I'm a JD who runs a company that creates SEO-focused content for law firms. (1) It's very difficult to say how work efficiency has improved since AI tools became widely available, as the work has transformed completely. No longer is perfect the enemy of the good; AI can generate grammatically perfect, well-structured content in a matter of minutes. That said, that has freed us up to create better content, faster. We now spend more time with tasks other than writing, such as humanizing, creating new data for the LLMs to ingest, checking for compliance, and SEO optimization. In other words, it takes just as long to create content, but the work has transformed. (2) We use AI to brainstorm topics, generate outlines, rewrite boilerplate sections (e.g. calls to action), edit, and optimize. For legal content specifically, we've found AI is better at structure than substance. It can organize an argument well but still needs human oversight on accuracy and nuance. (3) In the legal industry, it's critical for businesses to consider issues related to client confidentiality, legal accuracy, compliance with advertising rules, and whether the costs associated with vetting AI-generated work product outweigh the efficiency gains in the actual production of that work product. (4) For what we do, we have found that off-the-shelf AI tools have been sufficient, provided we customize the tools to our specific needs - either through standardized prompting, customized projects, or custom GPTs. That said, I suspect that for legal professionals in practice, off-the-shelf tools may be too risky to use in their current iteration. Between reports of chats with confidential private client information being made public, chatbots confidently hallucinating case law in pleadings (resulting in sanctions), and other ethical issues, I would hesitate to rely on non-specialized legal AI tools.
1) At our personal injury firm, our work efficiency has improved approximately 30% since starting using AI. We've found that we've been able to produce more work product in the same timeframe, even with the time used to double-check and verify AI's output and accuracy. 2) Our office mainly uses AI tools on on marketing, social media, and our website. We are able to use AI to produce charts and daily blogs for our website, with accompanying videos and posts for social media. We've found that the use of AI for marketing purposes has allowed us to significantly grow our presence online, especially through the posting of the daily blogs drafted by AI. 3) For our purposes, we use a general AI tool like ChatGPT. If we were using AI for more legal-based tasks, such as legal research or writing briefs, I believe a specialized AI tool would be more effective, as general AI tools tend to have various errors and case hallucinations.
As the founder of HeyOz, I am not a legal professional, but I work closely with counsel and regularly use AI legal tools to run and scale a tech company. From an operator's perspective, AI has significantly improved the efficiency of legal work, particularly for startups and small teams. Work efficiency has improved by approximately 30 to 40 percent. The most substantial gains come from reducing the time spent on initial drafts, reviews, and repetitive back-and-forth on documents. While AI does not replace legal judgment, it dramatically shortens the process. I primarily use AI legal tools for early contract drafts, comparing clauses, summarizing lengthy agreements, reviewing NDAs, identifying common risks in vendor contracts, and preparing questions before speaking with external counsel. These tools are also helpful for high-level regulatory research and compliance checklists, though not for final interpretation. Before subscribing, businesses should assess data privacy and confidentiality assurances, jurisdictional coverage, whether the tool is trained or fine-tuned on current legal sources, and how clearly it explains its outputs. Workflow integration is another crucial factor. A tool that works with existing document systems and review processes offers much greater value than a standalone application. General AI tools like ChatGPT are useful for understanding concepts, drafting initial language, and summarization. However, for matters involving legal risk, specialized legal AI tools are more effective. They are built with structured legal data, citations, and domain-specific safeguards, which leads to fewer errors and greater reliability. When used appropriately, AI legal tools serve as a force multiplier for founders and legal teams, rather than a substitute for qualified legal advice.
These are great questions. I have significant experience with AI tools/Generative AI tools over the past couple of years in the legal space. Answer to Question (1) I estimate our efficiency has increased by approximately 25%. In addition to greater efficiency, the quality of our work has improved. The AI tools we use have streamlined and enhanced specific niche tasks; that said, they do not replace professional judgment. Answer to Question (2) As a legal finance company, we use Theo AI to support our underwriting process, including the following tasks: * Organizing and summarizing case data * Flagging inconsistencies or missing information * Identifying whether the statute of limitations has passed * Supporting risk assessment and case evaluation workflows Theo AI serves as a support tool for underwriters rather than an automated decision-maker. Answer to Question (3) Key factors to consider before subscribing to an AI tool include: * Whether the tool meets your specific needs. While many AI tools show promise, some are ready for immediate use, while others may require further development before they become practical and reliable. * The presence of an intuitive graphical user interface * The ability to customize guardrails * Data security and confidentiality, particularly regarding legal or financial records * Transparency in the generation and use of outputs * To what degree the tool integrates with existing workflows without causing disruption, and the ways the integration can be implemented. Does it work with Zapier? Does it require a developer? These factors directly impact cost and potential downtime. Answer to Question (4) In my experience, general tools like ChatGPT or Claude AI are not sufficient for pre-settlement funding underwriting or legal analysis. Professional tasks such as these require specialized AI trained on relevant datasets and have been tested for anomalies arising from edge cases. General AI can assist with brainstorming or summarization, but professional legal and financial decisions demand more targeted solutions. One of our underwriters (who is a lawyer) wrote about AI and underwriting legal claims: https://expresslegalfunding.com/pre-settlement-funding-underwriting/#the-future-of-underwriting-personal-injury-claims Theo AI: https://theoai.ai/
1. For certain support tasks, I've seen efficiency improvements in the 15-25% range. That gain comes from organizing information, pressure testing ideas, and other administrative functions. 2. I primarily use AI tools for proofing, issue spotting, and organizing large amounts of written material. In my work, AI can be useful for clarifying questions, outlining analytical frameworks, and identifying areas that require closer human review. It is not used to reach conclusions, assess credibility, or substitute for fact specific analysis. 3. Businesses should be clear about what problem they are trying to solve. Data security, confidentiality, source transparency, and error rates matter more than raw speed. It's also critical to understand where human review is still required, particularly when outputs may influence legal decisions or risk exposure. 4. When used carefully, and properly, general AI tools can be sufficient for conceptual work and drafting support. Specialized legal tools can be helpful for research or document-specific tasks, but no tool replaces professional judgment/responsibility. The effectiveness of any system depends less on the tool itself and more on how well the user understands its limitations.
