Office Hours: I’m the CEO & co-founder of a digital behavioral health platform. I’m Angelia Muller. AMA!Featured

Posted in:Mental Health Office Hours Fundraising Startups

Hi Elphas!

I’m Angelia Muller, CEO and co-founder of Attunement AI, an assessment platform for better mental health.

Before Attunement AI, I was a partnership lead at Stripe and NVIDIA, working with ML startups to accelerate payment processing solutions adoption, synthetic data generation, data labeling, and annotation solutions.

During my downtime, I enjoy meditation, making digital art, and pilates.

Ask me anything about building your first startup, fundraising from investors you like, building products you are actually passionate about.

ElphaStaff·Jul 08

Thanks so much for joining us @AngeliaM!Elphas – please ask @AngeliaM your questions before Friday, July 12th. @AngeliaM may not have time to answer every questions, so emoji upvote your favorites 🔥👍🏾➕

anyakern·Jul 13

Attunement is on an awesome important mission! We have the same downtime preferences :) I go to Bodyrok on Haight. Lmk if you ever want a pilates buddy or explore a new studio

madhuchavva·Jul 12

Hi @AngeliaM, 1) What's your strategy to get the first few customers? 2) What's your suggestion to those female founders who have zero entrepreneurial experience but has the zeal & potential to be an entrepreneur? How do they go about fundraising?Thanks for taking time to inspire us with your invaluable inputs.

MandyWong·Jul 11

Hi @AngeliaM - thanks for being here. What was it about Attunement AI that inspired you to leave more established companies like Stripe and Nvidia? What are some of Attunement’s competitors in the market place and what does your brand awareness strategy look like? I’m also interested in the mental health space, so would love to know your thoughts. Thanks!Mandy

AngeliaM·Jul 12

Hi Mandy! Nvidia and Stripe are really great places to work at. Some of the best people I know still work there, and I would recommend the companies to anyone who's interested in working there. If it wasn't for Attunement, I'd still be quite happy and fulfilled at either of those places. Attunement feels much more aligned to me in terms of its mission, the problems it could solve, and it encapsulates many facets of my intrinsic interests. I am forgoing a lot of establishment and stability for intrinsic motivation, and there certainly is a good amount of risk-taking involved; it just didn't seem like it was a real "choice" at the end, because how much more appealing building Attunement is to me.We're 7 months old, and are building assessment software that complements a lot of existing workflows, so I can't call out any competitors that are taking the exact same approach that we are. In terms of brand awareness, it is providing a lot of value up front, and aligning with reputable organizations that have quite a bit of influence. Hope that helps!

Moriah172·Jul 11

Hi angelia! Thanks for your time. What are some of the ethical considerations when developing AI for mental health? Thx!

AngeliaM·Jul 12

Hi Moriah, thank you for your question! I'd approach the question with the most relevant dimensions we're personally thinking through right now. On a practical level, it's keeping up with HIPAA compliance, ensuring that BAAs are signed for any third party when dealing with sensitive information. Reliability and accuracy are also critical elements: is the AI system reliably and consistently returning accurate outputs that humans can rely on for their workflows. In mental health, there is already a lot of subjectivity. So could AI back up claims with more evidence? If so, you'd also want to think about transparency and communicating your process to end users.You'd want to avoid blackbox AI - you'd want to explain to people how the AI works, how it's generating the outputs, so people can trust that the system is accountable and transparent. You also want to make sure you are getting explicit consent from the end users (patients or clinicians) and that they are informed about what data it collects, and how it's being used.

Ila85·Jul 11

What were some of the biggest challenges you faced during your fundraising journey? And what qualities did you look for in investors?

AngeliaM·Jul 12

We're going through a pretty slow season for mental health investing. There is general investor fatigue with mental health in general, since a huge wave of companies emerged during the Covid era. We're seeing fewer generic mental health applications (like democratizing access to therapy) to more niche solutions that specifically target a patient demographic or disorder, or building for SMI applications (Serious Mental Health).Specifically with AI, there hasn't been any clear winners for the applied space (eg in AI therapy), even though LLMs have launched for a couple years now. My observation is that there is an overall sense of doubt/hesitation for any machine learning-focused applications in this space, because it's unclear how people would adopt a AI-first solution.Working with these industry dynamics can be challenging, so really taking the time to explain the thesis/strategy was key. In terms of qualities, I really value thoughtfulness, and advice from lived experience/observations. Investors who really take the time to listen and understand our approach are really great - you can tell they're very much invested in the thesis and the success of the business.

Shenna20·Jul 11

Hey Angelia! Love what you're doing over at Attunement AI, what inspired you to start a mental heath startup?

