InTouch

Inspiration

Initially, we noticed a few pain points when it came to outreach/automated outreach for companies. Namely:

Outreach content remains generic ⟶ fails to resonate
Even when personalised, outreach is timed poorly ⟶ low return engagement
Current marketing tools take days to update their data ⟶ execution is based on outdated information

We wanted to fix these problems, so we created InTouch.

What it does

InTouch creates, personalises, and pushes outreach content for businesses based on real-time data.

How we built it

Using an RAG architecture (with parallel auto-updating vector databases for faster information retrieval), as well as Python, Flask, MongoDB, Selenium, and Gemini, we developed a web app that helps automate personalized outreach.

Challenges we ran into

Scraping/crawling business websites ethically and efficiently. Blocked many times.

Accomplishments that we're proud of

Our well-rounded and overall complete app! We managed to create a login system, an efficient RAG, dynamic time-based company prioritization for outreach, a really nice-looking website, a scraping system, and more all in 36 hours.

What we learned

How to use MongoDB
How to make a hackathon idea that's also feasible as a real-world startup!
How to make an RAG architecture

What's next for InTouch

We want to expand InTouch with support for messaging businesses across more platforms on top of LinkedIn, such as Instagram or Twitter. We also want to expand the dashboard and web design.