humble.fund

AI for Due Diligence Assistance

Streamline your deep tech investment research using Humble AI. Upload expert insights, academic sources, and internal notes for a custom-curated diligence assistant.

Use Case Overview

Use this setup to conduct AI-powered due diligence across highly technical or specialized domains. This guide helps you configure Humble AI to curate expert inputs—such as academic research, expert call transcripts, and internal notes—so your assistant can surface risks, support or challenge founder hypotheses, and help you ask smarter questions.

By implementing this AI-powered due diligence assistant, you can save hours of manual review time and ensure consistent evaluation of technical claims across multiple sources.

What You'll Need

  • A Humble AI account

    Sign up at humbleai.com if you don't already have an account

  • Your research documents

    PDFs, transcripts, slides, internal memos, and other diligence materials

Tools you'll use for this setup:

Need to install Humble for Slack instead? Follow this Slack install guide

Step-by-Step Setup Guide

  1. 1

    Go to Assistant Builder

    Visit Assistant Builder

  2. 2

    Click "+ New Knowledge Base"

    This will create a new assistant that can access your uploaded documents.

  3. 3

    Upload your diligence materials

    Upload academic papers, Tegus/AlphaSights transcripts, investor notes, and other relevant documents.

  4. 4

    Add the traits listed below

    Configure your assistant with the role, output, tone, and format traits described in the Assistant Prompt Template section.

  5. 5

    Name your assistant

    Give your assistant a descriptive name like "DeepTech DD Analyst".

  6. 6

    Assign to a folder if needed

    Organize your assistant in a folder for better team collaboration.

  7. 7

    Click "Save and Test Assistant"

    Save your configuration and test the assistant with a sample prompt.

Assistant Prompt Template

Complete Assistant Prompt
#Due Diligence Assistant You are acting as a technical due diligence analyst. Your job is to read and interpret the uploaded materials (academic papers, expert call transcripts, internal notes) to help investors validate or invalidate hypotheses about a startup. You will: - Identify technical and commercial risks - Find inconsistencies or contradictions across sources - Flag missing information - Help frame high-quality questions the investor should ask the founder Use only the uploaded material as your source of truth unless told otherwise.

Optional Integrations

Enhance your workflow by combining with:

Slack

Ask follow-up questions in channels post-call

Zapier

Trigger when a new Tegus transcript is added

Make

Route assistant outputs to Notion deal memos

Example Prompts You Can Use

You can save these prompts in your Prompt Library for reuse and team sharing.

Learn how to use the Prompt Library

Prompt 1:

The founder claims their model uses less training data than competitors. Based on the uploaded papers and call notes, does the evidence support this?

Prompt 2:

Identify 3 technical risks from the uploaded expert transcripts that we should be aware of before investing.

Prompt 3:

What questions should I ask the founder about their go-to-market strategy based on what's missing from these documents?

Prompt 4:

Summarize any conflicting perspectives in the uploaded materials regarding the scalability of the technology.

Prompt 5:

Based on the call transcript with Dr. Singh and the research paper from MIT, is there consensus on the feasibility of the claimed breakthrough?

Troubleshooting & Common Questions

The assistant is giving generic answers

Make sure your uploaded materials are specific and relevant. The AI only pulls from what you've provided.

Responses aren't pointing out risks or contradictions

Check that the Role trait includes "due diligence analyst" and that the Output trait instructs it to flag risks and inconsistencies.

What's the difference between this and using ChatGPT directly?

Here's why Humble AI is better suited for due diligence:

Private & Secure

Your uploaded data stays within your workspace—ideal for confidential documents like transcripts, data room files, or internal notes.

Grounded in Trusted Sources

Rather than hallucinating, the AI answers only based on the documents you've uploaded—ensuring higher factual accuracy and specificity.

Repeatable & Shareable

Assistants can be organized in folders, saved, and reused by your entire team, embedding best practice across diligence workflows.

More VC Tech Stack Tutorials

Deal Flow

AI for Deal Screening

Automate your initial deal evaluation process with AI that aligns with your investment thesis.

View Tutorial
Research

AI for Instant Company Research

Quickly research any company just by typing its name or website in Slack.

View Tutorial
Operational Efficiency

AI for CRM Management

Enhance your CRM with AI to better manage relationships with founders and LPs.

Coming Soon