Automated Form-Filler Agent for Grants, Tenders, or Events

A smart browser agent that automatically scrapes upcoming events or tenders, filters for eligibility, fills out multi-page forms using a local knowledge base, and presents the completed form to a human for a final one-click review.

  • The Problem: Manually searching for industry events, evaluating complex eligibility criteria, and repeatedly filling out long, tedious application forms was a massive time sink.
  • The Solution: A smart browser agent that automatically scrapes upcoming events or tenders, filters for eligibility, fills out multi-page forms using a local knowledge base, and presents the completed form to a human for a final one-click review.
  • Similar use-cases: Government grant applications, procurement tenders and RFPs, business directory listings, vendor onboarding, and award submissions.

Why fill forms?

Filling out forms has to be one of the most lucrative single activities a business can undertake. Whether applying for industry events, government grants, or corporate tenders, a single application form can mean a win worth thousands to millions, or a high visibility marketing opportunity that boosts your business, brand, and sales.

It's also one of the most tedious and fiddly jobs there is. Often these application forms carry a high chance of failure, and so the key is to fill out a lot of them and accept that most will go nowhere. Those rejections can feel personal and make it a soul crushing job when the application process took so much time and effort. The solution is to scale it up, make it fast and easy to apply, and fill out as many applications as possible so that any one application stops carrying so much importance as the success rate goes up.

The Brief

The brief for this job was to automatically discover the best opportunities from a particular website, match them to a product for suitability, and then automatically fill out the most suitable set for applications. In this case the payoff was acceptance into a high visibility marketing opportunity for the product. In other cases the opportunity might be a tender and the "product" could be the service the company provides. The shape of the problem is the same: filling out lots of forms.

The Agent Design

The implementation was fairly straightforward, using a lightweight browser agent powered by Python and Playwright and run on a nightly basis. The workflow breaks down into three distinct roles:

  • Researcher: Runs every night to scrape opportunity listings, extract requirements and the form URLs, and evaluate eligibility criteria so unsuitable listings are automatically filtered out.
  • Knowledge Base: A structured document containing detailed background information, product specifications, team bios, and answers from previous applications.
  • Form Filler: Opens each form URL in an automated browser window, reads every field, finds the answer in the Knowledge Base, and then inputs the response across multi-page forms.

Active Learning & Human Input

There are three touch points where the human admin is looped back in. First, when the list of opportunities is created, the human gets to decide which ones are actually pursued. Secondly, when the agent can't answer a question based on the knowledge base, it asks the human for input. Critically, any new information the human provides is then entered into the knowledge base, helping the agent to actively learn about the product for future application forms. Finally, the agent never clicks submit automatically. It completes 99% of the heavy lifting and then pauses on the final page so the human expert can briefly review before clicking submit themselves.

This process may sound longwinded, but you must remember an agent doesn't care how difficult or annoying it is, and will blaze through the forms in the blink of an eye. The human part of the job is reduced to a couple of minutes of review. This changes a task that can potentially take hours to complete, into a walk in the park that only lasts a couple of easy minutes.

The human review step is critical. Agents can make mistakes and hallucinate assumptions about your product or company that are wrong. The feedback loop updating the knowledge base is incredibly valuable as it helps the agent learn and makes your future forms more and more accurate, requiring less and less intervention.

With everything I build I try to put the human in the loop at the most high leverage points and involve them in the most critical decisions. The human is the expert. Nobody knows your company, service, or product like you and your team. The agent just helps you advocate for it efficiently and at scale.

The Payoff

This automation dramatically shifted the economics of applying. It took a fiddly, annoying job that was error-prone and took ages, and turned it into something quick, easy and high leverage. It lifted the number of forms that could practically be applied for and that worked amazingly well. The product was accepted into real marketing opportunities and was seen by thousands of people, leading to real sales. The investment into automating the application process was an absolute no-brainer (pun intended).

If you're interested in similar time saving opportunities, please feel free to get in touch.

Ready to automate your business?

Let's discuss how we can eliminate your manual workflows and save hours of admin time every single month.