Do Lawyers Really Need to Touch Every Agreement?

The business sends over a routine agreement. Legal reviews it. The next one looks almost identical, so legal reviews that one, too. Somewhere in the queue is a contract with a risk the business hasn’t seen before, waiting for the same lawyer’s attention.

That is often what we mean when we say legal is protecting the company. The lawyer is involved, so the work must be controlled. But a lawyer reading every agreement is a way of allocating attention. It isn’t, by itself, a way of deciding where that attention is most valuable.

What if the lawyer did the important work before the agreement arrived?

I recently spoke with Curt McDaniel, a former chief legal officer, about a contracting process his team built. They defined the templates, playbooks, internal approvals, and boundaries for lower-value, less complex agreements. Employees could then move those agreements through the process without a lawyer reviewing each one. At one point, he said, 70% of their agreements went through the contract management system without individual lawyer review.

I can hear the pause that number produces. Curt got it from his executive committee, too. The useful question is what those lawyers had done before those agreements moved. They had helped decide which work belonged in that lane, what terms were acceptable, and when an agreement needed to come back to legal.

The judgment was still there. Its timing changed.

Where should the lawyer enter?

Curt described legal work as a pyramid. At its broad base are high-volume matters with relatively low risk and complexity. At the top are the matters where a lawyer’s knowledge of the law, the company, and its customers can make the greatest difference. His aim was to give lawyers more time at the top by finding other ways to handle the base: processes, technology, legal operations, and, for some work, outside support.

That starts with a different question than “How do we get legal to review this faster?” Ask what the business needs to accomplish, what risk the company is willing to accept, and which decisions truly require a lawyer in the moment.

For a basic agreement, the answer may already be in an approved template and a clear path for internal approvals. For another agreement, a change to the terms or the risk involved may call for legal review. The point is to make that distinction deliberately, before the queue makes it for you.

There is real legal work in making those distinctions. Someone has to decide what the standard language should say, what can vary, which approvals belong to the business, and what sends the work to a lawyer. That work may be less visible than marking up the same clause for the fiftieth time. It can also reach far beyond one contract.

Boundaries need a feedback loop

Of course, putting a rule in a workflow doesn’t make the rule right forever. A team needs to see what happens when people use it.

Curt’s team randomly selected agreements that had gone through without lawyer review and checked the result afterward. Had the agreement belonged on that path? Was there an exposure they had missed? The review mattered because the team could move routine work without asking lawyers to approve every instance, while still testing whether its decisions held up in practice.

The same question applies to the exceptions. If an agreement repeatedly comes back to legal for the same reason, is the boundary catching a real risk? Does the template need work? Has the business started doing something the original process did not account for? A lawyer should be able to examine the pattern, not simply resolve the next item and wait for it to recur.

This is where legal operations earns its place alongside legal expertise. Designing the workflow, maintaining the data, seeing where work stalls, and organizing a review of results all take capacity. Curt was clear that his vision depended on people who could implement and improve it. A rule written by a lawyer has little value if the organization cannot put it into practice or tell when it needs to change.

Where a GC could start

Curt didn’t offer a formula, but his team’s approach points to a few things to settle before the first agreement moves without review.

• Draw the lane in writing. List which agreement types qualify and what sends one back to legal. Curt’s team started with lower-value, low-complexity agreements, such as basic service agreements. The ones that needed to be “close to perfect” stayed with lawyers.
• Settle positions in advance. Curt noted that GCs differ on questions like arbitration versus litigation, and that historically there was no way to control which one showed up in agreements. Deciding it once, in the template and playbook, takes it out of each reviewer’s hands.
• Put approvals in the workflow. On Curt’s team, approvals were part of the contract system itself. Employees were empowered to move basic agreements quickly instead of waiting on legal.
• Put the review on a calendar. Curt’s team ran its random-sample review every summer, usually with legal interns, and he expects AI to handle much of that now. He also floated a cadence of every six months or a year for comparing completed agreements to the templates. He didn’t give a sample size, so that’s a number each team will need to set for itself.
• Fix the process before you automate it. Curt’s advice was to make sure processes are lean first: “Don’t automate a bad process.” He also said governance needs policies, training, and escalation, with another layer for AI.

None of this requires AI. It requires deciding, in advance, what lawyers shouldn’t have to decide again.

Read more Thought Leadership

AI changes the scale of the review

AI can help examine completed agreements against templates and playbooks. Curt described a possible use that I think is especially useful: looking back across agreements to find where terms differed from the standard, then asking whether those differences created unnecessary risk or revealed a reason to update the template.

That is a different assignment from allowing AI to approve agreements on its own. The team still has to define what it wants to find, govern the data the tool can access, and decide what to do with the results. Curt also emphasized that lawyers remain responsible for work product and need to understand the tools they use. In some lower-risk workflows, oversight may mean setting boundaries and reviewing results rather than examining every action as it happens.

There is no universal percentage of contracts that should bypass a lawyer. Curt’s 70% describes what worked within the boundaries his team built. Another legal department would have to draw its own lines around its agreements, risk, people, and processes.

But every department can ask the underlying question: Are we using lawyers to make the decisions only they can make, or are we using them to repeatedly apply decisions they have already made?

The answer should change how we think about legal involvement. Sometimes the best evidence of a lawyer’s judgment is a careful intervention on a difficult matter. Sometimes it is a routine agreement that moves safely without ever landing in their inbox, because they helped design the path it followed and the team knows how to check it.