Dr. Jeremy Sharp (00:00.568)
Hello everyone and welcome to the Testing Psychologist podcast. I’m your host, Dr. Jeremy Sharp, licensed psychologist, group practice owner and private practice coach.
Many of y’all know that I have been using TherapyNotes as our practice EHR for over 10 years now. I’ve looked at others and I just keep coming back to TherapyNotes because they do it all. If you’re interested in an EHR for your practice, you can get two free months of TherapyNotes by going to thetestingpsychologist.com slash therapy notes and enter the code testing. This episode is brought to you in part by PAR, the Personality Assessment Inventory Bariatric.
compiles the results of the PAI into a useful report for bariatric surgery candidates. Available on PAR iConnect, PAR’s online assessment platform, you can visit parinc.com backslash P-A-I-B-R.
Hey everyone, welcome back to the podcast. I’m excited to be here with you. I’m to be back doing the podcast. have to say, just reflecting back over the summer and accidentally taking a couple months off, which I talked about in a previous episode. You can go check that out. It has been great to get back to the podcast and be in research mode and planning mode and recording mode and just know that, get some episodes out and diving deep into some of these.
topics that are so fascinating in the world that we live in and getting to connect with all of you again via the podcast. Now, today’s topic is near and dear to my heart. We’re talking about AI. So you may have caught the previous episode that was entirely generated by AI. There was none of me in that episode whatsoever, except for my ideas. If you didn’t catch that one, go check it out. If nothing else, novelty’s sake, just to
Dr. Jeremy Sharp (01:54.924)
see what AI voice generation can look like these days. But that episode was meant to kick off our short AI series. It’s gonna be a three part series. Today is part one, where I’m talking about real world AI for testing psychologists. I’m gonna talk about practical tools and ways to utilize AI. I’m gonna talk about how to get started with AI if you have no idea what to do and there are, I know a lot of you out there in this world and there’s no shame around that.
Hopefully, it’ll break it down and make it pretty easy for you. We’re also going to talk about common tasks that clinicians can use AI for both within and outside of report writing. As usual, I’m sure I’ll deviate off into some other topics related to this area. If you are interested in AI like I am, or just curious, or maybe you’re an expert, hopefully there’ll be something here for all of you.
So let’s get to this discussion about practical ways to use AI in your practice right now.
Dr. Jeremy Sharp (03:15.47)
All right, everyone, we are back and we are talking about AI and its practical applications for testing psychologists. Now, I occupy a couple of roles, I think, that are pretty relevant for this discussion. The first, of course, is that I’m a co-founder of Reverb, which is an AI report writing platform. And I want to disclose that, of course, right off the bat. I will mention Reverb throughout these next few episodes, but there are plenty of other tools out there that can be helpful depending on what stage and
what your needs are in your practice. just to throw that out there, I am a co-founder of Reverb. So I have an investment in this space, certainly. My other role, of course, is as a practice owner and a larger practice owner, where there are several of us who are doing a number of evaluations every month. And so we’re constantly looking for ways to become more efficient. And I think AI is a piece of that puzzle. So I’ve got the…
practice owner hat on, and then of course I have my consultant and podcast hat on as well. this really serves as an intersection, I think, all the things that I’m doing in my life right now. And that makes it super energizing for me. So I’m excited to chat with y’all about all of this. Now, I’m gonna go through a lot of questions that have come up over the last few months around AI. Many of these questions came from
an AI panel that I participated in, which was fantastic. This panel occurred at the AAPDN virtual conference back in April of 2025. I was on that panel with Chris Barnes and Liz Angoff and Mandy White-Ajbani. And so a lot of these questions come from a panel that I participated in back in April of 2025. This was a panel for the AAPDN virtual conference and the panel was
Fantastic. It was really dynamic and we got tons of great questions and I got to hang out with some fabulous psychologists, Dr. Chris Barnes and Dr. Liz Angoff and Dr. Mandy White-Ajmani who moderated the panel. So it was a lot of fun. We got to have a great discussion around AI, but a lot of questions popped up that we didn’t get to and weren’t able to answer. So I brought some of those in and then of course augmented with some of my own experience and questions that popped up through my consulting work with folks. So let’s just dive
Dr. Jeremy Sharp (05:37.548)
right into it. I want to start with kind of a 30 second elevator pitch, so to speak, for what AI can and can’t do for psychologists right now. This is not meant to be all inclusive or the end all be all, but this is way that I think of it. I think of AI as being pretty excellent at generating and summarizing text, like pulling key points from an interview transcript or translating your report into a different reading level. It’s good at organizing your ideas and other large amounts of information into coherent themes.
