Getting Down to Business: Practical Ways Attorneys Can Leverage AI Today
March 2, 2026
This article was originally published in the Topeka Bar Association Briefings (February 2026).
It was August 2022, and I was in the middle of an interview for a 2L Summer Associate position with Foulston Siefkin…and I was showing my interviewers AI-generated images I made on my phone with a new tool called DALL-E-2. I admit, showing off AI-generated cat pictures was an unorthodox interview strategy, but fortunately, nerding out over technology wasn’t enough to deter my current employer.
Even then, with all the flaws of early AI content, I knew I was watching something significant unfold. Some readers may remember early chatbots like Cleverbot, popular in the late 2000s. Cleverbot is to ChatGPT what a unicycle is to a Tesla. Cleverbot was a clunky, useless distraction for a couple of ten-year-olds goofing off after school. When ChatGPT was released in November 2022, I immediately realized I was looking at something totally different. To say a lot has changed since then is an understatement.
The idea of AI went from a mostly useless gimmick, to an entertaining but untrustworthy conversationalist bad at simple math, to a powerful tool capable of analyzing hundreds of documents in seconds. The promise of AI-driven value creation will make leveraging AI the rule, not the exception—like knowing how to use Microsoft Word or “google something” (yes, I used an em dash; no, I did not let AI write this article).
Now it’s January 2026, and I’m drafting this article in Google Docs while I prompt Gemini with questions about the content I’m writing: fact-checking, grammar, spelling, etc. First, it was talked about. Then, it was a fascinating novelty. Now, it has arrived. Although there is still a significant user gap, AI is quickly becoming the next office technology, rivaling the significance of the personal computer, internet, and email.
A Brief Comment on AI and Ethics
A lot has already been said on ethical risks posed by attorneys using AI tools in their practices, but I wouldn’t feel comfortable writing an article advocating for using AI if I didn’t repeat some of those dangers here. First, AI hallucinations are still a problem (though less and less every day), and the models make mistakes, they’re only…artificial…after all. Second, attorneys have an ethical obligation to verify a platform is designed to protect confidentiality before including client information in a prompt. Additionally, attorneys need to be aware of any ethical obligations, including District Court Rules, that apply to the use of AI, such as Shawnee County DCR 3.125. Finally, we’re attorneys; using good judgment is kind of our job.
Four Ways AI Creates Value
So, if AI has really progressed from novel distraction to powerful tool, how exactly can attorneys leverage AI today to benefit their practices? Here are four ways attorneys can use AI to create value for their clients:
1. Legal Research
There are few areas where AI has caused more headaches or embarrassment than the area of legal research for our profession. Yet, legal research remains and is steadily proving to be one of the key areas where AI value-creation is the strongest. Even though some attorneys continue to get egg on their faces when using AI models for research (e.g., filing documents with hallucinated case citations), in the right context, AI can be an invaluable research assistant.
For now, AI isn’t “one-shotting” complex legal research prompts, but certain models can save the attorney time by providing a good starting point with relevant legal citations and potential answers to an initial prompt. Additionally, models that rely on curated legal databases like Lexis+ Protégé or AI-Assisted Research for Westlaw are significantly better, because the AI-generated output links to real legal citations. Although there may be some novel scenario where these models could hallucinate (i.e., cite to a bogus case or statute), I have yet to discover a fake citation. For now, the greater limitation lies in the models misinterpreting and misapplying a real case rather than fabricating a citation whole cloth. Because these models actually interpret cases and statutes, going a step beyond merely copying the content, attorneys should always confirm the AI has properly synthesized case citations.
AI is particularly useful when attorneys need to research areas of the law they have limited knowledge in. As one example, I was recently asked to review an issue regarding transfers between related entities in a highly regulated industry. I had little background on the particular area of law, so I started by prompting Protégé with the legal issue. Within a minute, it found two statutes that controlled the transfer, saving me a significant amount of research time. The point isn’t that I wouldn’t have found those statutes without AI-assisted research—it’s that starting with AI reduced the amount of time it took me to get to the answer, creating value for the client.
