RFP Response Facilitator

RFPs and their long list of questions can be tricky and cumbersome to answer, having to invoke many stakeholders. Here is an experiment to tackle the issue

Description

Inspired by our past experiences at FCTG, where we encountered the time-consuming and repetitive nature of handling RFPs, we designed an experiment to streamline this process using AI.

The experiment aimed to address the inefficiencies of gathering and answering RFP questions that require input from multiple departments.

Here is the approach for our simple yet powerful experiment:

• Document collection: compile a list of documents about the company, product specifications, technical information and security protocols. It should include as much information as possible to answer frequently asked RFP questions from past proposals. This should be an ongoing process: adding and updating documents as they are created, to have an always up-to-date knowledge base. Note that this process could also be facilitated!
• Embed documents: embed the documents in a vector database(the kind of database LLMs can read from) to allow the AI to search into your knowledge.
• AI Integration: the LLM is set to accurate respond to the listed questions based soly on the documents provided. Otherwise, it states that it doesn’t have the answer.
• Apply it to a potential RFP: list some frequently asked RFP questions from past proposals in a Google Sheets. It only has 2 columns: question and answer. We then automated the trigger to run the AI Automation whenever needed.

The result? In a matter of seconds, you could have a first draft of the answer, with the material needed to tailor the answer to the exact case. In the end, you know how to close a bid!

By automating the response process, we aimed to save time and reduce redundancy, allowing the team to focus on more strategic aspects of the RFPs.

Do you want a sneak peak on how it’s built?

Yes, that’s it. With a bit of work put in each node!

Potential

It is very simple and could be iterated on multiple times, based on a company specificities. One thing that would be key to work on is the source of the answer to 1. give trust to the employees that there was no hallucination, and 2. to help dig deeper if needed.

Another one is that it could not only for RFPs, but also for all sorts of documents that requires collecting documents. Another example could be to streamline the process of creating proposals!

Lastly, the process of answering RFP is iterative and requires a lot of tweeking. Humans are great for that, and AI can give a little bit of help too. In our case, AI could help tweaking the responses based on human input. For example, using comments: think of it as human-AI collaboration.

👉 Does it ring a bell? Does it take ages to answer all these questions? Maybe we can help your team!