Intellectual Property Litigation, Protection, Prosecution and AI Legal Matters

Listen to this Blog Post with Additional Commentary from the Author Henry M. Sneath - 49:01 min

The Daily AI Routine: A Practicum for Using AI Tools in Context - in Business and IP Litigation

This blog post and audio file is another in the series "Defending the Algorithm™" written, edited and narrated by Pittsburgh, Pennsylvania Business, IP and AI Trial Lawyer Henry M. Sneath, Esq. and was authored with research assistance by Claude® from Anthropic Sonnet Edition 4.5 Pro and research confirmation and assistance from Google Gemini 2.5 Flash and Westlaw Precision AI. This series focuses on AI, the legal practice, and the intersections of AI and substantive law. AI platforms can make mistakes but the Author has taken care to check them all for accuracy.

Introduction

In this Defending the Algorithm™ blog and audio series, we've examined the important Bartz v. Anthropic copyright settlement, explored the expanding legal battleground over AI training data beyond copyright claims, and calculated trade secret litigation probabilities. Throughout these analyses, I've disclosed that I use AI tools extensively in my legal practice—from researching and writing pleadings, briefs and strategy memos, to writing these posts and creating marketing materials. Obviously, we all use AI to manage our daily litigation workload, to sort and filter emails, to store documents to files, and to manage and generate large document reviews, analyses, coding and the production of documents. Consider this a practicum—a practical guide demonstrating that thoughtful AI implementation—involving proper information management, Bayesian thinking and rigorous user training—can enhance rather than replace or compromise the quality of legal services.

This raises an important question that I frequently hear from lawyers at CLEs, from business clients, and in conversations with other practitioners: "Exactly how are you using AI in your practice day-to-day? What are the specific tools and how far can you really go with AI in the law?" The answer matters because lawyers and clients seek efficiency, cost savings and great litigation results, but we all face unique professional responsibilities relating to competence, confidentiality, and candor that shape how we can ethically deploy these powerful tools.

This post shifts the series from theoretical analysis to practical application. I'll walk you through my actual daily workflow as a trial lawyer actively using AI across nearly every aspect of practice, from predictive analytics that inform strategic decisions to context-aware research tools that transform how I analyze cases.

The critical insight I've learned through trial and error: successful AI implementation depends on how well you've prepared your practice's information ecosystem to support it. Many firms invest heavily in AI platforms while overlooking the foundational work of organizing their data, standardizing their processes, and training their people. AI isn't magic—it's a sophisticated tool that requires thoughtful integration into existing workflows and information structures. The firms seeing real results aren't necessarily using different AI platforms; they're using them differently, with better preparation and clearer understanding of how information flows through their practice.

Bayesian Analysis: The P(lawyers will achieve AI competence to maintain their duty of technological competence under ABA Rule 1.1 et. seq. | The need for a dedicated approach to building smart and communicative data architecture) = the amount of hard work which law firms are willing to engage in to make it happen correctly.

The Foundation Problem: Why AI Implementation Often Disappoints

Before diving into my daily workflow routine, we need to address a frustrating reality: many lawyers trying AI tools find the results underwhelming. After some years of intensive AI adoption in my practice, I've learned that disappointing outcomes rarely stem from inadequate AI technology. Instead, the problem typically lies in how firms structure their information and workflows before introducing AI into the mix – and the sophistication with which lawyers use AI prompts. Consider how experienced litigators naturally work. When evaluating a case, we don't examine documents or evidence in isolation. We instinctively connect dots across multiple dimensions—linking factual allegations to procedural requirements, assessing how specific judges have ruled on similar issues, considering the opposing party's litigation history, and evaluating how various legal theories align with our client's broader business interests. This holistic Bayesian analysis draws on years of accumulated knowledge about how different case elements interconnect. We take prior probabilities of an outcome, blend in the new and gathered likelihood evidence and factors -- and project a posterior probability – an outcome prediction - to our clients and litigation foes.

AI systems need access to similar interconnected information to provide useful analysis. Yet many law firms offer AI tools in a fragmented landscape: separate systems for document storage, disconnected email archives, case management software that tracks deadlines but not strategic context, time-keeping databases with no link to substantive legal work or a client CRM, and research platforms operating in complete isolation from active matters. This fragmentation in data storage, architecture and IT coordination represents the fundamental challenge. When AI can only access isolated pieces of information from disconnected systems, it cannot provide the integrated analysis that makes it valuable. It's like hiring a brilliant associate but only allowing them to read one category of documents while keeping everything else locked away.

