Hi there! Today I want to share a bit more about Iro - the project I started at Antler, with my cofounder Nalla.
Before I dive into that, a quick note on the new blog frequency: I’ll now try to share bi-weekly updates, instead of weekly. You can also now subscribe to be notified when there’s a new post. Just add your email here (you’ll receive an email asking you to confirm - make sure you do 🙂)
🤖An introduction to Iro
Why build yet another (non)therapy chatbot?
First, I need to share that Iro will start as more of an emotional companion that is trained with therapeutic methodologies. Claiming something is actual therapy is heavily regulated in most countries, and takes certification work for which the regulation is not even there yet.
Moving on, my and my cofounder’s motivation to work on this is simple. We are both passionate about how people think and feel, and how they can do this in the healthiest way possible. Mental health is something we deeply care about - both for ourselves and for others. If everyone’s mental health was even 1% better, we’d be living in a wildly different world 🤩.
But we’re living in a world where at least an eight of the population live with a mental health disorder. People want to get help, but almost half are prevented from getting it either by cost or stigma.
We don’t aim to replace existing solutions such as therapy or medication. But we want to create a high-quality, affordable solution that can help the people whose access to care is currently limited. We believe that the latest advancements in artificial intelligence (i.e. large language models such as ChatGPT) allow for a much better solution that the existing therapy-like chatbots.
Progress so far
Our first milestone is to validate Iro’s value proposition. We ran around 40 user interviews with potential users as well as with experienced therapists to understand how such a solution could look like. They told us about their struggles with mental health, the solutions they have tried, and how they would envision using an AI emotional companion.
Based on these learnings, we launched a survey with 100 people to confirm and quantify these learnings. This helped us go from “people want/struggle with Y” to “x% of people want/struggle with Y”. We used this to focus and refine our messaging, and build mockups for how Iro could look like.
At this stage, a validated value proposition means a reasonable ratio between the cost to acquire a customer (CAC), and the revenue this customer could bring. We can estimate the revenue the revenue part with publicly available data. To learn about the CAC, we’re now running an ad campaign asking people to sign up to Iro’s waiting list. The cost we’ll end up paying/signup will be a proxy for the CAC.
Check out one of our ads below, and let me know what you think!
Next steps
If we manage to validate the value proposition, the next step is to build a proof-of-concept prototype. Something useable and valuable, but without all the bells and whistles. To do this, we’ll need to involve clinicians to help us design Iro in a way that makes therapeutic sense. Luckily, we’re already in discussions with some really knowledgeable (and excited) therapists.
We’ll then iterate this prototype until the end of year with the initial users that sign up to the waiting list. This will require further marketing and communication effort. To speed up development, we’ll also be looking to bring more people on the team, including a CTO who can both develop a mobile app, and work with large language models (hit me up if you know anyone)
In the next update, I’ll share more about how the marketing campaign went, and how (if) we’re building toward a proof of concept.
🗿Personal update: How’s it going, Stefan?
I’m very happy with the progress we’re making so far with Iro. Aside from that, I’m recovering well from my ankle sprain, so things are looking up for doing the Eindhoven marathon in October.