like finding research-supported recommendations from the internet or finding similar clinical threads in clinical data. It’s also good at serving as what I will call a thinking partner to bounce ideas off of. Those are some things that I think AI does really well, and we’ll translate that to testing tasks here shortly. What it’s not great at is generating rote templated text that needs to be the exact same verbatim wording every single time.
That’s simply not the nature of AI and generative AI. It’s also not good at quote unquote thinking for itself. So AI is not really doing any independent thinking. It’s just doing really, really sophisticated text prediction. I’m not going to get into the nuances and mechanics of machine learning and text prediction and language models and that kind of thing. But suffice it to say, the way I think of it is that AI or language models specifically are basically just doing text prediction like on your iPhone when you
go to write a text and it suggests the next word for you. It’s basically doing that, but just in a really, really sophisticated way with a ton of nuance and detail and training. So that’s a 30 second elevator pitch, much longer than 30 seconds, of course, but that’s a good intro, I think. Now I think we can spend some time just talking about how AI can help within assessment, and then we’ll transition to some ways that it can help outside of the assessment specific realm. So some of the most common, I would say,
writing tasks that clinicians are successfully offloading to AI at this point. It covers a lot of the report writing process, which is, I think, good news for us. But I think of, you know, background and history is an almost perfect application of AI. AI does really well at reading a transcript or, you know, handwritten or typed notes, organizing those notes and turning them into a really, you know, nice, like, pretty organized and coherent background or history.
Dr. Jeremy Sharp (08:04.462)
place that I feel like it falls short here is to, you know, sometimes it can have a lot of difficulty offering direct quotes. Sometimes that’s tough. But otherwise, back generator and a background in history is one of the best uses of AI as far as I’m concerned. Another way that it can help with report writing is to generate score tables. This is also pretty basic. Now, through running reverb, we have found that psychologists have a lot of
ideas and feelings about how our tables should look in our reports. So I will say that AI is really good at generating what I would call kind of basic tables. And if you want to do customized tables, that’s a little bit tougher. But if you just want basic score tables where you can feed in a report or score report and then get a table with those scores and percentiles and descriptors, that’s pretty doable. Another place that I think it does really well is summarizing collateral interviews. Okay, so again, this is just
essentially text analysis. So if you have a transcript of an interview, an audio recording of an interview, written notes, again, either handwritten or typed notes from a collateral interview, doesn’t matter how messy it is or how disorganized it seems, AI can do a really good job of organizing and summarizing those interview records. Similarly, it can do pretty well at summarizing documents and other records. So
PDF documents, Word documents, faxes, things like that. So we’re talking about medical records, prior evals, IEPs, imaging even. I had a, I guess, off-label use of AI outside of my own practice where I was able to analyze a family member’s medical records that had many years of complexity and imaging and so forth. And I was able to come up with some pretty robust findings and summaries.
