Legal teams tend to build AI business cases around legal outcomes. They describe efficiency gains, time savings, reduced risk of error, and improved compliance tracking. These outcomes are real and meaningful. They are also, from a CFO's perspective, insufficiently quantified, strategically underpositioned, and easy to defer.
The GCs who are winning budget approval for legal AI investment are not winning because they have better tools or bigger teams. They are winning because they have reframed the conversation in financial and strategic terms that sit naturally in a CFO's approval framework.
This article gives GCs that framework. The underlying research is drawn from the Plexus Future-Ready General Counsel 2026 report and from the commercial outcomes Plexus customers have documented after implementation.
|
32% of GCs cite budget as the top AI adoption barrier |
40% of legal time currently spent on tasks AI can handle |
60% average reduction in contract cycle time reported by legal AI adopters |
3x typical ROI on legal operating platform investment within 18 months |
Source: Plexus Future-Ready General Counsel 2026 Survey, n=150, January 2026; Plexus customer outcomes data, 2024 to 2025
CFOs approve investments for two reasons: because the investment generates a return that exceeds its cost, or because the cost of not making it exceeds the cost of the investment itself. Everything else is a preference, not a business case.
Most legal AI business cases fail because they are built around the first reason but not quantified rigorously enough to be convincing, and they ignore the second reason entirely.
Here is what a typical legal AI business case looks like in practice:
• The legal team will save approximately 10 to 15 hours per week on contract review
• Compliance tracking will be more consistent and less manual
• The system costs $X per month and will pay for itself in efficiency gains
• Other organisations in our industry are already doing this
A CFO reading this sees four things: an estimate with no methodology, a qualitative claim with no metric, a payback period with no supporting calculation, and a competitive anxiety argument. None of these meet the approval threshold for a capital investment.
The CFO is not being obstructive. They are applying the same financial rigour they apply to every investment request. The legal team has simply not yet provided the information that rigour requires.
|
What legal teams typically present |
What CFOs need to see instead |
|
Time savings (hours per week) |
FTE cost avoided or redeployment value: hours x fully loaded cost per hour x weeks per year |
|
Faster contract turnaround |
Revenue acceleration: contract value x number of deals per year x cycle time reduction as a percentage of deal close timeline |
|
Better compliance tracking |
Risk-adjusted exposure reduction: cost of a regulatory breach or compliance failure x probability reduction from systematic tracking |
|
Reduced external counsel spend |
Documented external counsel cost last 12 months x projected reduction percentage from insourcing enabled by AI tools, with named firm rates and historical volume |
|
Scalability without headcount |
Headcount avoidance: cost of the next legal hire (salary plus on-costs plus recruitment) compared to the cost of the technology that replaces that capacity need |
The most powerful element of a legal AI business case is often the one that goes unbuilt: the cost of doing nothing. CFOs are trained to evaluate investment decisions against a baseline. If the baseline is presented as acceptable or as simply the status quo, the investment appears discretionary. If the baseline is presented as costly, the investment appears necessary.
Here is how to build the current-state cost calculation:
Pull actual data from your team. How many contracts does the legal function handle per month? What is the average time spent per contract on drafting, review, negotiation, and execution? Multiply that volume by hours per contract by the fully loaded cost per hour for the relevant team members. This gives you an annual figure for the labour cost of manual contract management.
For a legal team handling 200 contracts per month at an average of 3 hours per contract, with a fully loaded hourly rate of $150, that is $1.08 million per year in contract management labour. This is not a cost that disappears with AI. But a 40% efficiency gain on that activity represents $432,000 per year in redeployable capacity. That is the number that belongs in the business case.
Pull 12 months of external counsel invoices. Identify the work categories. Segment them into work that could be insourced with AI assistance and work that genuinely requires external expertise. The insourceable portion, multiplied by the external counsel rate, is your external counsel reduction opportunity. Most legal functions find this number is larger than they expected.
This is the number most GCs leave out because it feels speculative. It is not. It requires two inputs: the regulatory or contractual penalty exposure that the organisation currently carries, and the probability reduction that systematic compliance tracking would achieve.
If your organisation operates under contracts that carry termination clauses for non-compliance, and you currently track compliance manually, the probability of a missed obligation creating a dispute or termination event is materially higher than it would be with automated tracking. Price that exposure against the cost of the technology. The comparison is usually stark.
How many deals per year does your organisation close that have a legal review component? What is the average contract cycle time, and what proportion of deal close time does it represent? If legal review is on the critical path for revenue recognition, reducing cycle time has a direct revenue impact. Even a conservative estimate of 2 weeks reduction on a 12-week cycle, applied to 100 deals per year at an average value of $500,000, produces a revenue acceleration figure that changes the conversation.
