Skip to main content

How Lawyers Get More Accurate Results From AI (Our ALFN Article)

Mosaic automation interface

In the American Legal & Financial Network's (ALFN) spring 2025 publication, a colleague, Adam Ringel, and I wrote about how legal professionals can get better, more accurate results from AI. Our answer: the difference comes from how you work with the model (specific context, a clear request for the output you want, and a willingness to iterate), not from which tool you buy. The whole issue centered on AI, with other articles on knowledge management, when to trust automation, and how to implement it.

For our piece, Adam and I wanted to focus on the legal user and provide some insights into AI that could help maximize productivity, with the added benefit that the techniques shared would also increase AI accuracy in a highly skeptical industry.

Our article, "Unlocking AI's Power in Legal Practice," doesn't shy away from being technical, but tries to offer practical insights for legal professionals approaching their work with AI models.

You can check out the full publication here -- our article can be found on page 32: ANGLE Vol 12 Issue 2 (Spring 2025)

What the Article Covers

The central premise of the piece is that the gap between "AI that kind of works" and "AI that actually improves your work" is almost entirely about how you engage with it — not about the tool itself. Legal professionals tend to approach AI the same way they approach new software: get access, figure out the basics, move on. But AI rewards a different kind of engagement, and most users never get there on their own.

Adam and I focused on prompt quality as the highest-leverage skill a legal professional can develop. Not in the sense of memorizing templates, but in the sense of understanding what makes a prompt useful: specificity about context, clarity about the desired output, and a willingness to iterate rather than accept the first response. These habits compound. A practitioner who develops them works faster and gets more accurate output — which matters considerably in an industry where accuracy isn't optional.

We also spent time on the skepticism question directly, because it's the right instinct applied in the wrong direction. Healthy skepticism about AI outputs is appropriate. Blanket skepticism that prevents adoption means falling behind peers who are doing the same work in half the time. The article tries to give legal professionals a framework for knowing when to trust AI output, when to verify it, and when the task isn't suited for AI at all. That distinction is more useful than any general endorsement or dismissal of the technology.

The piece is deliberately practical — it's written for the attorney or paralegal sitting down with an AI model to draft a memo or research an issue, not for a technology audience.

Why This Matters for Law Firms Right Now

Legal is one of the industries where AI productivity gains are most measurable and most uneven. Firms that have invested in structured adoption — even informally — are seeing real time savings on research, drafting, and document review. Firms that haven't are doing the same work they were doing two years ago, at the same cost.

The barrier usually isn't access to AI tools. It's the absence of anyone helping the team actually use them well. That's the gap Fractional AI Leadership is designed to close — not by deploying technology for its own sake, but by building the practical habits and internal clarity that turn a tool license into a productivity advantage.

If the themes in the article resonate with where your firm is, I'm happy to talk through what a structured adoption approach would look like for your team. Reach out here.

Frequently Asked Questions

What is the ALFN article about?

"Unlocking AI's Power in Legal Practice," by Drew Bloom and Adam Ringel, appeared in the spring 2025 issue of ANGLE, the American Legal & Financial Network's publication. It argues that the gap between AI that kind of works and AI that improves legal work comes down to how practitioners engage with it, and gives attorneys and paralegals practical techniques for better, more accurate results.

How can lawyers get more accurate answers from AI?

Treat prompting as a skill. Give the model specific context about the matter and the task, say exactly what output you want and in what form, and iterate instead of accepting the first response. These habits compound: a practitioner who builds them works faster and gets more accurate output, which matters in an industry where accuracy isn't optional.

Should lawyers trust AI output?

Neither blanket trust nor blanket skepticism works. Healthy skepticism about any single output is right, but skepticism that prevents adoption leaves a firm behind peers doing the same work faster. The article offers a way to decide when to trust AI output, when to verify it, and when a task isn't suited to AI at all.

Where can I read the ALFN article?

The full spring 2025 issue of ANGLE (Volume 12, Issue 2) is available as a PDF on the ALFN website, and the article starts on page 32. The whole issue focuses on AI, with other pieces on knowledge management, when to trust automation and how to implement it in legal practice.

Drew Bloom, founder of Mosaic Solutions

Drew Bloom

Founder, Mosaic Solutions — Business Systems, Software & Automation

Drew is the founder of Mosaic Solutions. He spent several years as an automation consultant and Innovation Lead for Affinity, where he worked with law firms and enterprise clients on AI strategy, change management, and custom software. He works with small and mid-sized businesses in Iowa and beyond to fix the systems that slow them down — with better processes, software, and AI that actually works.

More about Mosaic LinkedIn Work with Drew

Train Your Team on the Work They Already Do.

AI training for staff and leadership, onsite across Iowa or virtual anywhere, built around your own tools and workflows.