In drafting problems, we have found that once your workflow is clear, your templates are set up, the product is usually highly accurate, to your standards.
Here, if it is client instructions, you can directly bring it into the workflow.
A simple example of this is the Termsheet App, where the workflow consists of a series of questions that the client or your professionals must answer and the draft is ready.
Some of you want your own template, you have your own client touch points (questions), and you can build your own to take it into account.
This doesn’t cost much, which means that adoption is going to be much faster than we expected.
Many drafting exercises on the other hand start with review.
Reviewing documents. Reviewing instructions.
This is what makes review an attractive entry point into any workflow.
Let our AI agent sit with you in a client meeting. Hand over the notes to another agent who reviews the notes. This agent then identifies what needs to be escalated to other agents or humans as the task requires.
The agent receives client documents. It passes it on to another agent to review it, and a third to draft the outcomes.
In this world of AI agents, we will prize not just accuracy and efficiency but also benefit from peace of mind due to lack of inconsistency, and our own workflows becoming more systematic.
For the review problem, we have launched an experimental interface for those of you who want to play with a legal AI review tool.
What we are looking for is wide ranging feedback, in exchange for free access and full data security.
What you can do with it.
Reviews upto 5 of your documents.
It answers questions based only on what’s in the documents.
All documents are in temporary storage and deleted after a session.
You can complain if you don’t like the answer.
Make sure to use the feedback tool so we know what is going on.
Here is where you can access it - https://docreview.scripters.app/
If you don’t like the tool, it is reason to build one that you like, because anything is possible.
Review is interesting.
It can be the entry point to handing over to your next agent, the drafting agent or an agent that communicates with your client, or even a report or list that you need to make for your work.
If your job involves making sense of a box of documents, then you might need an AI document analysis agent.
If it then involves drafting something, or doing something, you can consider building an agent for that workflow.
I believe it will increasingly be important to own your own agents, because our agents will define who we are, much like our work, our brand, our colleagues and our partners.