This African Founder Wants You Alive

Meet Therabot, Your AI Therapist

Welcome to The Afro Pivot Point

Showing You What’s Art and What’s Not in African Tech

Hello to the gentle readers joining us since the last edition. You’ll love it here ❤️

If you haven’t already, please subscribe and join over a thousand curious minds reading about tech and culture in Africa.

In honor of Suicide Awareness Month this September, we bring you an important conversation with Caleb Jephunneh, CEO of BrickLabs AI, who shares his personal journey with mental health and the inspiration behind Therabot, an AI-driven e-therapist designed to provide accessible mental health support. Caleb’s story is a powerful reminder of the importance of seeking help when struggling. Whether through professional therapy or innovative solutions like Therabot, there are resources available to those in need. If you or anyone you know is struggling, know that reaching out for help is a vital step. Mental health is a journey, and being brave enough to take the first step can make all the difference.

Hey, thanks for meeting with me today. Please tell us more about you and what you do 

My name is Caleb Jephunneh. I’m a software developer, startup founder, and entrepreneur. Currently, I’m the CEO of BrickLabs AI, where we leverage AI to solve some of Africa’s most pressing problems using a modern approach.

Nice. How long has Bricklabs been in operation? 

For about 1.5 years now. In that period, we’ve been able to execute the idea, grow our user base, and even received a US $150,000 grant from Microsoft in April 2024.

Congratulations! Do you work closely with the Microsoft team?

Absolutely! Apart from the grant, Microsoft has also helped us in other ways. They’ve given us free access to Microsoft’s resources, such as computing power, resources and expertise needed to scale the company, office space, and mentorship.

And how exactly does BrickLabs AI work? 

We operate in three ways, actually. The first is that we look for young founders with great ideas and help them bring their ideas to life by supporting them with research, technical support, and tools to run their business.

So, an incubator of sorts?

Yes. But we also make our own solutions, like Therabot – an e-therapist. And then we also integrate our services and products. Currently, we are a team of 21 and have three key departments: Research, Technical, and Marketing.

Part of the team at BrickLabs AI

Wow! And are you guys focused on specific niches?

Not at all. The only thing we focus on is a good idea that solves a problem practically.

I get it now. Let’s talk about Therabot specifically. Was it your own idea or an idea from one of the founders you support? What’s the story behind it?

Therabot was inspired by my own lived experience. I was going through a terrible heartbreak in January this year.

Happens to the best of them

It’s life. I was going through this heartbreak while also working full-time for a startup. I tried talking it out with friends and family, but I’m one of those guys who generally struggle to open up about my problems. I decided to look for a therapist, but the most recommended ones charged an average of KES 3,000 per hour, which was too high for me. More often than not, the appointment slots available would coincide with my working hours at the startup. So, as I kept drowning, I was desperate to save myself. That’s when I started thinking about how AI could help me solve this.

Wasn’t it easier to just look for alternative therapists with flexible hours and who were cheaper? 

(Laughs) It was. But honestly, I felt like I needed specific help. And believe me, I tried. It was so hard to find a good fit due to scheduling conflicts, cost, and biases like religion, etc.

What was it about AI specifically that gave you hope about building this? Why not a different technology?

My heartbreak just happened to coincide with the AI gold rush. I stumbled across some iterations of the same idea in other countries, but there wasn’t anything like it in the local market. So I started toying with the idea: How can we make an AI that can talk and act like a therapist? At the time, the main idea was to create an AI companion you can talk to, and only proceed to a therapist if you felt like the help you got wasn’t sufficient.

How did you start working on it?

After working on the MVP back in March, I mobilized developers, and we started working on this AI chatbot. We consulted with therapists locally to understand their thought processes and principles, and then we scraped the internet for data related to the mission we were on. We kept improving the idea until we got the opportunity with Microsoft in April.

That was quite an impressive thing to do in just a month. What happened next? 

With the money from Microsoft, we were able to host models, compute power, train different models better, fine-tune therapy data, get better access to peer-reviewed data, and use part of it for operational costs.

When did you go to market?

We rolled out the first version in April after a social media marketing run. Therabot was launched on WhatsApp as a chatbot. Mainly because, in Kenya, WhatsApp could still be accessed even after your data bundles ran out. It was, and still is, completely free.

I’m actually curious. How do you personally take care of your mental health? Apart from Therabot, of course.

I hang out with loved ones, disconnect from the internet, and I’ve recently gotten into meditation.

Nice. Why should people try Therabot? What are the benefits?

Not only is it free therapy that you can access without the time limits that come with real-life therapists, but it’s also an extremely personalized experience. In real life, it would typically take a couple of sessions before you finally connect with your therapist. The same is true here. The more you use Therabot, the more you tell it what you’re feeling and experiencing, the more personalized the experience becomes. It remembers you, personalizes answers, asks about previous interactions, and continues to fine-tune the output. Compared to the cost of a real-life therapist, Therabot is much more accessible, with close to zero cost implications.

Why should investors invest in Therabot?

If they’re looking for a health-tech startup with strong ESG principles, that has already overcome government barriers like licensing, operates in a competitive market with a lot of potential, and solves a real and pressing need, then Therabot is the perfect fit.

What do the numbers look like currently?

To date, over 800 people have visited our site, with 600 of those users being returning users.

There isn’t a one-size-fits-all approach to therapy. How do you ensure the bot doesn’t get mixed up and only gives relevant answers? How do you ensure that the users get exactly what they need? 

We have been heavily reliant on user feedback. Luckily, our users have given us valuable feedback that has allowed us to improve the answers with every iteration. Also, we are working on a new feature that will allow users to have different threads to avoid mix-ups.

Why should users trust you with their data? 

We operate within GDPR, and our privacy policy is also very iron-clad.

