ApplyX

Inspiration

The inspiration for ApplyX came from a simple observation: applying for opportunities is not just about filling forms.

Candidates repeatedly enter the same information, manually read eligibility criteria, compare requirements with their profile, deal with missing or uncertain information, and decide whether an opportunity is even worth pursuing.

We wanted to build an agent that could handle the thinking before the form filling.

This led to ApplyX - an Al application agent that understands opportunities, verifies eligibility, identifies uncertainty, asks for clarification when needed, and makes an explicit Apply, Review, or Skip decision. What makes the idea especially important to us is that automation should not mean giving an Al unrestricted control. ApplyX can reason and prepare an application, but human approval remains the boundary before real-world submission.

Our goal is simple: reduce repetitive application work while making the decision process smarter, transparent, and human-controlled.

WE DIDN'T WANT TO BUILD ANOTHER FORM FILLER. WE WANTED TO BUILD AN AGENT THAT UNDERSTANDS WHETHER APPLYING IS ACTUALLY WORTH DOING.

What it does

ApplyX understands job, internship, and scholarship opportunities, analyzes requirements against the candidate's profile, verifies eligibility, identifies uncertainty, and decides whether to Apply, Review, or Skip. It then prepares the application and keeps the human in control by requiring approval before real-world submission.

How we built it

We built ApplyX using Strands Agents for Al reasoning, Python for orchestration and decision logic, SQLite for application state and duplicate protection, and Playwright for controlled browser automation. A deterministic state machine manages the workflow, while human approval remains the final boundary before submission.

Challenges we ran into

The biggest challenges were handling uncertain eligibility, maintaining reliable application state, preventing duplicate applications, and safely connecting Al reasoning with browser automation. We also had to ensure that Al decisions remained structured, traceable, and never bypassed human approval before submission.

Accomplishments that we're proud of

We built a working agentic application workflow that goes beyond form filling. ApplyX can analyze eligibility, handle uncertainty through clarification and re-evaluation, prevent duplicates, prepare versioned applications, and enforce human approval before real-world submission.

What we learned

We learnt that building a reliable Al agent requires more than intelligent reasoning. ApplyX taught us the importance of structured decisions, deterministic workflow control, uncertainty handling, state management, traceability, and keeping humans in control of critical real-world actions.

What's next for ApplyX

Next, we plan to make ApplyX a scalable opportunity-to-application platform by adding intelligent opportunity discovery, supporting more job and scholarship portals, improving document understanding, and expanding real-world application automation. Our focus is to scale automation without compromising reliability, transparency, or human control