Jennifer AI is a new associate starting work at your law firm. Orientation has completed and it is time for Ms. AI’s first assignment. Let’s see how she does.

Memorandum

From: David Johnson, Partner

To: Jennifer AI, Associate

Date: Tuesday, November 2, 2023 at 4:00 pm.

Regarding: Upcoming Depositions in the Bottled Water Matter

Hello Jennifer and welcome to the firm. We need your help right away.

Our client is a citizens group concerned about how Florida administrators are allowing private companies to extract water from the aquifer and sell it as bottled water. There are concerns that the companies are depleting the aquifer, causing sinkholes to swallow homes and more.

We sent out this request for production: “All documents evidencing, reflecting, or pertaining to the extraction of water in Florida for bottling by commercial enterprises.” Depositions are not yet scheduled but we need to know who to depose from which companies and which government officials are involved in the decision making process. Ideally, we would like to know who is in whose pocket in this business.

We need your response ASAP, hopefully by the end of the week. Depositions start next Monday, November 6th.

Memorandum

From: Jennifer AI

To: David Johnson, Partner

Date: Tuesday, November 2, 2023 at 4:30 pm.

Regarding: Upcoming Depositions in the Bottled Water Matter

The issue of water extraction for bottling in Florida involves several key entities, including commercial water bottling companies, landowners, citizens groups, and government agencies.

Commercial water bottling companies such as Zephyrhills Natural Spring Water and Nestle Waters North America have been extracting water from various springs in Florida for bottling purposes (Bush050189, Bush249878, Bush470991, Bush550978, Bush071366, Bush061164, Bush618047). Robert Thomas, the owner of Two Rivers Ranch and Crystal Springs Preserve, has been involved in water extraction activities, applying for permission to increase water extraction limits for Zephyrhills Natural Spring Water (Bush050189).

Citizens groups such as Save Our Springs (SOS) have been actively opposing increased water extraction and bottling. SOS, led by Terri Wolfe, has been involved in disputes over water extraction and has been vocal about the potential depletion of the aquifer (Bush050189, Bush292191, Bush550978, Bush061164, Bush618047).

Government agencies such as the Southwest Florida Water Management District (Swiftmud) and the Florida Department of Environmental Protection (DEP) have been involved in regulating water usage and granting or denying applications for increased water extraction (Bush050189, Bush071366, Bush470991, Bush292191).

In terms of potential depositions, the following individuals and organizations could be considered:

• Robert Thomas, owner of Two Rivers Ranch and Crystal Springs Preserve (Bush050189)
• Representatives from Zephyrhills Natural Spring Water and Nestle Waters North America (Bush050189, Bush249878, Bush470991, Bush550978, Bush071366, Bush061164, Bush618047)
• Terri Wolfe, leader of Save Our Springs (Bush050189, Bush292191, Bush550978, Bush061164, Bush618047)
• Representatives from the Southwest Florida Water Management District (Swiftmud) and the Florida Department of Environmental Protection (DEP) (Bush050189, Bush071366, Bush470991, Bush292191)

As for the question of who is in whose pocket in this business, the documents do not provide direct evidence of any inappropriate relationships or financial transactions between the involved parties. However, the ongoing disputes and controversies over water extraction for bottling suggest a complex interplay of interests and influences that may warrant further investigation.

Memorandum

From: David Johnson, Partner

To: Jennifer AI

Date: Tuesday, November 2, 2023 at 4:45 pm.

Regarding: Upcoming Depositions in the Bottled Water Matter

Thanks for your quick response.

Who should we depose and why? What topics should we cover and what documents should I include for each?

Thanks again for your help.

Memorandum

From: Jennifer AI

To: David Johnson, Partner

Date: Tuesday, November 2, 2023 at 5:00 pm.

Regarding: Upcoming Depositions in the Bottled Water Matter

In preparation for the upcoming depositions in our case concerning water extraction for bottling in Florida, I have identified several key individuals who could provide valuable insights and information. Below are the individuals, along with the reasons for their selection and potential questions to ask during the depositions.

1. Robert Thomas, Owner of Two Rivers Ranch and Crystal Springs Preserve

Robert Thomas has been directly involved in water extraction activities, applying for permission to increase water extraction limits for Zephyrhills Natural Spring Water (Bush050189). His property, the Two Rivers Ranch and Crystal Springs Preserve, is a source of water for bottling by Zephyrhills.