Experts have come to think of this as the information architecture challenge. Before worrying about which AI platform to purchase, firms need to examine how information flows—or fails to flow—through their practice. Where do gaps exist? Which systems don't communicate with each other? What critical context gets lost when lawyers manually transfer information between platforms? The most successful AI implementations share a common characteristic: they occurred in firms that first invested time organizing their information infrastructure. They established consistent naming conventions, created centralized matter repositories, developed standardized intake procedures, and built connections between previously siloed systems. Only after laying this groundwork did they introduce AI tools—which then performed remarkably well because they finally had access to the connected information they needed.

Chapter 1: Workflow – Litigation Research, Analytics, and Case Assessment

My typical day begins rather typically I suspect, with opening several software systems in our IT network and on the internet. That’s not unusual for most lawyers. Outlook, Juris, NetDocs, Google Gemini, Adobe Premier Pro, Ring Central, Teams and maybe Westlaw Precision depending on the plans for the day. Much depends as well on whether I am working on actual files or doing work for marketing, business development, website content or social media. As a senior lawyer I had gotten away from much of the actual drafting of documents and spent my time generally supervising, editing and commenting upon the research and writing work of others. While I am hands on in every aspect of a case from matter opening to trial (or being an appellee!) I delegated legal research and writing to more junior lawyers with lower billing rates as part of cost management for the matters, and to the benefit of clients. I also generally delegated 1^st level large volume document review, coding and analysis to junior lawyers, paralegals and contract document review lawyers.

Cost of Using AI – Accretive? Efficient?

While still being mindful of litigation spend for clients, some of that dynamic has changed due to a number of AI tools in our kit. I find myself doing more and more front-end legal research to start our team’s analysis for the client, and more document review albeit generally at the 2L or 3L review level – depending on the volume and complexity of coding, tagging and privilege issues. I am enjoying doing front-end research to set the parameters for more detailed research on complex commercial and IP matters. I will explain this Westlaw front-end protocol in a moment. In order to make a more accurate prediction to a client at the front-end, I can conduct legal research very quickly with AI myself, to get the analysis moving forward very quickly. I prime the research pump at very little cost to the client and we can therefore create and more quickly agree on a strategic case assessment. My role in document review in Relativity is quick and efficient and takes far less time (less cost) than lugging boxes to a conference room and shuffling paper like we used to do. Document review can be done more efficiently with Active Learning AI and reports can be generated showing me the pace of review by others, the number of times that documents have been marked “hot docs” or “privileged” and the tagging of documents by issue or witness designation. All of that work makes prep for depos and trial much more efficient and less costly. While this may not be a scientific cost analysis, there are certainly areas where AI has fundamentally transformed the efficiency of my team through both AI research and writing tools and predictive analytics.

Chapter 2: Building the Tech Stack - Specific AI Tools For The Daily IT and AI Routine

Using your static IT tools and dynamic AI platforms in a coordinated and purposeful way makes all the difference. I do NOT put any client or case related data on any public platform but I can use these platforms to hone language or presentations in a non-named, case generic way. Here is a standard suite of AI and IT platforms that can be effectively used at the same time on multiple screens, to create a more seamless work and information flow – and better AI output:

• Google Gemini: Google itself at the premium level has built-in AI which will summarize your search results and provide narrative which helps to focus your search. The Gemini module is more advanced and can generate images, audio and all things AI. I tend to use it as a primary resource for getting started with some ideas on either a legal research project or quick check of the state of the law in a particular jurisdiction. It learns from our interactions but will not display my personal information unless directed to do so. I also use it as a secondary resource to fact check other AI generated content and non-client materials. It is also a good interface with Google Drive in the Cloud which I use, for marketing and networking articles, blogs and presentations, as a portal for document storage, document sharing and content creation. In other words, I connect various Google apps to improve efficiency, learning and content generation. Gemini has various internal modules which put Gemini into “Research” mode or “Deep Thinking” mode. I use these for my more complicated prompts and the program will display notes about its “Thinking” as the process moves on. Claude will do that as well so that you’re getting a running account of what the AI is analyzing moment by moment so that you understand its methodology. This is great so that you can explain the methodology if you ever need to Defend the Algorithm™. Of course, I use various security settings to decide what I want Google to do with data.
• Westlaw Precision AI-Assisted Legal Research: Our firm upgraded our Westlaw account to the level of Westlaw Precision with the following modules:
  • Westlaw Precision AI Preferred with CoCounsel Premier
  • Westlaw: People Map
  • Westlaw: Company Investigator
  • Drafting Assistant
  • Practical Law AI Connect Preferred Dynamic