So summarizing documents and records is also a task that AI can do really well. Similarly, organizing behavioral observations. Great task for AI. So essentially you’re getting a theme here like anything where you are taking notes and need to organize and write those notes into some kind of prosaic format is gonna be a great task for AI. Another thing that it can do relatively well is generate visuals or presentations. I think
Dr. Jeremy Sharp (10:22.626)
This takes a little bit of finagling at this point. It’s gonna be a little tougher to get exactly what you want right out of the box, but it is possible with AI. And depending on how you format those images or those presentations, some methods or formats can be easier than others, which is great. So you can tell, mean, just listening to that list, there are many things in the report writing process that I think AI can do really well. I did not mention
generating recommendations, it can generate pretty basic recommendations, I think really well and can even search the internet and find local referral sources. The trouble with recommendations for me is that then we have to go in and provide, I think, a fair amount of nuance for those recommendations just based on the client that we’re seeing and making it a little more personal. But it can also generate recommendations relatively easily. So moving on to another question, one of the questions that came up
I touched on this a bit, is can AI summarize past evaluations or records accurately? I accurately being the active word here. And what are the limits and benefits there? So I would say definitely. I can’t speak to other platforms necessarily aside from ChatGPT. Like I said, I was able to use it to analyze some family members’ records. Our software, Reverb in particular, allows you to summarize records by telling the software exactly what you’re trying to extract.
from those records. I can’t speak for others, but this is possible and it’s highly capable to do something like that where, you you might feed it an IEP and say, focus just on the cognitive testing results, the ultimate, you know, educational classification and the primary accommodations or interventions that are going to be in place for the student. And it can go do that. IEPs are particularly dense, as I think many of us know, and you don’t want
summary of the entire thing because it could be overkill. So you can tell the software how to summarize those records. So the answer to that question is yes, I think that’s very, very doable. Limitations can be, you know, if there’s a ton of imaging or just images in the records, those can sometimes be difficult to synthesize and analyze for AI. But otherwise, I think it does really well. It can get lost a little bit in higher page counts, you know, so if we’re looking up at like
Dr. Jeremy Sharp (12:45.806)
you know, 500, 800, a thousand pages, things like that. You do have to be pretty specific with, in my experience anyway, pretty specific with what you want from it.
Y’all know that I love TherapyNotes, but I am not the only one. They have a 4.9 out of 5 star rating on TrustPilot.com and Google, which makes them the number one rated electronic health record system available for mental health folks today. They make billing, scheduling, note taking, and telehealth all incredibly easy. They also offer custom forms that you can send through the portal. For all the prescribers out there, TherapyNotes is proudly offering ePrescribe as well.
And maybe the most important thing for me is that they have live telephone support seven days a week. So you can actually talk to a real person in a timely manner. If you’re trying to switch from another EHR, the transition is incredibly easy. They’ll import your demographic data free of charge. So you can get going right away. So if you’re curious or you want to switch or you need a new EHR, try therapy notes for two months, absolutely free. You can go to the testing psychologist.com slash.
therapy notes and enter the code testing. Again, totally free, no strings attached. Check it out and see why everyone is switching to therapy notes. The Personality Assessment Inventory Bariatric or PAI Bariatric compiles the results of the PAI into a useful report just for bariatric surgery candidates. Log in to PARI Connect, PAR’s online assessment platform.
or can administer the PAI and select the bariatric score report to view results based on bariatric presurgical candidate norms. Learn more at parinc.com backslash P-A-I-B-R.
Dr. Jeremy Sharp (14:37.006)
Another question that came up is how are people using AI with existing report templates and what’s a smart way to structure that workflow? It’s a great question. I think there are several ways to do this. The simplest for me is just using AI a la carte. This is not the, I would say, most efficient way, but it is the simplest. So this is what a lot of people are doing where you just redact the PHI or protected health information and you paste
you know, put that into the AI and paste the output from AI back into your report. I would argue that this isn’t actually like using AI with your existing template. It’s more just, you know, cobbling together an AI solution and then kind of copying and pasting, which like I said is pretty inefficient. The next step up is you could do something like using Microsoft Copilot within your Word document, your report template, though just a
Asterisk here, mean, you absolutely need to have the HIPAA compliant version of Office to do that. So that’s the version where you sign to BAA with Microsoft. But you can, know, that copilot icon pops up next to every line that you’ve written in a Word document and you can give it prompts and have it do things for you within the document. So that’s an easy way to utilize AI within your report. You can create templated prompts and sections and something like bastions. So
You know, that’s maybe the next step up is using something like Bastion, a GPT, which is a HIPAA-compliant AI language model wrapper that utilizes the chat GPT technology. But you can’t upload your template per se to Bastion, at least at this point when I was planning a podcast. You can create templated prompts and sections and save those. But as far as I know, you cannot upload your actual report template. And then there are other
platforms out there that do allow you to use your own template. some of the AI solutions out there that I will call like all in one AI options are out there. Our platform, Reverb, know, specifically builds your template into the software so that your output, you know, matches the headings and the layouts that you’re used to. So I’d say that’s kind of the pinnacle, like the top of the mountain, so to speak, if you are trying to integrate AI into your existing template and, you know, have it exist within.