Questions to ask:

• What agreements do you have with Zephyrhills Natural Spring Water regarding water extraction from your property?
• What measures are in place to monitor and manage the impact of water extraction on the aquifer and surrounding environment?
• Can you provide details of your application to Swiftmud to increase the water extraction limit?

2. Representatives from Zephyrhills Natural Spring Water and Nestle Waters North America

These companies have been extracting water from various springs in Florida for bottling purposes (Bush050189, Bush249878, Bush470991, Bush550978, Bush071366, Bush061164, Bush618047). They would have direct knowledge of their companies’ water extraction activities and policies.

Questions to ask:

• What are your company’s policies and practices regarding water extraction in Florida?
• How do you monitor and manage the impact of your water extraction activities on the aquifer and surrounding environment?
• Can you provide details of any applications your company has made to increase water extraction limits?

3. Terri Wolfe, Leader of Save Our Springs (SOS)

Terri Wolfe has been leading SOS in opposing increased water extraction and bottling. She has been vocal about the potential depletion of the aquifer (Bush050189, Bush292191, Bush550978, Bush061164, Bush618047). Her insights could provide valuable context on the concerns and arguments of those opposing commercial water extraction.

Questions to ask:

• Can you provide details of SOS’s opposition to increased water extraction and bottling in Florida?
• What specific concerns does SOS have about the potential depletion of the aquifer?
• Can you provide any evidence or examples to support SOS’s concerns?

4. Representatives from the Southwest Florida Water Management District (Swiftmud) and the Florida Department of Environmental Protection (DEP)

These government agencies have been involved in regulating water usage and granting or denying applications for increased water extraction (Bush050189, Bush071366, Bush470991, Bush292191). They would have direct knowledge of the regulatory framework and decision-making process around water extraction in Florida.

Questions to ask:

• Can you provide details of the regulatory framework for water extraction in Florida?
• What factors do you consider when reviewing applications for increased water extraction limits?
• Can you provide any examples of applications that have been denied, and the reasons for denial?

I believe these individuals will provide valuable insights and information for our case. Please let me know if you have any questions or need further information.

Is This a Real Life Story?

It could be. You have correctly deduced that our associate Jen AI is a large language model (LLM), which many call generative AI.

Using manual techniques, this project might take several days to complete with a (really smart) human associate doing the work. At a billing rate of $800 an hour, the associate’s analysis might cost $15,000 or more.

With DiscoveryPartner–powered by Sherlock AI–these reports can be completed in minutes. Our hypothetical associate actually held off sending her responses for several minutes to give the appearance of being hard at work. The whole project cost about $1.00 in AI software costs.

GenAI-integrated systems like DiscoveryPartner can do the following–in minutes rather than hours or days:

1. Help create a search across two million Jeb Bush emails (using our evaluation site for this exercise);
2. Run the search, identifying the 100 most likely relevant documents (this is configurable);
3. Use a GenAI model like Claude Instant to summarize the documents based on your topic description and interest (prompt);
4. Rank the documents based on likely relevance (0 to 10);
5. Sort the summaries in relevance order, top to bottom;
6. Send the summaries to a GenAI model like GPT 4 to use for its analysis; and,
7. Display the analysis complete with live links to the summaries and original emails.

Welcome to a new era of discovery. It’s a world where we can seamlessly transition from search hits to discovery answers, thanks to the extraordinary capabilities of machine learning and, particularly, generative AI.

By seamlessly integrating generative AI and other machine learning algorithms into a full-featured investigation and discovery platform like DiscoveryPartner, legal professionals can harness the immense capabilities of large language models like GPT and Claude to streamline discovery processes and quickly dive deeper into the key documents that are most relevant for the case.

A real life story? It is now. Reach out to us at Merlin.Tech to see our new generative AI discovery platform in action.

Behind the Curtain

Our peerless associate, Jennifer AI, distinguished herself with those promptly-delivered memos and appears to be a shoe in for the partner track. If you want a closer look at how she managed to create these memos, and to peer behind the curtain as the Wizard of Oz might say, we offer this closer look at DiscoveryPartner and the steps Jen had to take to prepare her memos for the Bottled Water Matter.

The Initial Question

To start, we inserted the text of our document request: “Bottled Water — All documents concerning the extraction of water in Florida for bottling by commercial enterprises” into DiscoveryPartner’s (DP’s) search window.

