Why AI Alone Does Not Improve Legal Workflows

A written interview with Keao Caindec, Co-Founder and CEO of Clarra

Artificial intelligence has become a frequent topic of discussion across legal practice and legal operations. How would you characterize the current stage of adoption?

Interest in artificial intelligence has accelerated as general-purpose models have demonstrated the ability to perform language-based tasks that appear directly relevant to legal work. Many organizations are actively experimenting with applications across research, document review, contract analysis and litigation support.

What we are seeing right now is experimentation happening across multiple functions at once, often without a single coordinated implementation strategy. Teams are testing where AI can reduce manual effort while also assessing reliability, governance considerations and workflow fit.

This pattern is typical when a capability becomes widely accessible. The conversation has largely moved past whether AI has relevance to legal work. The more immediate question is where it produces dependable value and what conditions allow that value to extend beyond isolated use cases.

Some commentary suggests that advances in large language models could reduce the need for certain types of legal technology. How do you interpret this perspective?

New technology rarely removes the need for software altogether. More often, it shifts where the real value sits. Many legal tools were built to support specific tasks, such as drafting documents or tracking deadlines. As general-purpose models become more capable of handling those types of activities, the value of standalone task automation naturally begins to shift.

However, legal work does not happen as a set of disconnected tasks. It unfolds across a lifecycle shaped by procedural requirements, strategic decision-making, outside counsel coordination and ongoing evaluation of risk. Those relationships require context that persists over time.

As AI capabilities become more widely accessible, the question becomes less about the ability to generate text and more about the ability to connect outputs to the broader matter environment in which legal decisions are made. The emphasis shifts from isolated tools to systems that maintain continuity across matters and make information usable in context.

Where are organizations encountering friction as they begin applying AI tools in real legal workflows?

In many legal organizations, information relevant to a single matter exists across multiple environments that evolved independently over time. Documents reside in shared drives, deadlines are tracked in spreadsheets, financial information lives in billing systems and matter status is communicated through email.

These approaches can function for individual tasks, but they often make it difficult to maintain a consistent view across matters. When multiple versions of key information exist, synthesizing portfolio-level insight becomes more complex and time-consuming.

Introducing additional tools can further increase the number of places where information is created or updated. Without clarity on where the definitive version of information lives, fragmentation can increase rather than decrease.

AI reflects these conditions. When inputs are inconsistent, outputs often require verification. Teams may find themselves spending as much time validating information as they save generating it. For many organizations, improving consistency across matter information produces immediate operational benefits, independent of any AI initiative.

How does information structure influence the reliability of AI-generated insights in legal environments?

AI performs most reliably when the information it draws from is captured in a consistent way. In legal environments, that often means using shared definitions for matter types, aligning on how exposure or risk is categorized and clearly assigning responsibility for keeping key information current.

When matter information is structured consistently, relationships among deadlines, documents, financial data and case status become easier to track and interpret. This allows teams to look across multiple matters and identify patterns, priorities and potential risks without needing to manually reconcile conflicting information.

When AI is applied within systems that already maintain this context, outputs can be evaluated against information teams already trust. When the underlying information is scattered across separate tools or maintained inconsistently, additional effort is required to verify results before they can inform decisions.

Consistency alone does not create insight, but inconsistent information makes reliable insight difficult to achieve. For organizations seeking dependable analysis across matters, structure becomes a practical prerequisite.

What practical steps can legal leaders take to improve the usefulness of AI within their organizations?

Progress begins with relatively straightforward operational disciplines:

• Maintain a centralized system of record for matters

• Define consistent naming conventions and key data fields

• Clarify responsibility for maintaining matter information

• Reduce duplicative tracking across spreadsheets and email threads

• Establish consistent reporting structures across matters

• Identify which information sources should be relied upon

  
  

These steps are not specific to AI, but they significantly influence whether AI-generated outputs are interpretable and dependable.

When these foundations are in place, AI can operate within clearer parameters and contribute to analysis rather than introducing additional uncertainty. In this sense, operational maturity often determines how much value organizations derive from new technical capabilities.

How do you see the role of legal professionals evolving as AI capabilities continue to develop?

Legal work involves judgment, interpretation and accountability within defined professional frameworks. AI systems can assist with information synthesis and identification of relevant materials, but the responsibility for decision-making remains with practitioners.

In many environments, reducing administrative effort allows more time for analytical work, strategic evaluation and client interaction. This shifts how time and effort is allocated without changing the underlying nature of legal work.

Developing familiarity with how AI systems generate outputs, where those outputs are reliable and where additional validation is required will become an increasingly relevant competency.

AI may change how work is performed, but professional responsibility for legal outcomes remains unchanged.

Looking ahead, what distinguishes organizations that are deriving sustained value from AI initiatives?

Organizations seeing sustained value from AI tend to treat it as part of a broader effort to improve how matter information is captured and maintained over time. Rather than approaching adoption as a standalone technology decision, they look at how new capabilities fit within existing workflows and reporting practices.

This often leads to a more incremental approach. As matter information becomes more consistent and easier to access, additional uses of AI become easier to introduce and evaluate without disrupting how teams already work.

Over time, results tend to depend less on choosing a particular tool and more on maintaining an environment where information is current, connected and usable across matters.

Access to AI is expanding quickly. The organizations benefiting most are those making sure the underlying information it relies on is reliable in the first place.

About the Author

Keao Caindec is the CEO and co-founder of Clarra, a fast-growing legal practice management platform redefining how firms and organizations manage matters. A veteran technology leader and entrepreneur, he has built and scaled companies that have transformed mature markets across legal, finance and technology. At Clarra, Caindec introduced a litigation-focused, docket-driven approach that has quickly gained traction among midsize plaintiffs’ firms, AmLaw 100 firms and corporate legal teams. Before founding Clarra, he held leadership roles at Farallon Technology Group, Mocana, 365 Data Centers, OpSource, Yipes, and CyberCash, all of which achieved successful acquisitions. An active member of ILTA, ABA, and CLOC, Caindec frequently writes and speaks on legal technology and the business of law.