AngeliaM·Jul 12

Hi Shanna. Without sounding overly kitschy, this felt more like what I was getting ready for in the last couple years; I mentioned in another response that this all started at a regular hackathon I participated in last year. The code, the problem the product solves, the go to market etc, all kind of "clicked" in one day and my body was like: this is absolutely what you should be doing for as long as you can. I don't think I spontaneously combusted into an entrepreneur overnight, but rather many skillsets/learnings from disparate jobs and projects came to a head then. Psych and neuroscience have always always been my number one passion (and education) in terms of their goals in promoting human flourishing, and I have always stayed current with its papers, and trends in the industry. However, I could not find any careers within the fields that felt right for me (I didn't love the day to day of therapy, dispensing medication, or doing research), and my brain always gravitated towards systems thinking. A career in tech in different roles has really suited my particular brain structure. Looking back, I have always tried to find a pathway back to mental health, and looks like I'm figuring it out.

anonymousvictoria·Jul 18

Hi Angelia. Thanks for be willing to offer advice! I see that you were not a tech person but is starting a tech company startup, what would you say is some of the challenges in that? And would you spend lots of time reading ML papers or something to try to understand the technical aspects of things?

DrKingaMnich·Jul 09

Angelia, did you code your prototype, or did you hire an engineer? I have no tech background and always wonder how people do it without funding or technical experience.

AngeliaM·Jul 10

Hi Kinga, great question, and one I hear more often from non-technical female founders! I built all the prototypes we've ever demoed, and my co-founder recently built a fairly more complex transcription tool that goes beyond the MVPs. I started picking up the technical skills for prototyping/iteration the last few years (which goes beyond code), because I was getting quite frustrated by having ideas in my head that wouldn't translate into the real world unless I relied on someone else. I also relied heavily on no-code tools for iteration, and a lot of very technical folks do too! It speeds up execution, and tools have gotten quite advanced in the last two years.A helpful framework that may help here is that being technical isn't really a binary yes/no thing - technical skills lie on a spectrum, and entirely dependent on previous projects that you have worked on. Having an idea for what you want to build is a great starting point, since that gives you a scaffolding to know what you need to learn, and gives you a map for what is required to build it. I am definitely not saying you should switch career into engineering, but it is critical for you to grasp the main concepts and know how the product works on a fundamental level: even if you end up hiring a team or getting a technical co-founder.In case it's helpful: a front-end eng would not be able to wake up one day and do back-end, and a person who has full stack experience doesn't actually mean they're great at everything or can rapidly prototype and iterate. Founding a product-led company actually requires a lot of continuous learning for everything; having a technical background sets you up for the ability to learn, as new skills need to be picked up all the time.What would really help you get started is to grasp the basic foundations for building an application, and remember that what you don't know, you can always learn. Also, try to start cultivating a network of growthy people around you who can jump in and help/motivate you. Best of luck!

elisegatsby·Jul 09

Hi Angelia! Thanks for taking questions! I'm curious about the similarities between two seemingly distinct industries: finance and mental health. What skills crossover have you experienced or witnessed? When building out a team, are you more likely to look to one industry or the other?

leeshiela·Jul 08

@AngeliaM, thanks for your time! When and how did you make the leap to founding your startup? Did you wait to see if there was market fit by doing surveys of your target population, or did you begin to build your MVP and then ask for funding? What was the process like in the very beginning of building your startup? I'm trying to determine when to make the leap for a startup idea that I have and would appreciate hearing your story.

AngeliaM·Jul 09

Hi @leeshiela! I love this question and I'm so excited for you. The short and sweet answer is get out there and talk to your prospective users. If you can sell it and get paying users (you can do this before building out the full product) - you have a startup! But let me share my story really quickly.I went to a mental health hackathon sponsored by an investor I really liked last year, and built what I really thought was required in the space. It was just a hacky prototype that was made in a day, but I got an overwhelmingly positive response from this community (who were mostly engineers with a few psychologists who graced us with their presence) - and that gave me _some _confidence that there may be something to what I was building.I was still at Stripe at the time, but I couldn't shake off the feeling that I may be onto something. So after work and before work hours, I would use the same deck I built at the hackathon and bugged many friends who were shrinks/psychiatrists, their friends who were shrinks/psychiatrists, and I wanted to get in front of them and get their feedback. Everyone seemed to love it! I was so excited. But! Turns out I really was not asking the right questions- I knew nothing about user research then, and didn't get a single sale. I wish I knew about the book the Mom's Test--it's a book that hypothesizes that your mom loves you, so she will always lie to you about loving what you build. So you want to ask the real questions to reveal whether these users would really buy what you're building.With this assumption that people really wanted what I built at the hackathon, I applied to YC and got in! The partner knew the product wasn't quite ready, but was convinced of the thesis. So, I spent another few months in YC actually asking the right questions, iterating on the product a bunch, I finally began selling, and landed my first paying customer. This led to my first investors.I'd say there is no linear path in the beginning - the main insight here is to really talk to your users and get someone to buy something from you. Some people decide to jump once they get here, and others decide to do that full time. I didn't get a paying customer until after I made the leap. Hope this helps!

remoterebel·Jul 09

Really cool business @AngeliaM! I know a few people that would love to work in this sector. Let me know if you have plans to hire in the near future and I'll happily make the introductions :)

AngeliaM·Jul 10

Thank you Michelle! Let's keep in touch so I can reach out when the time is right. :)

remoterebel·Jul 11

Lovely. Very best of luck with everything :)

suzhang·Jul 08

Hi Angelia,Great to meet you here! I noticed even you've had education in psychology/neuroscience field, you've never worked in mental health space, how did you discover the problem you're solving at Attunement AI? Also, how did you overcome the technical challenge to build the product?

