of that changed with the process modification.

Marie D'Aiuto: That's what I hear a lot. Organizations tell me growth gives them the opportunity to serve more patients, which is great, but it also sends more volume and complexity through a manual process. You need a workflow that can grow with the practice without making patients wait longer or asking the team to absorb more repetitive work.

Where the manual process slowed down

Marie D'Aiuto: Let's stay with the day-to-day experience for a second. A referral arrives. What happens next, and where did the process slow down?

Jensine Joseph: When a referral arrives, creating a chart can be a very manual process. A lot of the time the referral doesn't come in with a chart attached, so our staff has to manually create that chart and place calls out. That was all before Assort came on to help us. It was very manual labor, and the speed was sometimes slow because the team also has to manage incoming phone calls. Or maybe an outreach had already been done, and they'd have to figure out where the referral went, what to do with it next, and what stage it was in. There are a lot of manual things that have to be done daily.

Marie D'Aiuto: Do you have any insight into what that looked like from the leadership side, and how the delay affected patient access and growth?

Jensine Joseph: Absolutely. Patients would say, "I found someone else in the time it took you to contact me," and they'd already be scheduled somewhere else. We started seeing a decline in the number of referrals that were actually being scheduled.

We also experienced frustration because sometimes fractures would come through the referrals and we weren't able to take care of them in a timely manner. As many people know, fractures can heal quickly, and damage can sometimes be irreversible if patients aren't scheduled correctly.

Marie D'Aiuto: I remember talking to Connie about this. With all that growth and expansion, there was no clear view of where referrals came from, where they stalled, and how each one performed. That delay is easily described as a queue problem, but the patient experiences it differently. Another provider has already told them they need care, and then they wait several days without knowing what happens next. Jensine, what did that mean for patients and the team?

Jensine Joseph: Patients were frustrated. A provider had told them they needed to be seen urgently, and then we basically had to delay that because of the high volume of referrals. At one point we had to tell patients, "We've received your referral, and it's going to take time." We even posted something on the website saying there would be a delay, but that we hadn't forgotten them. That was something we were really frustrated with.

Marie D'Aiuto: And by the time you reach them, maybe they've gone somewhere else, or they feel it's been too long and they don't need help anymore. If you had called a week ago, they would have come in. Those are the pain points of not getting to a referral quickly enough. The same pattern appears across many specialty practices when growing demand runs into a process that depends on people completing every step.

How Assort approaches referral automation

Marie D'Aiuto: Now I'm going to show you how Assort approaches that work, starting with a quick introduction to Assort and our platform. We build AI agents for patient access, trained around the workflows and operating rules of specialty care. That's what really sets us apart. The value isn't a single interaction in isolation: as more steps connect across the patient journey, the workflow can carry forward context the practice already has and reduce repeated work for patients and staff. Referrals are one of those connected workflows.

This is the foundation behind what you'll see in the demo. The platform connects AI agents with EHR and phone integrations, analytics, multilingual support, and tools for staff. Patient journey memory sits over all of that and carries context forward, so patients don't have to start from zero at every touchpoint. For referrals, that means the workflow can use specialty-specific rules to move from a document toward the right appointment, while keeping staff involved where judgment or a human touch is needed.

Here's how Assort Referrals works, left to right. First, Assort ingests the referral from an EHR feed, fax, email, or an upload, and turns that document into structured patient information. Next, it classifies urgency. Based on your practice's specifications, Assort can determine whether something like a fracture or an ACL tear is urgent, versus pain that's been going on for five weeks.

It checks benefits and eligibility and sends exceptions to a manual review queue, so a person can handle that patient and make the outbound call instead of the automation. It then applies the practice's protocols and payer rules to recommend the best-fit provider. Finally, it begins outreach by voice, text, or email, books the patient into the EHR, and closes the loop with the referring provider. The predictable path moves automatically; missing information, ambiguity, and exceptions still go to a person.

Jensine, how does that division of work feel to your team?