Dr. Jeremy Sharp (17:00.514)
the software platform that you are using. So there are lots of different options here and you can pick and choose depending on what works best for you. But again, just be mindful of the PHI and security component and know that you are not allowed to paste client information into chat GPT or Claude at this point. Next question that came up is, are there any examples where people have used AI to adjust summaries for parents versus teachers versus teens? This is a
Definitely. In fact, I think this is one of the best uses of AI. This is also place where Dr. Angoff has a ton of materials and options to facilitate this process. So you can use AI to adjust the reading level or summarize information for a particular audience really easily. Liz and many others, I think, are using AI to, for example, generate stories for kids to explain their diagnoses.
generate fables, map the results into kind of like a Star Wars arc, you know, that kind of thing. These are fantastic ways to utilize AI. We should not overlook. So again, on the software side, you know, our software allows for adjusting tone and reading level and many other variables just based on the client or the audience. And it’ll also generate a visual report that simplifies results that you can share with other parties as well.
So the answer again to this question is absolutely. And if nothing else, I think this is a great way to utilize AI to kind of translate, report results, or evaluation results to different audiences and come up with metaphors and analogies and illustrations and stories and things like that to really bring the results to life. There are many other examples of this. Like I said, you can visit Dr. Liz’s website.
which I will link in the show notes. And there are other resources out there as well to get more ideas around this. That leads me to the question of, you know, how can AI help us generate visual summaries? And there are many ways to do this. You can do Dr. Liz’s handout generator. Again, you can use an all-in-one platform like ours and maybe some of the others. I’m not totally sure if the others do this, but, you know, generating a visual report or presentation that will help translate those results and help people understand a little bit more.
Dr. Jeremy Sharp (19:22.712)
So those are many ways that we could use AI within the report writing space. I like to think of AI outside of the report writing space as well. And there are so many ways to apply it in the business realm that we could talk about. So first one that comes to mind, my friend Uriah Guilford, who’s been on the podcast many times, he’s from Productive Therapist, at this point has built what I would call a fully responsive AI receptionist for mental health practices using voice technology and custom prompting. We haven’t put it into practice yet, but
We are messing with it behind the scenes and it is pretty awesome, I have to say. So I don’t know if it is publicly available yet. I think he might still be in beta when this is released, but that’s something to keep an eye out for. And I’ll definitely try to promote that once it is totally live. Another tool that I’ve been using a lot is Notebook LM. So this is a tool that’s built into Google workspace that essentially allows you to build a custom language model.
and other resources using only your specific sources. whereas most language models pull all of their information just from the internet, Notebook LM allows you to build your own sort of custom language model, again, using only your sources. For example, if I’m doing research for a podcast, I could add peer-reviewed articles, YouTube videos, other podcasts, web pages, all those kinds of things to that particular notebook.