AngeliaM·Jul 09

Hi @suzhang! That's a really great question! Through schooling and my other social networks, I have kept a pretty robust network of friends in psychiatry and psychology. Many are academic practitioners, or they have a private practice. When I started, I asked anyone who would be keen for me to moonlight them- and I had two friends who were kind and literally brought me to work, who I badgered for months on end, and I'm still in contact with them. I did not have access to their patient conversations, but they would debrief their experiences with me afterwards and walked through all the tools and resources they are (and are not) using. Without those experiences, I think it would taken a lot longer to understand who I'm building for - I'm really grateful for those initial people who later converted to my first customers.I just went through YC, an incubator program with 300 other companies -and there were a lot of founders who were outsiders of the fields they were building in, so it's not that uncommon, and it has been done more often than you think. I know one team who literally got second jobs during YC just to moonlight to fast track an understanding to the people/problems they're building for.I'd say humility and being user-centric are the most important elements here - we definitely know nothing going in, and also - we can solve problems by really listening to the end users, and leveraging experts in the space to help you navigate.Technical challenge - there are so many tools for easy prototypes/ easy iteration until you hit features that really resonate, so starting out is so much easier now than it used to be. But also - know your strengths and hire the rest! One of the key responsibilities for founding a company is recruiting a world class team, the product is really as good as the team. So make sure you always keep tabs on the smartest person you know/have worked with, and maybe you can convince them to join you. :)

laurenvriens·Jul 08

Can you tell us more about Attunement AI? What are you building technology-wise? I love the idea of more data being applied to the mental health space!

AngeliaM·Jul 09

Sure thing! At Attunement, we help psychologists do behavioral health assessments with AI and cut their waitlist by 50%.We are only focused on assessment psychologists, who are a subtype of psychologists (they can diagnose, so are usually PHDs/PsyDs). The main product has two components: a software platform for patient coordination, and a HIPAA compliant scribe/transcription for speeding up the assessment workflow. We picked this specific space because it's a *really* good use case for large language models - all neuropsychological assessments have consistent coding/outputs, and you're taking a lot of unstructured data (clinical interviews of patients) and asking AI to code them into the right categories.

noushink·Jul 10

Hi @AngeliaM, thank you for being willing to do an AMA. Attunement AI is such a great product and I have recently been brainstorming a similar product, focused on women's health, in Canada. We have a different medical system altogether and no real private system but waitlists are through the roof and access to knowledgeable specialists to support perimenopausal women is rapidly deteriorating. As a woman entering her 40s, the struggle is real. Gen AI has the potential to help re-prioritize waitlists. Do you have technical co-founders, and if so, how did you guys form a partnership (pre-existing relationship or outreach)? How did you target which assessment psychologists you would go after and how did you get them on board? Did they already have patient coordination platforms and did that pose a barrier? What are your biggest challenges as a business, both technical and otherwise? And, is there anything you are particularly excited about regarding the future of generative AI and behavioral health?

AngeliaM·Jul 12

Thank you @noushink! The space you're building in has a ton of opportunities, simply because the perimenopausal space is often overlooked--so huge congratulations for being astute in your observations. I'll try to respond to all your questions, but contact me at [email protected] if I don't get to all of it.I do have a technical cofounder, and the initial partnerships were design partnerships (see here; https://a16z.com/a-framework-for-finding-a-design-partner/), where the initial partner got the software for free if they'd give us tactical feedback. You'd want to actually sell the software eventually, but when it's unclear what features you want to really hone in on in the beginning, design partnerships are a great way to figure out what to build.For assessment psychologists, there has been a ton of trial and error. It's a fairly fragmented space, as many do not have patient coordination platforms/EHR, and some are still doing everything with pen and paper. The general advice that I heed is to hone in on one specific feature that is really needed in the market, and build from there.Biggest challenge: I'd say the human oversight/ manual review of a lot of AI generated outputs. Scaling manual review is unsustainable and I fundamentally believe there is a startup that could solve this for many of us.Future of Gen AI and behavioral health: I believe we are at very early innings for what I believe to be a very transformative journey for the mental health industry. Let's take therapy as an example. If you think of what genAI does best: extracting information from unstructured conversations and other sources, and consolidating it for analysis, I simply can't think of another use case that sincerely needs this. Therapy is inherently conversational can become a lot more effective with real time, evidence-based insights.

noushink·Jul 15

Thank you so much for your response @AngeliaM and for your willingness to continue the conversation via email.

share this article