The Board Wants AI Right Now, but Nobody Left a Credit Card on the Desk

The prevailing consensus in corporate suites is that generative software will immediately strip efficiency drag out of legal departments, allowing in-house counsel to absorb massive workloads without adding staff.

It is a tidy corporate narrative, but it ignores the cold mechanics of how software gets funded. A new survey of 510 in-house legal and finance leaders published by Axiom reveals that 92% of corporate AI mandates arrive with exactly zero dedicated budget attached. To satisfy executive directives, general counsel are forced to quietly strip resources from their existing technology stack, reallocating dollars from functional infrastructure to pay for unbudgeted vendor seats.

This is not a digital transformation. It is operational cannibalization.

This is not a digital transformation. It is operational cannibalization.

The Budget Shell Game

When corporate leadership orders a department to adopt machine learning without expanding its budget line, the money must come out of the room somewhere else. In-house departments are re-routing funds from contract management, document repositories, and security software just to purchase user licenses for generative tools.

The financial strain compounds at the exact moment corporate legal departments try to lean on external counsel for help. Rather than offering relief, law firms are treating software deployments as an excuse to raise their bills. Axiom’s research shows that 59% of in-house respondents reported their outside law firms hiked their fees while framing their internal AI usage as a premium value-add. The client pays for the software mandate twice: once by gutting their internal tech budget, and again when opening outside counsel invoices.

Absorbing the Unmanaged Risk

Unfunded deployments do not just destabilize tech stacks; they introduce severe liability that under-resourced legal teams are ill-equipped to manage. As vendors rapidly push generative capabilities into courtrooms and corporate workflows, regulatory and judicial boundaries are shifting underfoot.

Consider how quickly courts are acting on synthetic media and algorithmic liability. In Minnesota, a federal judge recently turned back a challenge by xAI against state deepfake regulations that impose penalties up to $500,000 for nonconsensual sexualized media. Meanwhile, international bodies like China's Supreme People's Court have begun issuing national guidance on algorithmic IP infringement and training data disclosures. Across our own tracking database, judges have issued 775 AI-related sanction incidents against legal practitioners who failed to verify their software outputs.

When executive suites demand instant software integration without funding proper governance, specialized personnel, or testing protocols, the resulting ethical and regulatory exposure sits squarely on the general counsel’s desk.

The Price of Inaction

Skeptics will argue that pushing departments to adopt technology under lean conditions is simply standard corporate discipline, forcing teams to trim legacy software fat. If a tool works, the argument goes, it will rapidly pay for itself in saved hours and reduced outside counsel spend.

That argument overlooks the physical mechanics of software integration. Enterprise tools require configuration, security vetting, data sanitation, and continuous human oversight. You cannot starve the operational engine and expect the car to run faster. When you force a legal department to absorb complex software integrations without funding, you get uncalibrated search runs, unverified briefs, and overworked lawyers attempting to audit machine outputs on lunch breaks.

If the C-suite wants the productivity gains of enterprise legal technology, it needs to write a check for the infrastructure required to run it safely. Until then, general counsel should stop playing the budget shell game and hand the mandate back.