We are using a unique, AI-powered process, which we call Freeform Search, to find highly relevant documents. It will ignore search syntax and even punctuation, instead focusing on the most important terms for its retrieval efforts.

In practice, we would enhance the query by removing stop words like this phrase: “All documents evidencing, reflecting, or pertaining to.” We might also add other keywords or phrases, as many as we can identify, to improve the search.

With an algorithmic search, you don’t have to worry about syntax or even adding the wrong keywords. The algorithmic search engine will retrieve the most likely relevant documents based on the entirety of your search.

Initial Results

Here are our initial results:

This is our snippet view, different from the traditional grid layout. It allows us to extract relevant information from the documents in a Google-like view to give you an idea of your search results. You can scroll through the page or jump into individual documents to review.

The snippet results appear to be on topic so our next step is to copy the top 100 results into our Analyze folder so Sherlock can summarize and analyze the files to answer our questions.

Alternative Approach: Asking Sherlock to Suggest Keywords

Although we didn’t use this technique for this paper, you can also ask the GenAI algorithm to suggest keywords based upon its general knowledge and the documents it has reviewed. We did that and got these results:

“water extraction”, “Florida”, “aquifer”, “bottling”, “commercial enterprises”, “Zephyrhills Spring Water Company”, “Perrier Group of America”, “Nestle Corporation”, “water permit”, “Suwannee River Water Management District”, “Florida Department of Environmental Protection”, “Madison Blue Springs”, “water policy”, “drought”, “aquifer depletion”, “sinkholes”, “water conservation”, “water resources”, “water management”, “water bottling”, “water withdrawal”, “water use permit”, “consumptive use permit”, “water rationing”, “water shortage”, “water bottling plant”, “water bottling facility”, “groundwater extraction”, “water supply”, “water rights”, “water regulations”, “water legislation”, “water policy”, “water governance”, “water law”, “water conflict”, “water dispute”, “water controversy”, “water crisis”, “water issue”, “water problem”, “water challenge”, “water concern”, “water risk”, “water threat”, “water impact”, “water effect”, “water consequence”, “water result”, “water outcome”, “water damage”, “water harm”, “water detriment”, “water destruction”, “water devastation”, “water ruin”, “water loss”, “water depletion”, “water reduction”, “water decrease”, “water decline”, “water diminution”, “water shrinkage”, “water contraction”, “water lessening”, “water dwindling”, “water draining”, “water sapping”, “water emptying”, “water exhaustion”, “water exploitation”, “water overuse”, “water overexploitation”, “water overextraction”, “water overpumping”, “water oversiphoning”, “water overdraining”, “water overdepletion”, “water overconsumption”, “water overharvesting”, “water overwithdrawal”.

While we wouldn’t suggest using all of these terms, they provide additional keywords and keyword phrases we could enter into the search window for DiscoveryPartner to consider.

Summarizing the Documents

The next step is to submit our top documents to Sherlock for summarization and analysis. We start by describing what we are looking for to the GenAI engine. Our description is used to guide Sherlock’s summarization efforts and to provide an initial analysis of our topic question. In this case, let’s input the key paragraphs from the partner’s request memo to our GenAI engine.

In this case, we are asking Sherlock to read, analyze, and summarize 100 likely relevant documents which we found from our search (shown in the snippets page above). The chosen GenAI engine (Claude Instant or GPT 3.5) will quickly read all of the documents and present summaries for each, like this top-ranked summary below.

The summarization process for all 100 documents takes about 30 seconds in total.

Click on View Text to see a text version of the underlying document.

Or click on View Document to see the underlying native file and tag it as may be appropriate.

Sherlock’s Analysis

After finding, summarizing, and ranking the top 100 documents, our GenAI engine provides its answers to your topical questions.

We can scroll down to see more of Sherlock’s answer:

Who Should I depose?

Our next question was:

I need to write a memo to my senior legal partner answering this question. Please write the memo:

Tell me who should we depose and why that person? List each person in bold by name and title in a separate heading. For each, tell me what questions should I ask and why with links to any and all relevant documents. List each in a numbered list.

Here is Sherlock’s answer:

And here is more of Sherlock’s answer.

As you can see, Generative AI can be a powerful tool for helping Jenifer find and report on answers for her first litigation assignment.

Assisted by GAI and LLM Technologies per EDRM GAI and LLM Policy.