Want more content like this? Sign up for our monthly newsletter.
There’s a figure that keeps showing up in contracting research: somewhere between 9 and 10 percent of contract value gets lost. That number has been in the data for the better part of a decade. It predates the investment wave in contract technology. It predates AI. World Commerce & Contracting and the CCM Institute have been tracking it. Their latest study, drawing on 120 organizations and more than 20 practitioner interviews, examines why it still hasn’t moved. The answer is that most organizations were losing value through the same mechanisms long before AI arrived, and those mechanisms are still intact. Getting faster at a process doesn’t change what the process loses.
AI is being aimed at the wrong work
Two-thirds of organizations are using AI for individual productivity: drafting, summarizing, reviewing. The process around those tasks hasn’t changed. Only 12% have actually redesigned a process around AI. And when researchers asked what measurable effect AI had on value lost, risk exposure, or revenue captured, not one respondent could name a result.
The reason isn’t that AI doesn’t work, it’s that it’s pointed to the wrong place: pre-signing work, or the same tasks that already got the most investment. Post-signing work gets far less attention, even though poor oversight after signing is a top-three cause of value erosion in every segment of this survey. The gap between what a contract promises and what actually gets delivered lives in execution, not negotiation.
Getting real value out of AI requires a sequence: get faster first, then redesign the process around what AI is surfacing, then connect those insights to people who can act on them. Most organizations are stuck between steps one and two. The real question is whether your process is built to do anything with what it finds.
The metrics most teams use can’t see the problem
Most contracting teams measure efficiency: cycle time, volume, negotiation duration. Those numbers tell you how much work is moving. They don’t tell you whether any of it created or destroyed value.
One organization in this research redesigned its NDA process using approved templates, workflow automation, and e-signature. Turnaround improved 65%, and the organization estimated 800K–1.2M in annual value from personnel-cost avoidance and faster time to revenue. The 65% is not the result. The avoided cost is. Without measuring the business outcome, a team running this process would see a 65% improvement and have no way of knowing what it was actually worth.
The organizations that have demonstrated real ROI measure outcomes: margin protected, obligations met, disputes avoided, change-order value recovered, and they validate those numbers with finance. Without that, any efficiency gain is ungrounded. You can automate contract review while nobody tracks whether the obligations in those contracts actually get met. You can cut negotiation time while creating problems that surface as disputes six months later. The metrics won’t catch it.
If you’re on the buy side, 10% is probably an undercount
The headline figure hides a real gap. Buy-side organizations estimate their losses at roughly 13%; sell-side clusters near 8%. Buyers are further from the revenue line and closer to delivery risk. They’re also doing less to manage it. None of the cross-functional AI platforms in this research belong to pure buy-side organizations. Only 10% of buyers give finance full access to contract data, compared to 43% on the sell side.
If you’re a buyer using the overall figure to gauge your exposure, you’re probably working from the wrong number.
Your finance team probably can’t see your contract data
Only 19% of organizations give finance full access to contract data: what was committed to, how pricing works, what’s owed, when things renew. Almost none of it flows automatically into the systems where financial decisions actually get made. It arrives manually, when it arrives at all.
AI can surface a renewal risk, an unexercised pricing right, a missed obligation. But if that surfaces inside legal or procurement and never reaches anyone with authority to act on it, nothing changes. The same disconnect runs across the organization itself: even among companies operating on both sides of deals, only one in five review their buy-side and sell-side commitments together. What you owe as a customer and what you’re owed as a supplier are invisible to each other.
Most organizations still haven’t done two things: given finance real access to contract data, and put one person in charge of keeping it clean. Without both, AI insights stay inside legal.