In our high-volume divorce mediation practice in Massachusetts, we use AI for two purposes: 1) to write blog posts that appear on our website and as guest posts on other websites, and 2) to update and cross check financial details across as many as five court documents and to check consistency of asset names and identifiers across these five documents. Clients regularly give us updates and corrections on this information, and AI helps us to update all documents simultaneously and find inconsistencies. This streamlines document preparation and reduces manual cross-checking in our filings. For this clerical task, AI (we use Claude Sonnet 4.5), cuts time and effort by 75%, and so far it, is more accurate than a paralegal doing the work.
Is AI Transforming Your Legal Work Legal work has traditionally been time-intensive, expensive, and heavily dependent on manual processes. From drafting contracts to conducting legal research, even routine legal tasks often require significant time and resources. Today, small businesses, startups, and law firms are increasingly turning to AI-powered legal tools to streamline operations, reduce costs, and improve efficiency. As legal demands continue to grow, AI is emerging as a practical solution for handling repetitive tasks while enabling professionals to focus on strategic and complex legal matters. According to Thomas Reuters 2025 Future of Professional Reports, 80% of legal professionals believe AI will have a high or transformational impact on their work within the next five years - up 10 percentage points from 2023. Critically, as of 2025, artificial intelligence has emerged not merely as a technological advancement but as a strategic ally. Why AI Is Gaining Momentum In the Legal Industry The force behind the adoption of AI in the legal industry is driven by faster, more efficient, and scalable legal solutions. Traditional legal processes often involve long processes, high fees, and manual effort. For startups and small businesses in particular, establishing consistent legal support could prove to be challenging. As organizations operate in increasingly regulated and competitive environments, the demand for quicker contract execution, faster legal research, and proactive compliance has intensified. Businesses can no longer afford delays caused by slow legal workflows. AI tools in the workplace have emerged as a way to improve both workplace efficiency and add value to their products and client services. Most interestingly, the report suggests that AI-powered tech tools could free up the average professional as much as four hours per week within the next year. How AI Legal Tools Improve Efficiency Across Key Legal Tasks Legal professionals, small business owners and startups have reported measurable productivity improvements after adopting AI-powered legal tools, with efficiency gains of more than 40% of usual. It's also no surprise then that 78% of SMB leaders using AI believe it's going to be a game changer. In fact, 71% of small businesses say they plan to increase their AI investment over the next year. Why? Because the payoff is amazing.
For what it's worth, my AI grind has been heavy on the prep side of things. Drafting first draft summaries from scratch PDFs, cleanup of redlines of contracts, side-by-side comparisons of NDAs or settlement agreements. I'll say I've leaned on tools recently that do 1 thing really well more than anything. They're efficient, they're to the point, and they leave me with a clean baseline to work from. Nothing too fancy......just tools that know how to make the machine work without rewriting the machine. One question I ask everyone before they jump into the tool, but who owns the output? Where is it stored? Some of these tools have hairy privacy policies when you start feeding them client sensitive information or proprietary draft language related to litigation. If the backend is logging even pieces of that for training you're hosed. Surface level stuff like organization, structure, basic grammar....general AI can chew through that pretty well. But if I had to be honest, I wouldn't use it for case assessment or jurisdictional analysis unless it was something tailor made for the task. Some lines of risk just can't be rushed; they need to be traced.
I am a commercial and technology law attorney, CPA, and chief executive officer of the law firm Cummings & Cummings Law (https://www.cummings.law) with offices in Dallas, Texas and Naples, Florida. I also teach business and AI law at Florida Gulf Coast University. AI tools have not transformed legal work. They have flooded it with risk. The time savings are marginal when measured against the review and correction required to prevent malpractice. Any perceived "efficiency" vanishes once the tool generates a wrong citation, misses a jurisdictional carveout, or includes language that waives rights under governing law. I have seen tools insert Delaware-style fiduciary language into Texas LLC operating agreements. I have also seen AI-generated contract summaries omit survival clauses, fee-shifting provisions, and indemnity carveouts. This creates unacceptable malpractice risk. The only permissible use case in my practice is limited summarization of raw, non-privileged text, such as extracting dates or counts from deposition transcripts. Nothing involving client strategy or legal import should be entrusted to AI at this stage, no matter how advanced the marketing materials may appear. Before subscribing to any AI tool, we ask one question: who is legally responsible when the tool is wrong? If the answer is "no one," then it does not belong in our toolbox. Most subscription agreements disclaim all liability. Many grant the vendor perpetual rights to your data. That includes sensitive contract terms and deal documents. Once uploaded, you may have no practical way to claw that data back. The bottom line is that AI is transforming the practice of law, but at the end of the day, agentic platforms can't be sued for malpractice. Attorneys can. My profile and credentials can be viewed on my Featured profile and on my website above. Should you have any follow up questions or wish to schedule a Zoom conference to discuss, please email me at chad@cummings.law.