2. Document Summarization and Analysis
By this point, everyone should be aware that AI is good at document summarization and analysis; this was probably one of the earliest use cases for AI. Now, most AI models can analyze large amounts of data, providing useful analysis in minutes. Further, models like Protégé are designed to protect client confidentiality, enabling users to upload client documents to these platforms directly, without risking exposing sensitive information. Although batch document review provides a quick way to get a handle on large volumes of documents, attorneys remain ultimately responsible for reviewing those documents and verifying any AI-generated analysis for accuracy.
3. Legal Drafting and Proofreading
Many AI models are increasingly becoming more competent at drafting standalone clauses, paragraphs, and even simple, one-page contracts. Although more companies are building products designed at complex contract/brief drafting, most of these products lack the refinement and utility necessary for widespread adoption. On the other hand, AI proofreading is already very good and can offer great support for legal drafting by reducing editing time and finding more errors.
For example, I recently asked Microsoft Copilot to review a document and it flagged the typo “Manger.” Spellcheck missed this typo, but Copilot correctly determined the word should have been “Manager” based on the context of the document. In the era of AI, I predict typos will become a thing of the past. If we’re still submitting briefs that cite to the “statue” by 2030, something’s gone horribly wrong. Frankly, most of the mainstream AI models already excel at catching these basic errors; however, there is still room for improvement in other areas, such as proper use of punctuation and application of other nuanced grammatical rules.
4. All Those Things That Don’t Require a Law Degree
One of the things I’ve been most surprised by since starting legal practice is just how much time attorneys spend doing things that don’t actually require a law degree: herding cats, collecting documents, reviewing financial information, etc. So much of legal practice is ancillary to those core functions that define the practice of law. For now, I may still be a better attorney than ChatGPT, but ChatGPT is already a much better mathematician than I could ever hope to be (or have any interest in becoming for that matter). And if that is true, then it makes sense that there are ways I can use AI to cut down on the amount of time spent on activities that don’t actually require a law degree. Sometimes, it just takes a minute to stop and ask yourself if this thing you need to do can instead be done by AI.
As one example, I recently had to review a massive .csv file containing tens of thousands of rows of data. I needed to determine if there was a relationship between certain columns that held financial information. Initially, I thought about randomly adding numbers together to try to find the relationship. Then, it occurred to me that AI could find the relationship for me. I copied a sample of the numbers and prompted AI to identify any related formulas. AI correctly identified the relationship between that set of numbers in under a minute, saving me time and the client money. The more familiar and comfortable we get with how this new technology works, the easier it will be to identify real opportunities to create value for our clients—something attorneys should be eager to do.
Conclusion
Until Jarvis gets a bar license, attorneys won’t be out of a job, but that shouldn’t stop us from learning how to effectively leverage AI. Not only will it help you deliver more value to your clients, it will also give you a competitive advantage over those who refuse to learn.
Established in 1919, Foulston Siefkin is the largest Kansas-based law firm. With offices in Wichita, Kansas City, and Topeka, Foulston provides a full range of legal services to clients in the areas of administrative & regulatory; antitrust & trade regulation; appellate law; banking & financial services; business & corporate; construction; creditors’ rights & bankruptcy; e-commerce; education & public entity; elder law; employee benefits & ERISA; employment & labor; energy; environmental; ERISA litigation; estate planning & probate; family business enterprise; franchise & distribution; government investigations & white collar defense; governmental liability; government relations & public policy; healthcare; immigration; insurance regulatory; intellectual property; litigation & disputes; long-term care; mediation/dispute resolution; mergers & acquisitions; noncompete & trade secret litigation; oil, gas & minerals; OSHA; privacy & data security; private equity & venture capital; product liability; professional malpractice; real estate; renewable energy, storage, and transmission; securities & corporate finance; startup/entrepreneurship; supply chain management; tax-exempt/not-for-profit organizations; taxation; and water rights.
