Has AI solved its ROI problem?
AI is now a material part of the risk and compliance technology stack. A directional comparison of two Parker & Lawrence Research studies suggests that AI spending is equivalent to roughly 30% of wider risk and compliance technology budgets.
In previous years, AI had a clear ROI problem. This year, the reported returns are encouraging, although uneven. All 300 senior risk and compliance professionals in our 2026 study selected a positive realised ROI band. More than a quarter, 27.7%, report returns above 50%, while 13.3% sit in the 1–10% band.
These are self-reported estimates rather than audited profit measures. Almost one-third of respondents still identify difficulty measuring ROI as a barrier to AI adoption, particularly where the value comes through avoided losses, stronger detection or resilience.
Spending per employee does not explain the difference. ROI leaders spend around 36% more per employee than ROI laggards, yet firms in the highest spending quintile are no more likely than the rest of the market to report returns above 50%: 26.7% versus 27.9%.
For this analysis, ROI leaders are firms reporting returns of 51% or more; ROI laggards report 1–10%. Our report derives five key lessons from the success of ROI leaders. Here are three:
Connect knowledge to action. Leaders are more likely than laggards to use information retrieval (+20.2 percentage points), structured text generation (+18.2 points) and workflow orchestration (+17.5 points).
Target business outcomes. Leaders are more likely to report faster onboarding and approvals (+15.9 percentage points), faster product launches or market expansion (+15.6 points), and lower technology costs (+12.6 points).
Deploy beyond isolated use cases. Leaders score 4.9 points higher on AI Deployment Intensity, which combines deployment depth, autonomy and decision impact.
The largest differences sit in the operating chain: finding the right evidence, turning it into a usable output and moving it through a process tied to a measurable business result. Model access and a larger budget do not create that chain by themselves.
These are descriptive relationships, not proof that any one capability causes a higher return. They do show where the distinction between access and implementation becomes visible. The full report examines all five lessons and a lot more, including the model mix used by ROI leaders and the implementation barriers that appear as programmes mature.
Download AI in Risk & Compliance 2026 for the full analysis of ROI leaders, laggards and deployment patterns.
