How ALSPs And Generative AI Can Work Together

Alternative legal service providers, or ALSPs, are entering a make-it-or-break-it moment to revamp their value proposition for the first time in a long time.

Traditionally, in-house counsel and law firms would turn to ALSPs to help manage dockets of discovery, contract redlines and other high-volume yet important tasks. However, the ascent of generative artificial intelligence is causing many ALSP stakeholders to review their service offerings and products from a fresh perspective.

AI technology, without a doubt, will force ALSPs to rethink how they frame their competitive advantages to law firms and in-house departments. But will this technology make ALSPs irrelevant? The answer is a resounding no.

There will always be a need for a human lawyer’s perspective when offering accurate consulting and analysis. If anything, the technology can help ALSPs build upon and deliver their services even more quickly, efficiently and effectively.

The most significant opportunity for ALSPs with generative AI is determining how to marry the best attributes of artificial and human intelligence to deliver streamlined client services. Fortunately, there are a few areas where ALSPs can make significant strides to leverage AI to expedite turnarounds and deliveries — all while maintaining, if not boosting, their value proposition for clients.

Contract Review

For ALSPs, reviewing and managing standard contracts such as master service agreements, or MSAs, and nondisclosure agreements, have been a bread-and-butter service. Under traditional ALSP delivery models, teams of legal reviewers powered this process, with some technological assistance.

Generative AI, however, will force ALSPs to adjust their approaches to these kinds of projects. This is because generative AI tools can now automate 50% to 60% of contract review tasks — and even provide redlines to boot.

AI can boost the capabilities of ALSPs to help them track inconsistencies, summarize material terms and calendar key performance dates and deadlines.

Once given the time to study and learn a law firm or company’s common contracts, current generation AI can even help generate automated playbooks that lawyers can consult when discussing material amendments or changes to the scope of various NDAs or MSAs. These capabilities include identifying and organizing fallback language, which can commonly arise from choice of law conversations, arbitration provision negotiations and more.

Still, a robot can only partially pilot the contracting process.

When agreements become due for renegotiation, a company or law firm’s best bet is to continue letting attorneys and business development personnel drive the process to give talks a more compassionate, informed touch. The nuances associated with back-and-forth negotiation and relationship building are things a robot cannot replicate.

However, AI can prepare strategies, terms and questions negotiating teams could raise based on the data it digests.

ALSPs must rework their contract review pricing models accordingly — and adjust the role of human intelligence — to account for potential lost revenue. With the adept deployment of AI, ALSP leaders and project managers could allocate more personnel to other growing service lines while reducing overhead around their contract review services.

They also must consider adjusting their pricing to reflect AI technology speed and efficiency.

E-Discovery

Picture receiving piles of digital files for pressing litigation — a common sight in complex cases.

Opposing parties tend to bury each other in digital paperwork rife with nonresponsive data, duplicate correspondence and mixed media. Even parties receiving discovery requests have their work cut out for them, especially when reviewing large data troves for finding and excluding privileged, nonresponsive documents.

Current use cases around generative AI can significantly conserve billables spent around document review and greatly streamline the discovery process. After exposure to common and sporadic document types during the machine learning process, a generative AI module can comb through voluminous documents, videos, audio files and PDF files to flag potentially responsive information based on common keywords and text-audio attributes.

The byproduct of the AI module’s work is a narrowed-down field of responsive documents that parties can review.

While these capabilities can help attorneys streamline the more rote aspects of e-discovery, a generative AI module could fall short in some areas in data review.

For example, lawyers will still need to act as a second line of review to screen out and redact personally identifiable information that the AI module could overlook. In addition, only lawyers are qualified to decide whether to withhold specific digital evidence or correspondence under the various evidentiary privileges, including but not limited to the attorney-client privilege.

ALSPs accustomed to using project staff to review and tag responsive and nonresponsive discovery should rethink how they approach e-discovery for their law firm and in-house clients. After all, the usage of AI will impact turnaround times, the quality of documents under review and the fees an ALSP could charge.

With AI, an ALSP no longer needs to review a quarter-million responsive and unresponsive docs for a high-stakes matter. ALSP teams can instead analyse 5,000 relevant docs from that original trove and charge a fraction of the amount — all while arriving at the same or better results.

ALSPs can use generative AI to have an excellent opportunity to conduct e-discovery far more efficiently for their clients. After all, success in e-discovery hinges on detecting document patterns, flagging potentially nonresponsive items and organizing them by issue and topic focus — all of which an AI module can accomplish with appropriate machine learning processes.

While these adjustments could dent profit margins for e-discovery to some extent, they can help ALSPs take on more work with a more streamlined e-discovery service offering.

Legal Research

ALSPs can also turn to large language models, or LLMs, to help facilitate and expedite their legal research projects.

LLMs work by collecting, analyzing and synthesizing extensive data collections and then distilling it into written responses to prompts and questions. This technology, which already powers emerging tools like OpenAI and legal-sector programs like Casetext, can process complex data and opinions to find and summarize relevant cases and statutes.

LLM-powered research can go a long way toward expediting and streamlining legal research, allowing ALSPs to reduce their internal headcounts on legal research projects without sacrificing effectiveness. However, even with associated staffing reductions, ALSPs can still explore ways to offer premium research services to their clientele, particularly given jurisdictional shortcomings around mainstream tools.

For example, many of the larger case research tools are built on robust access to PACER and federal dockets — making those tools unhelpful for someone with a predominately state-focused docket. Therefore, ALSPs can still deliver value by focusing on niche research areas, be it by locality or dispute focus, while leveraging LLM programs to expedite their turnarounds.

In addition, recent malpractice lawsuits are exposing a notable flaw with generative AI: its ability to hallucinate facts and cases to suit specific inquiries. In one New York case from earlier this year held in the U.S. District Court for the Southern District of New York, Mata v. Avianca, two lawyers were fined after they submitted OpenAI-generated legal briefs that cited fake cases the popular chatbot generated.

Cases such as this emphasize the importance of human intelligence in putting AI’s florid busywork in check — and the value lawyers at ALSPs can bring to these reworked research processes. At the same time, ALSPs using generative AI can devote more efforts toward more substantive legal work while allowing AI to handle the basic research components.

ALSPs Can Accomplish a Lot — If They Pair AI and Human Intelligence

Generative AI does hold promise and potential, so long as ALSPs know how to use it best.

As more and more companies and firms dip their toes into generative AI programs, it will only behoove ALSPs to follow suit — and come out even stronger and more sophisticated with their service offerings.