Advancing behavioural healthcare from inside the castle wall

By Published On: September 12, 2025Last Updated: October 6, 2025
Advancing behavioural healthcare from inside the castle wall

Andy Flanaghan is CEO of Iris Telehealth, a telepsychiatry company operating across the US.

The company recently launched it’s Iris Insights tool which leverages AI to reduce behavioural health escalations and ensure patients get access to timely care.

We spoke to Andy to get his thoughts on the challenges of working in behavioural health and find out more about how the tech is improving efficiencies on the ground.

 

Hi Andy. Please tell us a bit about yourself and your background.

I’ve been in technology and healthcare my entire career and have led Iris Telehealth for almost five years now.

We’ve been through some tumultuous times – Covid, post-pandemic and now the mainstream rationalisation of mental healthcare.

It’s a pretty challenging world. My job is to stay focused on our strategy and make sure we’re always delivering high-quality care at the same time.

You mentioned Covid. I imagine that was something of a game-changer for you, as it was for the health tech industry as a whole. It exposed a lot of issues within the healthcare system.

It has exposed a lot of things, but not all of them are negative. The pandemic firmly established telehealth for mental health.

Virtual appointments have proven to be as clinically effective as in-person appointments.

Of course, there are things that need to be done physically. But on the flipside, a blended model of in-person and virtual is emerging as a very natural care delivery method which the world didn’t believe [was possible] pre-Covid, or at least in the US.

As a result, there’s now better access, it’s much more consumer-centric.

Healthcare AI has proven to be a game-changer too. What can you tell me about your new Iris Insights platform? What problem are you hoping to solve with it?

Yeah, it’s really exciting for us.

We’re a large medical group with roughly 800 employees, 650 of which are clinicians, and so AI can have some tremendous administrative benefits and some personal benefits in terms of efficiency and quality of life.

Andy Flanagan

What we’re learning is that the dysfunction in our medical system, whether it’s in the US or around the world, can be improved by small, incremental improvements, and technology is the right tool for that.

Our job with Iris Insights is to streamline and improve the patient experience, the clinician experience, and reduce what is widely acknowledged as 30 per cent of the waste of the dollars that we spend, through technology.

It’s a complementary platform for medical groups, health systems, clinicians and patients.

We are a medical group, so we’re our own customer. That’s how it started. We built it for us.

The first precious resource is the clinician.

We use the insights to optimise their schedule, [taking into account] the demographic of the patient, the history and diagnosis.

[The challenge then is] to optimise that schedule so that the clinician isn’t working 18 hours a day, and is only seeing patients that are appropriate for their licensure.

How did you develop the tech?

We had a lot of clinicians and a lot of patients and realised that there had to be a better way of managing them.

A patient would no-show and then call back later saying ‘I’m free now’. But now the clinician is double-booked, so that’s not going to work.

So we started using the technology to see if we could find another clinician who had a no-show at that time and swap the patients over.

But the key was to ensure that the swap was both clinically-appropriate and safe.

Ultimately, we wanted the software to help us be more agile in supporting the daily life of the patient and the clinician, and it worked.

I guess underpinning this is data, and an important part of the work that you do is ensuring the security of that data and that it’s being used by the right people and interpreted in the correct way.

First of all, we understand data very well. We understand the ontologies, we understand how data is structured.

In behavioural health, it’s unstructured data. The note is a narrative, so it’s very hard, compared to, say, cardiology or oncology, where you have biometrics involved.

We adhere to all the data policies of our partners.

We’re in the systems of our health systems and our community partners. That allows us to avoid transmission errors and hacking.

We don’t hold the data for someone to come and hack our account. It’s securely in their electronic health record and in their data warehouse.

It took more engineering on our part to keep it there. It would have been easier just to take it and put it in our cloud.

But strategically, we made the decision not to take the data, rather to build in their environment.

Whenever you’re introducing a new technology, there’s always that challenge of trying to get it to integrate with other existing systems.

How does that work with Iris Insights?

We are on the inside of the castle wall, as a friend of mine always likes to say.

We are a remote user of their health system and an employee of their organisation, so we get to work with the data in their environment, natively.

We send our tools into their environment through a virtual machine. We completely broke that whole barrier to getting data. We don’t have that friction.

It was more expensive to do it our way. We certainly would be touting our ability to train on a huge dataset if we had every patient record for every single customer.

That’s not our goal. We’re a medical group. We’re about delivering care and data is an enabler for efficient care delivery.

What do you see as the big challenges in behavioural health at the moment, and what can you do to help tackle those?

It’s hard to make money in behavioural health. You have to be an operator. You have to grind.

We’re grinders here. We don’t have corner offices. We’re doing that hard work, and that’s what it takes.

Reimbursement is pretty low, and as a result, you have to be really focused on the problem that you want to solve.

There are bright lights in our industry, but for many of the people who showed up during the pandemic to create a virtual team to solve this problem with data or software, it’s not ending well.

That’s because they’re not really solving the whole problem. And that’s the journey.

It’s a personal journey, unfortunately, usually for a long time in mental health, certainly behavioural health.

Our job is to be very present on the problems we solve and not confuse ourselves with what we think we have the right to solve.

And so we’re always trying to say ‘no’, fast. It’s working for us.

How tech solutions can support the UK’s Life Sciences Sector Plan and reboot clinical trials 
How AI is reshaping medical imaging – and what it means for healthcare