Jensine Joseph: We still have people working on things, of course. We don't take away jobs just because we have Assort helping us. But it has definitely relieved a lot of their burden. They used to log off at the end of the day and text me, "I'm stressed about seeing all these numbers piling up, and I feel like I haven't done enough today." Now they're more comfortable. They're not stressing all the time or taking work home, and at the end of the day they know work is being done for them in a more efficient manner.

Closing the loop with referring providers

Marie D'Aiuto: So your staff can focus on the referrals where a person makes a bigger difference. The workflow also closes the loop with the referring provider's practice. CSMO can automatically notify the referring provider when the referral is received, when the patient is scheduled, or when the referral can't move forward. Jensine, why does that communication matter?

Jensine Joseph: Communication is so important, just like in everyday life. It's important to connect the dots between a practice and what happens after they've sent a referral. Being able to show which step the referral is in keeps everyone engaged, and if something changes or isn't up to par, we can take that feedback too. It keeps the relationship open.

Marie D'Aiuto: Absolutely. The referring practice no longer has to wonder what happened to the patient they sent over. Were they taken care of? Did they decline care? Now they can see that the referral was received and what the result was. That visibility strengthens the relationship between the referring provider and the rendering practice, and it shows the referring provider that you value the referral and follow through.

Pipeline visibility and attribution

Marie D'Aiuto: The other major change is visibility. CSMO can now see the pipeline and conversion rates by referring practice, referring provider, receiving provider, and payer. Jensine, what can you act on now that was hidden before?

Jensine Joseph: Now we can see exactly why patients aren't scheduling. We can break it down and see what went wrong and what we could have done differently. That gives us insight into whether we need to revisit some of our policies. Have we put unnecessary restrictions in place? Is it something the referring physician needs to weigh in on? Did we miss something within the referral? It gives us a lot more insight into things we might not have caught before.

Marie D'Aiuto: And with the data attribution, if I'd had this at one of my old practices, seeing where referrals come from and their conversion rates month over month, I'd hand it right to my marketing team and say, "I see a 10% decrease in referrals over the last few months from this practice. Let's bring them some bagels, introduce one of our newer physicians, and build that relationship." You can really see how referral patterns are changing, and it provides visibility into growth and strategy opportunities.

One more capability coming soon is online self-scheduling for referred patients. I know you're excited about this one too, Jensine. The outreach message will include a link, and because the information is already extracted from the referral, it will prefill the experience so the patient doesn't have to re-enter everything from scratch. The practice still controls all the routing rules, including whether a patient may choose a provider beyond the one named on the referral. Patients get another convenient way to book, and CSMO keeps the guardrails it needs.

Offering different avenues is so beneficial. Even though I ran call centers for years, I don't want to call anywhere. If I hurt my shoulder, I'd much rather get a text. Being able to meet different types of patients where they are is crucial.

Live demo: from fax to booked appointment

Marie D'Aiuto: Let's make this concrete. I'm going to show one referral going from fax to a booked appointment. As you watch, notice three things: what the system extracts from the source document, which steps move automatically, and where a staff member can review or step in when needed.

This is the referrals dashboard in Assort Health. Right away you can see all the referrals that came in, how many are pending outreach that requires a person, how many the automation has already reached out to, and any where a task is required, like a second opinion or a request for additional information.

Because we're integrated with the EMR, this is the information extracted from each referral: patient information, whether it's urgent or non-urgent, and the patient type, new or existing. We also automatically identify duplicate charts, so if a patient already exists, we make sure they're treated as an existing patient rather than new.

In this example, outreach has already started and is in progress. Selecting the patient shows the attempt cadence that occurred: day one, day two, and whether it was voice, text, or email. All of this is customizable to the practice. Assort also has a lot of best-practice data from other practices, so when you're ready to integrate, we can recommend what a best-practice cadence looks like, but you can fully customize it. We can also show transcripts, so you can see exactly what the AI agent said on the phone during outreach.