  It has many features which make it a very robust AI platform and which therefore helps to augment, not replace our legal analytics and client service. AI tools like Gemini and Claude will admit that they cannot find certain case opinions and analysis like a human legal researcher because they do not have access to closed databases like Westlaw, Lexis and Bloomberg Law. Therefore, it improves the process to work with both your favorite AI program and Westlaw together to produce superior work product.

• Microsoft Outlook, Teams, Ring Central, and NetDocs: These tools guide me through the day with their basic functions and AI integrations. NetDocs stores case files and client documents, while Teams and Ring Central manage internal and external communications integrated with Outlook. Microsoft Co-Pilot AI assists with programming, meeting management, and generating meeting summaries.
• Claude® by Anthropic Sonnet 4.5 Max: This AI tool has become a workhorse for my practice, particularly in business development efforts. I use it to build topic-specific databases, upload case opinions and topical articles, and generate analyses for ongoing projects. While I do not upload client materials, Claude allows dynamic AI interaction with stored research. Courts now increasingly require certifications confirming that AI-assisted work has undergone human review, which I ensure in all instances.

As you can see, the goal is integrated workflow, excellent communication, and organized data management through AI and IT coordination. These tools, when properly connected and managed, provide efficiency gains and greater accuracy in every aspect of legal practice.

Chapter 3: Document and Data Collection, Review and Production

• Relativity®: While document management systems of various kinds are well known to litigation lawyers, I want to focus on the management aspect of the large document review and production projects and the coordination of document management systems like Relativity® AI with the law firm’s other IT. Managing large document projects for gathering, housing, sorting, batching, reviewing, coding, tagging and producing documents is an art and a plan must be included in the first substantive conversation with a client. This process management by a lawyer is a valuable skill because if done right, it often gives your side a huge advantage if the other side is not so savvy with large document management. If it goes wrong, it can really spoil a case both on the cost side and the results side. It also impresses the Judge when you are defending your discovery compliance if it is challenged by the other side. Judges now mandate and enforce detailed and written ESI protocols as a routine part of at least Federal cases and they do not have much tolerance for sloppy or old-style document management.

Because of our firm size (40 lawyers), we do not house Relativity in-house, but work through a dedicated vendor bit-x-bit (BXB) in Pittsburgh. I’m not sure if I would ever want it in house because there is a large staff, in multiple time zones at BXB to work with you at all times to create and customize searches and productions. They work at the very front end of a trade secret misappropriation case for example, to forensically examine, download and maintain chain of custody of all relevant data which exists on certain drives, like those of a client’s departing employee who is believed to have breached his fiduciary duty to his employer and/or is believed to have downloaded company confidential or trade secret information. BXB works directly with the client and their IT folks to either remotely or in person gather all data into their servers and they create an evidentiary initial set. They then create a working version that can be examined, reviewed and coded, so that the original set remains – original. The working set can be put into Relativity once there is a document management strategy established. BXB personnel will provide affidavit support to the Court if needed on issues like evidence gathering, chain of custody and data examination. With other AI tools, they can run initial and very detailed reports on every type of computer activity in which the departing employee engaged in the time leading up to the departure. They can tell us which websites he accessed, which dates and how many times access was made. They can survey email traffic and more importantly the Activity Logs which are behind the scene in Microsoft Outlook - which logs show all of the internet activity triggered by every communication. They can trace everything and can recover data from deleted spaces or in the cloud. Then we decide how much of that gathered data to load into Relativity® for review and processing.