Well, no one actually reads the Terms and Conditions.

(Laughs) I get that.

But?

We only collect the data that will make the model better and useful. We don’t track any personal identifiers like phone numbers or stuff like that. And truly, most of our models are improved based on the ratings the user gives on the feedback forms. So by the time we are using this information to improve the models, the users are fully aware of how we use this information. We want to make it as easy as possible for users to state their problem and get appropriate help while protecting their anonymity. So we only collect information like what most users are going through, the symptoms they report, and so on. We use this to rate the answers they got and then fine-tune it.

So based on the user patterns and data you have, what issues are people mostly seeking help for from Therabot?

Relationship issues and navigating romantic breakups are the most common reasons.

Is this deliberate or connected to your own experience? Is Therabot being understood as a relationship therapist by its user base? Or was it built that way? 

That might be true. But the goal is obviously to make it more cohesive and cover a larger number of issues such as depression, anxiety, and other issues like addiction.

A strong tenet of therapy is the human aspect. People want to feel seen, heard, and have that human connection. What attempts have you made to humanize Therabot and make it sound like my human therapist? 

The responses are modeled to sound like a physical therapist based on our consultations with therapists. We are still working on making the model more human-centric though. And we also try to bring the human approach by doing physical events where users get to interact with Therabot and the team in person as they share insights. Also, just like in physical therapy settings, Therabot is able to recommend a certified human therapist based on your presenting symptoms—should you not be satisfied with the help you get.

Do you think Therabot will get to a point where it can solve the user’s problems effectively without needing to refer them elsewhere? 

Absolutely! Our research team has been working tirelessly to help us get there sooner. We also have amazing therapists that came on board to help in the research process.

What kind of therapists are you working with?

For now, relationship therapists and those versed in this area. Obviously, this is hugely influenced by the user feedback and patterns, and we want to make sure the feedback given is as human as possible. But, we are also looking to work with therapists who focus on other niches so the model can be more helpful.

I love that. So can therapists who are interested in this cause join you and how can they offer their expertise?

They can choose to volunteer for the research through our website or participate in the training programs we hold occasionally for our users. Hopefully, we will get to have certified medical professionals on the team in the long term.

What services are you planning to introduce to make Therabot more accessible and effective? 

We are rolling out a voice feature soon so that users can have a conversation if they prefer that over the chatbot. We plan to also integrate it with other health apps and mood trackers to be able to better predict what the user is going through based on data across platforms. Sometimes, you might just be sleepy or hungry, not grumpy. So we want Therabot to be able to comfort you and assure you of the same, which means better responses over the course of time. Without saying too much, Therabot will soon be available across wearable devices, smart home gadgets, and USSD. Obviously, the goal is for Therabot to reach as many people as possible globally.

Currently, how can users access you?

They can visit our website and try it. It will redirect them to WhatsApp, and it’s pretty straightforward. However, we will be rolling out our mobile app by October 2024.

What can you say about access to mental health care in Africa?

Most of our research has been conducted in Kenya, so I can only speak about Kenya for now. Currently, there are about 2000 licensed and registered therapists in Kenya. For a population as big as ours, I think that is very imbalanced. Also, the cost of therapy isn’t very friendly to the average Kenyan. And because of that, people continue to treat therapy as a luxury compared to the vital need that it is.

But do you think how we talk about mental health has changed? 

Yes, but we still have a lot to do.

What do you wish you knew before you started building this?

Users are everything. It doesn’t matter how shiny, cutting-edge, and groundbreaking your product is. If they hate it, they hate it. You have to take their feedback seriously.

What do you want to see in the African health-tech space?

More technologies, more founders, more products—more of everything, basically. There is still a lot in healthcare that needs solutions, and I think a huge part of those solutions will be technology-based.

What are some of the challenges you faced as you tried to scale this?

Apart from not talking to users enough, I would say the usual systemic issues in healthcare.

How do you mean?

The fact that there are only 2000 registered therapists in a country with such a huge number of educated and unemployed graduates screams FAILURE. The government is failing its citizens and is highly contributing to the rise of mental health cases because of its role in things like unemployment, inflation, and lack of access to basic needs.

In an ideal world, the government should be at least trying to make sure citizens have access to proper healthcare. We don’t live in an ideal world. And so we built Therabot. Also, it has been a nightmare trying to get licensed, especially for those related to data privacy. The signatories are notorious for asking for bribes to give you the license. Luckily, NGOs have been very helpful in bridging that gap and helping us get these licenses.

What should the Kenyan government do to encourage innovation in healthcare?

They should make the space more lucrative for innovation and investment by making processes easier. It would also be nice if they managed to deal with these corrupt officials.

Have you thought of partnering with private and public hospitals to increase access? 

Yes, talks are underway to actualize that.

What other technologies can be leveraged by founders looking to join this space and make therapy more accessible? 

Blockchain and Web3 can solve this in a way. There are several use cases. I would also say there is a lot of potential in AR and VR in making the models more interactive. But still, AI has a lot of possibilities too.

How can people support Therabot?

The biggest way right now is by interacting with the model and giving us feedback so we can use it to improve. You can also support us by referring Therabot to your friends and family or by donating through the M-changa link on our website.

What’s your hope for Therabot in the next couple of years?

My hope is that we will be able to provide free, reliable, and accessible mental health support to anyone in the world.

I’d love to hear your thoughts on Caleb Jephune’s vision with Therabot. How do you see innovations like Therabot transforming mental health support in Africa? Share your insights and experiences with us!

Talk to me: [email protected]

  •  Follow me on X

  •  Connect on LinkedIn

  •  Reply to this email

    And if you haven't already, please subscribe to get this in your inbox next week. Till next time, cheers!