Most organizations are fixing symptoms, not causes
The pattern across this research is consistent. Organizations invest in AI before fixing measurement, so there’s no way to know if it’s working. They deploy AI before fixing how contract data is stored and shared, so insights never reach decision-makers. They try to improve post-signing execution before fixing the handover from negotiation, so the same problems keep recurring, just faster now. It’s a sequencing problem, not a tech one. The sequence that works inverts that order. It starts with measurement: understanding how much value is being lost, in which contract types, through which mechanisms, in terms finance can validate. Data infrastructure and ownership come next. AI-led process redesign comes last. Before that foundation exists, AI just makes individuals faster.
There’s also a design problem that neither measurement nor process alone will fix. Contracts need to be built to absorb disruption without requiring full renegotiation, but the clauses that make that possible, trigger-based reopeners and benchmarking rights, exist in fewer than one in ten agreements today. Only 2% of respondents are very confident their major contracts could absorb a significant disruption without renegotiation or dispute; 59% are not very or not at all confident. The mechanisms are well understood, but they’re not being written in.
By the numbers
The findings above rest on a consistent pattern across 120 organizations. Here’s the data behind it, organized by theme, if you want a closer look. All figures are from the ROI of Contracting Excellence 2026 by World Commerce & Contracting and the CCM Institute.
Value erosion
The 2026 estimate of roughly 9–10% compares with 8.6% in 2023 and roughly 9% in WorldCC’s long-run benchmark — it hasn’t moved in a decade. Half of respondents report no change in their own erosion rate over three years. Two in five estimate losses of 10% or more, almost all from project-based industries. Disputes and claims rank last as a source of erosion in every segment — only 6–9% of respondents identify them as a major cause. The damage accumulates in weak post-award governance, missed obligations, and unmanaged scope changes.
AI deployment
62% of respondents say individual AI gains haven’t translated into organization-level results; among organizations on both sides of deals, that’s 67%. Just 4% describe AI as a cross-functional platform connecting contract data to business decisions — and none of those are on the buy side. Among buyers, 80% use AI for drafting and generation and 65% for review and redlining. Obligation tracking and post-signing management trail well behind.
Measurement
No organization in the survey has a formal ROI framework connecting contracting to business outcomes; several interviewees said they had no measurement framework at all. Interest in outcome-based contracting is growing — 40% of buy-side, 38% of sell-side, and 71% of both-sides organizations report increased interest over three years — but the leading barrier is not being able to measure outcomes reliably, cited by 76% of buy-side and 82% of sell-side respondents.
Contract data and finance access
Less than 8% of organizations have contract data that enables portfolio-level analysis. A single point of accountability for commercial data quality exists in only 15% of buy-side organizations. Roles that span legal, procurement, sales, and finance exist in 25% of buy-side organizations, 21% of sell-side, and 28% of both-sides organizations.
Contract design and resilience
Indexation is standard practice for 60% of buy-side organizations, but it adjusts prices by formula — it doesn’t reopen the contract. The mechanisms that require parties to re-engage when conditions shift are far rarer. Trigger-linked reopener clauses are standard for only 10% of buy-side organizations, 21% of sell-side, and 6% of organizations operating on both sides. Benchmarking rights are standard for 15% of buy-side organizations, 14% of sell-side, and zero both-sides organizations. Contract duration is not shortening in response to volatility — 64% of sell-side organizations report it unchanged.
Where to start
Every finding in this report comes back to the same root cause: organizations try to fix contracting problems at the execution layer without addressing what’s underneath. AI gets deployed before anyone has established what success looks like. Data gets generated before anyone has established where it needs to go. Processes get redesigned before the measurement exists to know whether the redesign worked.
The starting point, by this research, is measurement: not cycle time or volume, but commercial outcomes. How much value is your organization losing, in which contract types, through which mechanisms, expressed in terms finance can actually validate? Until that question has a real answer, everything else is optimization without a target.
To dive into more detail and get a ten-step action plan on how to start recapturing lost value, download the full report.
Ironclad is not a law firm, and this post does not constitute or contain legal advice. To evaluate the accuracy, sufficiency, or reliability of the ideas and guidance reflected here, or the applicability of these materials to your business, you should consult with a licensed attorney.