You can see the information presented in the referral, including demographics, insurance, and patient details. Under conditions, you can build in your rules. For example: "I don't want workers' comp or motor vehicle accident patients to be scheduled automatically by the AI agent. I want those to go to a dedicated team." We identify that information in the fax referral, extract it, and route the patient to the team that needs to handle it manually.

Here's an example of the referral itself. You can view the original fax in the Assort platform alongside everything we extracted: demographics, insurance, referring provider, who the patient is being referred to, and the chief complaint. If something is missing, like a phone number or a last name, we can build rules that route it to a person as a task, so they can call the PCP and say, "We received this referral, but we're missing the phone number. Can you send it so we can reach the patient?"

In closed referrals, meaning those that were scheduled or not, we document every step of the outreach cadence along with any other information. This is the PDF that was sent back to the PCP. That's the closing-the-loop process in Assort's referral product: once we receive the referral, we send the PCP a notification thanking them and confirming receipt. The message can be customized to your practice's needs.

Here's an example where we got the patient scheduled, so we're telling the PCP the appointment is with this doctor, at this time, for this chief complaint. And if the patient declined to schedule for whatever reason, we identify that reason and send it back as well, again with any custom language you'd like to include.

Now let's pivot to the analytics dashboard. This is a high-level view of the information we can present to your administrative team about referrals. There are a lot of key metrics, and we can customize them to whatever you'd like to see. A few big ones I like to call out are total received, conversion rate, and average time to process for your human team.

In the attribution data, which is what we were talking about earlier, you now have visibility into each referring practice and referring provider: how many referrals they send you and how many convert, down to the practice level. This is exactly the information I would hand to my marketing team to say, "We're seeing a decrease from this area. Let's do some outreach and build a strategy to continue our growth there." That's the referral workflow end to end with Assort.

What changed at CSMO

Marie D'Aiuto: Now let's come back to what changed at CSMO. Jensine, during a recent walkthrough, you saw that a staff member could click an extracted field and jump to the exact place it appeared in the referral. I think your reaction was, "Get out of town, this is amazing." What made that feature land so strongly with you?

Jensine Joseph: Sometimes you've got 20 to 30 pages to go through, and you're trying to figure out where a piece of information came from. When you have to verify whether it's accurate, you don't want to follow behind someone else page by page. Being able to click on something and go right to that field makes life so much easier. That's what technology is all about nowadays. I can tell my staff, "If you need to verify something, it's just a click away."

Marie D'Aiuto: If you put yourself in Connie's shoes, or leadership's shoes more broadly, what do you think was the biggest change from that perspective?

Jensine Joseph: I think it goes back to what I said earlier: being able to contact patients in a timely manner. It also allows my staff to work on things that need their time. They don't need to be sitting around creating charts for incoming referrals. They have more time to make sure they're properly scheduling the patient they're on the phone with, instead of rushing through and thinking, "Oh my goodness, I have to finish this referral." It lets us work on things that actually need our time.

Marie D'Aiuto: That makes so much sense. From Connie's perspective, I'd add that gaining capacity and visibility at the same time is one of the big things. The automatic work is now handed off, so your team can handle the complex patients who need more of a human touch. And from a leadership perspective, you have visibility into how the referral pipeline is performing, what could change, and where the opportunities are.

Results: hours instead of days

Marie D'Aiuto: Here are the headline results for your team. First outreach moved from three to four business days to hours. Approximately 90% of referrals now move end to end without any manual work, and about 77% of referrals convert to scheduled appointments. Jensine, what change do you feel most in a normal workday?

Jensine Joseph: A huge burden being lifted. I'm sorry if I sound like a broken record, but it comes back to the volume that comes through day to day. We receive anywhere from 700 to 1,200 calls a day, in addition to patient portal messages, internal referrals, and external referrals. The biggest change I've seen is lifting that manual process of going through each individual chart and faxing things back, and when something is missing, having to locate it and let it sit again until someone can get to it. It's a workload lift. It turns something that was so manual into something more feasible and automated.

