Would you trust AI to read a contract before you signed? — Docusign’s new Iris engine wants to eliminate “signer hesitation”

by

Let’s be real – how often do you actually read those terms of service before just clicking “Agree”? Or a legal document, for that matter? Do you truly read every single word before you sign it?

It’s crucial to carefully read and understand any document before signing it, but Docusign seems to suggest this isn’t always a priority.

The “Q-Tip” Problem

Docusign has introduced new AI tools to help people quickly understand agreements, providing instant answers instead of requiring them to read through lengthy legal documents.

Although the creators don’t present this tool as a way to skip reading, most people will likely use it for that purpose. It’s similar to Q-Tips, which are technically not meant for cleaning ears, but everyone uses them that way anyway.

I’m really excited about this new feature in Docusign! It uses an AI called Iris to basically explain agreements in simple, everyday language. And you can even *ask* it questions about the document! Docusign says it means we can sign things much quicker and feel really sure about what we’re agreeing to – which is a huge relief!

Confidence vs. Correctness

Docusign heavily emphasizes “confidence” in its advertising. A recent survey by the company revealed that 75% of people would be more comfortable signing documents with the help of artificial intelligence.

While many AI tools are available, Docusign argues its new feature, designed to help people understand and complete agreements, is more effective than simply copying and pasting a contract into a general-purpose AI platform.

It’s crucial to remember that AI tools can sometimes make things up – they might present false information as if it were true. While these errors can be minor, it’s important to verify AI-generated content because overlooking inaccuracies could lead to serious problems.

Police in the West Midlands, England, recently admitted a mistake in an intelligence report led them to wrongly prevent Maccabi Tel Aviv fans from going to a game against Aston Villa. The report incorrectly stated that a match had taken place and involved an incident.

As an analyst, I see significant risks here. If a police report contains inaccurate information generated by AI, it could easily become a civil rights concern. And in the business world, an AI ‘hallucination’ – essentially a fabricated detail – within a contract could lead to overlooking a harmful clause or an unexpected fee. It’s a real vulnerability.

More data, less problems?

Docusign explains that Iris stands out from typical chatbots because it focuses specifically on agreements, not general conversation. It leverages details from each contract and Docusign’s agreement management technology to provide more precise information.

Since last year, Iris AI has analyzed more than 100 million customer agreements, giving Docusign a significant edge through the insights gained from this data.

How well a large language model performs is heavily influenced by the data it’s trained on. Google, with its vast collection of images, has developed image generation tools that currently lead the market.

According to Mangesh Bhandarkar, Docusign’s head of product management, the latest AI features are designed to help people feel more confident when signing documents.

I just wonder if hesitation is a negative when it comes to signing agreements.

DocuSign believes many people find dealing with contracts tedious and are now open to using AI to quickly understand them. Are you comfortable skipping the detailed reading and letting their Iris AI engine do the work?

Read More

2026-01-16 01:10