and then have the AI summarize only that content and generate questions from it. It’ll even, I think a feature of Notebook LM that is pretty unique is that it will automatically generate a podcast episode based on the content to summarize and communicate that information. You can also do mind maps from the content. It’s really cool and really powerful and I think kind of an overlooked tool for a lot of us. I’m still…
You know, I talk to folks and many people have not heard about or really utilize Notebook LM. So if you have Google Workspace, it’s built into Google Workspace and you just have access to it already. So go check that out. And it’s also great for a lit review type of task where you’re trying to learn more about a clinical topic and you just, you know, dump in a ton of articles and ask for summaries and ask questions and so forth. You can also use it to develop a curriculum or syllabus based on specific sources of information. If you’re in academia,
Dr. Jeremy Sharp (21:43.65)
that kind of thing. So I think there are a lot of uses for notebook LM. Another thing I used AI to analyze our intake scheduling workflow and identify any points of friction and then suggest solutions for those points of friction. Love that. I use it to help write emails all the time. This is one of my shortcomings is when I’m like pretty fired up or having a strong reaction to an email, I will dump my reactions and thoughts and the original message into
an AI program and have it rewrite a response in a particular tone that’s usually a little more friendly than the tone I want to use. And so that helps quite a bit. I’ve used it to generate quote unquote stock photos from my website. I’ve used it to, you know, upload all of my practice P &Ls from the last few years and had it analyze.
trends to find ways to optimize my finances. These are just a few examples. I I think at one point I created a birthday card for my wife. So, you know, there’s lots of different ways that you can utilize AI and these are obviously just a handful. But these are some of the things that I found most useful outside of the report writing realm. So the question I want to close with is what are some simple ways for people to get started if you’ve never touched AI before? And again, I know there are a lot of you out there.
Maybe you have like literally, you know, messed with it, like logged into chat GPT or like asked a question or two. And I went through this exercise myself just to see what would happen. But I would advocate spending 30 to 60 minutes once a week with chat GPT or Claude. I think they’re both great. They have different strengths. Doesn’t matter. Don’t get into analysis paralysis around which one to use. Just go to chat GPT or Claude and you can type in, I am a total AI newbie.
you are helping me learn AI as quickly and thoroughly as possible. Guide me through some questions and exercises to help me learn how to use these tools. And from that point, it is off to the races. And when I did this exercise, we ended up, we, me and chat, we ended up down a pretty productive rabbit hole and actually ended up discovering, you know, a few ideas and topics that were new to me. So I think that’s a great way to get started.
Dr. Jeremy Sharp (23:57.07)
You don’t have to pay for anything necessarily. Just do something and let it guide you through the process. So with that, we will close. I hope this has been helpful for you talking about practical wins, but let’s be honest, there’s kind of a dark side, I suppose, and a lot to be cautious with in terms of AI. And so next episode, I’m going to be talking about ethical questions and just how it changes the way that we might teach the next generation of clinicians.
is very relevant for us having interns and postdocs in our practice and I think for many of you as well. So the next episode we’re going to talk about protecting clinical thinking, maintaining trust and just avoiding, you know, turning into testing mills, I suppose, in the age of AI.
All right, y’all, thank you so much for tuning into this episode. Always grateful to have you here. I hope that you take away some information that you can implement in your practice and in your life. Any resources that we mentioned during the episode will be listed in the show notes, so make sure to check those out. If you like what you hear on the podcast, I would be so grateful if you left a review on iTunes or Spotify or wherever you listen to your podcasts. And if you’re a practice owner or aspiring practice owner,
I’d invite you to check out the testing psychologists mastermind groups. have mastermind groups at every stage of practice development, beginner, intermediate, and advanced. We have homework, we have accountability, we have support, we have resources. These groups are amazing. We do a lot of work and a lot of connecting. If that sounds interesting to you, you can check out the details at the testing psychologists.com slash consulting.
You can sign up for a pre-group phone call and we will chat and figure out if a group could be a good fit for you. Thanks so much.
Dr. Jeremy Sharp (26:07.234)
The information contained in this podcast and on the testing psychologist website are intended for informational and educational purposes only. Nothing in this podcast or on the website is intended to be a substitute for professional psychological, psychiatric or medical advice, diagnosis or treatment. Please note that no doctor-patient relationship is formed here. And similarly, no supervisory or consultative relationship is formed.
between the host or guests of this podcast and listeners of this podcast. If you need the qualified advice of any mental health practitioner or medical provider, please seek one in your area. Similarly, if you need supervision on clinical matters, please find a supervisor with an expertise that fits your needs.
Click here to listen instead!