Once the current-state cost is documented, the ROI calculation follows a standard structure. The table below shows how to build it:
|
Value category |
How to calculate it |
Conservative benchmark |
|
Labour efficiency |
Annual contract management labour cost x efficiency gain percentage |
30 to 40% efficiency gain on automated workflows |
|
External counsel reduction |
Insourceable external counsel spend x reduction percentage achievable with AI assistance |
20 to 35% of external counsel spend insourceable with AI assistance |
|
Risk exposure reduction |
Total regulatory and contractual penalty exposure x probability reduction from systematic compliance tracking |
Use conservative 15% probability reduction if no internal data is available |
|
Revenue acceleration |
Number of deals with legal review x average deal value x cycle time reduction x deal close timeline proportion attributable to legal |
50 to 60% reduction in contract cycle time reported by Plexus customers |
|
Headcount avoidance |
Cost of next planned legal hire (salary plus on-costs plus recruitment fees) compared against cost of technology that provides equivalent capacity |
In most markets, the technology cost is 15 to 25% of the fully loaded hire cost |
Sum the five value categories and compare against the total cost of ownership of the investment: software licence, implementation, internal time to deploy, and ongoing management. The payback period for most legal AI investments, when this calculation is done correctly, falls between 8 and 18 months. That is a capital investment profile a CFO can approve.
The calculation is necessary but not sufficient. A CFO also needs to trust the person presenting it and the methodology behind it. How the case is structured and presented matters as much as the numbers themselves.
The most common mistake GCs make when presenting to a CFO is leading with the product. They open with what the technology does, then work backwards to why the organisation should buy it. This positions the conversation as a procurement decision rather than an operational necessity.
Lead instead with the current-state cost calculation. Show the CFO what the legal function's manual processes are costing the business in labour, external counsel dependency, risk exposure, and revenue delay. Once the cost of the current state is visible, the investment is not a purchase. It is a solution.
CFOs distrust single-point ROI estimates. They are trained to stress-test assumptions. Present three scenarios: conservative (30% of projected efficiency gains), base case (the central estimate), and upside (if adoption is high and cycle time reduction is at the upper end of benchmarks). This demonstrates that the business case has been stress-tested and that even the conservative scenario produces a positive return. It also gives the CFO a model they can interrogate rather than a number they are asked to accept.
Projections require the CFO to trust your assumptions. Comparables require them to assess evidence. Where possible, anchor your business case on documented outcomes from peer organisations or from the vendor's own customer data. Plexus publishes customer outcome data including contract cycle time reductions, external counsel spend decreases, and FTE efficiency gains. This evidence shifts the conversation from what might happen to what has already happened in comparable implementations.
Every quarter without the investment is a quarter of cost accumulation at the current-state rate. If your current-state cost calculation shows $800,000 per year in avoidable cost, each quarter of delay represents $200,000. Name this in the presentation. It converts the approval decision from a budget question into a timing question, and timing questions are easier for CFOs to resolve than budget questions.
The CFO sits in executive conversations where the CEO is talking about growth, efficiency, and competitive positioning. Connect your legal AI business case to those conversations. If the organisation is targeting 20% revenue growth, show how legal bottlenecks on the critical path of deal closing are a constraint on that target. If the organisation is running a cost efficiency programme, show how legal operating costs compare against the technology investment. The business case that is connected to existing executive priorities is the one that gets prioritised.
The GCs who cannot get legal AI approved are not typically in organisations where the CFO has considered the investment and decided against it. They are in organisations where the investment has been presented in terms the CFO does not have a clear framework to evaluate.
The fix is not a better tool or a larger budget request. It is a better business case: one that starts with the cost of the current state, quantifies the opportunity in financial terms the CFO already uses, presents scenarios that can be stress-tested, and connects to the strategic agenda the board is already tracking.
The GCs who have made this shift report that the budget conversation changes fundamentally when the business case is built this way. The CFO is no longer being asked to evaluate a technology purchase. They are being asked to approve a decision that reduces a documented cost and accelerates a documented revenue opportunity.
That is a conversation a CFO is trained to have. It is up to the GC to bring it.
Source: Plexus Future-Ready General Counsel 2026 Survey, n=150 General Counsels, January 2026. External citations: Thomson Reuters Generative AI in Professional Services Report 2025; ACC/Everlaw GenAI Survey 2025, n=657; Gartner Legal and Compliance Leader research 2025.