Reasoning is the New Services

Reasoning models will fundamentally change how we businesses run. It’s no longer just about automating tasks; now, entire professional services—like legal, financial, or even healthcare—can be completed by AI at scale. What’s truly exciting is that these models aren’t just glorified autocomplete engines. As these models get better at reasoning, it's almost a given that soon they will be able to reason about complex and specialized tasks better than we can.

That’s the real power of reasoning models: they’re on the cusp of making high-level decision-making capabilities available at scale, whether for legal audits, financial risk assessments, or beyond.

What Does this Mean for Professional Services

Some people claim that reasoning models will kill off service industries entirely—that they’ll steal jobs because they’re on the path to “AGI.” In reality, though, these models may be the best thing that’s ever happened to professional services.

Take a lawyer, for instance. Today, many lawyers spend maybe 10% of their time interacting directly with clients, while the remaining 90% is swallowed up by drafting documents, conducting research, and doing other manual tasks. As AI improvements continue at an incredible rate, it’s possible to imagine a future in which lawyers can devote 90% of their time to client interaction, delegating the tedious work to an AI model. What do you think they’d prefer?

Example: Scenario Planning

Scenario planning is a prime example within financial services. It involves mapping out multiple possible futures—each with its own risks, opportunities, and outcomes—so that businesses aren’t locked into a single forecast. Banks and advisory firms often rely on scenario planning to guide their clients through ever-shifting market conditions. In these scenarios, reasoning models excel because they can synthesize past data, current trends, and broader contextual insights to reason as a human would the range of possible scenarios and highlight the key factors driving each.

By automating the heavy lifting of data analysis and what-if projections, these models free up analysts to focus on interpreting results and advising clients on the most strategic path forward. In other words, reasoning models don’t replace human expertise; they enhance it—allowing professionals to deliver more nuanced, tailored services without getting bogged down in mind-numbing detail.

Is This Possible Today?

It’s partially possible today. The biggest roadblock is the feedback loop, because reasoning models need to learn how professionals think in real-world situations. For example, if we’re building a legal reasoning model, we need practicing lawyers to refine the system, providing examples of how they interpret cases, spot critical details, and arrive at conclusions. Essentially, the model needs a “legal verifier” that teaches it the logic and nuances behind each answer. Once that foundational knowledge is in place, reasoning models can begin to match the thought processes of a trained professional—whether a lawyer, accountant, or physician—making them incredibly powerful allies in complex tasks.

A Future that is More Human

Imagine a future where professionals in law, finance, healthcare, and beyond spend most of their day doing the most rewarding parts of their jobs—advising clients, building relationships, and tackling interesting, high-level problems—while delegating routine tasks and heavy research to AI. That’s the promise of advanced reasoning models. Far from making professionals obsolete, these systems stand to enhance their capabilities, serving as expert coworkers who absorb the drudgery and free up time for human creativity and insight.