Marie D'Aiuto: When I think about my past practices, imagine being able to reach a patient right as the PCP sends the referral over. By the time they're out to their car, Assort can reach out and get them an appointment, rather than leaving them waiting for days and wondering why the doctor never called, until they decide to go somewhere else. Reducing that lag time to almost immediate is crucial, especially in a competitive specialty like orthopedics, where you want to deliver a superior patient experience and get patients on the schedule quickly while still adhering to physician protocols.

So the numbers matter, but the day-to-day change matters just as much. Patients hear from CSMO sooner, and staff have more time for the cases that need human attention.

Since Connie couldn't join us, I'll share what she told me when I asked how pipeline visibility has changed physician outreach and marketing decisions. CSMO now connects referral sources with outcomes, instead of treating the queue as one undifferentiated volume of work. The attribution report gives leadership confidence in referral patterns: her marketing director can see when a referral pattern might be changing, focus attention there, and measure the effect over the following months.

Change management and building staff trust

Marie D'Aiuto: Jensine, what did the team need to see before trusting the workflow? You manage a big call center team, and whenever you implement technology, you probably hear, "Is this taking over my job? Why are we doing this?" Change management is so important. How did you introduce it, and what was your talk track with your team?

Jensine Joseph: Naturally, there was a lot of hesitation. The team needed to see it work and see it in action. They wanted to see live referrals and whether it was doing what it was supposed to do. If there was an exception to one of our rules, did it get to the right person? Was the patient being scheduled with the correct physician, especially for something we wouldn't normally treat? Seeing live demonstrations and live referrals being scheduled made a big difference. The Assort team was great about sending us the first 10 referrals they scheduled and letting us audit them. I had my staff go through those referrals and see it firsthand, and that really helped them see the process in real time.

Marie D'Aiuto: I've always found that when staff are part of the testing process, they start to believe in it faster and have more confidence in the system. They see it booking a patient exactly like they would have. It also helps to make sure your team knows this is help for them. They're inundated with back-to-back calls and referrals to work, and automating this high-impact portion gives them a breather so they can focus on the patients who need them. For anyone in the audience facing staff hesitation, talking about the technology as a teammate instead of a replacement is a big deal.

What made the partnership work

Marie D'Aiuto: Jensine, you've described the Assort and CSMO partnership as one of your strongest. What made the implementation feel different compared to others you've done in the past?

Jensine Joseph: The responsiveness, for sure. I tell everyone I talk to about Assort that their open mind to our feedback really helps. It's one thing to send feedback and hear, "We've never seen that happen before," or "Our test scenarios didn't do that." Instead, they took the scenarios we presented, looked at all the reasons it might have occurred, and were responsive. They provided timely solutions, and when they hadn't found a fix yet, they told us, "We're still working on it." That update piece was amazing. It felt so different from other things we've done, and I'm very appreciative. I tell them that all the time.

Marie D'Aiuto: I'm so happy to hear that. Any technology implementation has some messiness; I've been through plenty in my career, and you don't always know what you're getting into with a new platform. But our teams met weekly, sometimes daily in the very beginning, reviewing each step of the referral workflow to make sure we truly understood it, what made the most sense, and what we could automate. And the team stayed involved after launch. When an issue came up, we came together, and you were so great about working with us as a partnership to make the product the best it could be. Thank you. You've been an amazing partner, and that hands-on approach is so critical to early success.

One last question, Jensine: what did that look like from the team's perspective?

Jensine Joseph: They were hesitant when things were broken, and they were nervous about whether technology could do the same things they do. But in the end, the partnership with Assort, their reliability, and their willingness to change things made the difference. The team saw that when they had feedback, they could bring it to Assort, and Assort would help implement it and make it as close to human as possible. That really helped with their day-to-day reassurance.

Marie D'Aiuto: That's great to hear. Again, you've been an amazing partner in building this out, and we're looking forward to the enhancements we have in the pipeline and to continuing to work with you.