I’m applying for SouthParkCommons (-1 to 0 incubator) and am workshopping my application for their founder fellowship.
I’ve been working on Interviewssuck.com, a prototype (for v1, I geared it towards interviews) of a tool that coaches on conversations based on psychology research. Here is a screenshot! Yes, our recommendation algorithm is fairly simple at the moment.
A common misconception is that human conversation is just innate “chemistry” or “charisma.” Mountains of research tell us this misconception is wrong, and explains how there is an optimal and more effective format to have a better conversation in specific contexts, like work, or working through a disagreement with a romantic partner.
In romance, John Gottman studied couples for over 40 years and is able to predict with a > 99% accuracy, the chance that couples will stay together based on the presence of certain key conversation patterns.
What if we could represent his (or comparable) research as well-designed product that could meaningfully improve the EQ of the average user?
With that said, here is my application in its current form:
For each team member, share a product, project, or accomplishment you are most proud of and why.
Building iCantSpell, a spell checker Chrome Extension for Gmail that identifies and fixes misspellings of names and companies contextually (see pitch to Grammarly, below)
I’ve always have both a passion for entrepreneurship (ran DJ business in college) + resilience (failed CS in college) + later restudied CS via a coding bootcamp, almost failed out again, and got hired for software engineering at Google. iCantSpell was the first time I channeled those qualities to make a net new product that I use daily using recent tech.
How long has your team been actively exploring starting a company? Why did you decide to work together?
Sahil and I have been talking about company building (first casually and now more seriously) since 2018, when I moved to San Francisco (he also lived in San Francisco then) to work as an engineer at Google. We decided to work together for a couple of reasons:
He complements my expertise in product and design with deep engineering knowledge. He is very good at probing and making my (sometimes) idealistic ideas executable.
He is a contrarian and calm thinker, which contrasts with my extroverted and high-energy demeanor. I find our debates challenging but productive.
Like me, he comes from an immigrant background and has one of the strongest work ethics I've ever seen.
What ideas are you exploring?
I’m exploring AI models paired with speech to text that ‘coach’ conversations based on psychology research where there is an ‘optimal’ format. Examples here are managing, conflict resolution in couples, listening in general, sales, to cite a few. I built a prototype for interviews called interviewssuck.com
Existing coaching solutions fail. Therapy does not fix and address conversational patterns in real-time, or provide the needed practice and repetition to learn strong emotional habits.
Habit based apps provide repetition, but lack customization, feedback, and focusing on actual conversations between multiple people, which is where dysfunctional communication patterns usually lie.
For instance Ahead (leading emotional coaching app) focuses on habits like emotional regulation, and anxiety, but does not focus on conversations. Reading a list of tips and best-practices does not meaningfully help apply them because there is no feedback loop.
Trained LLMs can give specific feedback AND do so in real-time.
What expertise do you have that is related to the ideas you are exploring?
As a neurodivergent, I had to learn ‘unwritten’ social rules in a more formulaic way.
While challenging, it led me to realize that EQ can and must be coached. I both have the technical skills to build the core product + the knowledge around the user-experience of breaking ‘EQ’ into more formally defined logic.
I’m partnering with therapists to understand, fact-check and prioritize mental health research.
What progress have you made?
✅ I hadn’t coded in 3 years, so I … learned Python, React, Next, AI engineering in a couple months
✅ Built and shipped a prototype: interviewssuck.com. I started with interviews for a simple use-case to play with the technology.
🔄 (in-progress) User testing with technical interviewers on Exponent
🔄 Product research to determine the highest ROI decisions for:
the interface (will our product listen to conversations and coach vs role-play conversations and coach?)
types of conversations to focus on (for instance, sales vs work-place collaboration vs couples vs children with autism, to name a few)
competition: of the given conversation types, where are the biggest gaps not currently solved by competition (like the SPC company Speaksage
psychology research around conversations: the intent is to coach users to have higher quality conversations. That optimal ‘ceiling’ is determined by psychology research explaining how an ideal conversation would be. The more specific the research —> the better we can coach users
marketing strategy: What are the constraints around marketing to the above different segments? How would distribution inform the above choices?
How does the world look different if you succeed? How big is the market, or how does your idea unlock one?
We have countless missed opportunities for communication. We’re flawed communicators. We interrupt when we mean to listen. We’re too direct. We give feedback when we should be building trust. It’s hard & costly.
How many deals were missed because of unwritten social cues? How many missed dates? How much miscommunication slowed down teams?
Better conversations would help us tremendously. For instance, to take one segment we are considering, imagine if the U.S divorce rate were halved from 42% because couples learn to communicate and listen better.
Tech isn’t helping much with the above (currently). We can optimize the first 10 seconds of a TikTok using software, but the quality of the most important medium — human conversation — remains largely unchanged.
A common misconception is that human conversation is just innate “chemistry” or “charisma” that some people either do or do not have. This is wrong.
Mountains of research explain how to have a better conversation in specific contexts, like work, or working through a disagreement with a romantic partner. Example: John Gottman: studied couples for over 40 years and found conversation patterns that predict whether a couple will be together 15 years later with a 99% degree of accuracy.
What if we could represent his (or comparable) research as well-designed product that could meaningfully improve the EQ of the average user?
Let me know of feedback you might have! All the best. ~ Sasha
Hi Sahil - saw your post from FRCTNL. My advice from the other side (done a bunch of accelerators & YC, not specifically SPC) is:
- how does this team build conviction? you mention past accomplishments, and seem fairly high conviction. during the interview I'd try to dig into your projects to understand this and how you move
- what do you agree on as cofounders? if you're debating quite a bit but still working together, I'd want to dig in on that
other feedback
- congrats on the prototype / you guys def show the hustle
- SPC is builder-focused, so would want to see an interest into the "science" behind conversations & further work on LLM tuning / engineering. What is the gap btwn current coaching in terms of tech? Are state-of-the-art prompting techniques not good enough? Why not?
- user feedback: how do ppl feel about getting coaching during interviews? how do YOU feel with a tool giving feedback? I'm curious how the advice lands with people, esp knowing it comes from AI. Part of me is skeptical, but part of me is also very excited and eager to see how AI can noticeably improve my quality of life
Best of luck!
thank you Kai! Sahil is my co-founder :)
this is great feedback