Discovery Document Production

As the case progresses we receive documents from our opposing party and more documents from the client - based on the document custodians that we and the client have identified as most likely to have relevant and responsive documents under Rule 26. We then work with Relativity AI to create certain strategic searches and parameters. BXB can also create overview searches for me or our Team Lead to monitor document review, monitor the most hit-upon search terms and other management tasks. The goal is to meet a certain probability percentage that we have gathered a sufficient number of relevant or responsive documents to comply with the discovery rules, and to pass muster with opposing counsel and/or the Court. These statistics vary depending on the case and the volume and nature of the documents.

We use AI Active Learning in Relativity® to continually re-focus the AI on the most important and most likely responsive documents so that eventually it is feeding us only the most likely responsive documents. The AI is recognizing patterns and providing us with feedback and reports. When we make a production of documents, we store that production on our own servers in a searchable database. If a case is stayed or ends with any uncertainty as to whether it will start up again, we have lesser expensive document storage options with BXB to either archive or put the data on separate large hard drives. At case final end, we can analyze the scope of our responsibilities for destruction of confidential document data and preserve for a reasonable period of time, the remaining data. Relativity AI can assist with that as well.

This is an expensive process, but if it were done by a human reviewer only, it would be far more expensive and not likely to be finished in even a generous discovery time period. If AI reduces a database from 3,000,000 pages to 1,000,000 pages of the most likely relevant and responsive documents, it makes a huge difference in both review time and cost of review. Using AI in this context, while somewhat expensive, is a financially accretive process for the client and it allows 120 days of discovery time period not to be consumed simply by document review.

Document gathering, review and production needs to be managed from day one in a case and the coordinated use of AI tools at every step, if well managed, makes accurate and efficient production of massive ESI data, a more likely probability.  

Chapter 4: Looking Forward – The AI-Competent Lawyer

The legal profession stands at an inflection point regarding AI adoption. Within five years, lawyers who competently use AI tools will likely be the norm rather than the exception. The question facing today's practitioners isn't whether to adopt AI but how to do so responsibly while fulfilling our professional obligations. My daily workflow hopefully demonstrates that thoughtful AI integration—particularly with attention to overall systemwide information organization and Bayesian analytical frameworks—can significantly enhance legal practice without compromising professional responsibilities. The key insight: treating AI as a powerful tool requiring both professional supervision and proper information infrastructure coordinated by and with your IT professional. The specific tools matter less than the framework for thinking about AI adoption: enhance efficiency while maintaining competence, protect confidentiality, verify AI output, invest in organizational infrastructure, use predictive analytics for strategic advantage, and preserve the professional judgment that distinguishes lawyers from mere document production services.

This requires investments in unglamorous infrastructure work that doesn't generate excitement at partnership meetings. It means consolidating systems, standardizing processes, and organizing information before deploying the latest AI features. It means teaching your team consistent practices rather than allowing everyone to develop their own disconnected workflows. But firms that make this investment—that prioritize information organization before layering AI on top—will discover that AI actually delivers on its promise. Not because they bought superior AI, but because they finally gave AI the organized, connected information it needs to provide useful analysis.

The path forward requires neither blind adoption nor resistant skepticism, but thoughtful integration that serves clients while fulfilling our professional responsibilities. Welcome to the future of law practice—one where AI assists but doesn't replace the human judgment, ethical reasoning, and client relationship management that define our profession. But also, one where that AI assistance only works if we organize our information infrastructure to support intelligent analysis.

The revolution isn't primarily in the AI models themselves—those continue improving regardless of what lawyers do. The revolution is in finally organizing our information and workflows to support intelligent analysis— that comes from human attorneys, artificial intelligence, and coordinated IT systems across all firm tech platforms.

To conclude in Bayesian probability terms: P(infrastructure organization being essential | evidence of competitive advantage from AI) = Very High".

________________________________________

My Thanks to Claude® for its assistance with the Defending the Algorithm series. Houston Harbaugh's intellectual property and AI litigation team continues to monitor developments in training data and AI enterprise litigation across all legal frameworks. For questions about how these evolving legal standards may affect your AI deployment, data acquisition practices, or enterprise AI risk management, please contact the Author Henry M. Sneath, Esq. in our office at email address sneathhm@hh-law.com or 412-288-4013. Thanks for reading or listening. See you next time.

This blog post represents the author's views and does not constitute legal advice. Readers should consult with qualified legal counsel regarding specific legal issues.