Can AI detect franchise royalty underreporting?

It can flag anomalies, not prove underreporting. Pattern detection across comparable units is a real capability, but the widely-quoted claim that AI lifts royalty collection to 95% has no published basis and should not be treated as a benchmark.

Royalty assurance is a legitimate AI application and a magnet for unverifiable statistics.

What the technology genuinely does

Compares a unit's reported sales against its own history and against comparable units on traffic, seasonality and mix, then flags divergence for human review.

What it cannot do

Establish that underreporting occurred. A flag is a reason to audit, not a finding. Divergence has many innocent explanations — local construction, a competitor opening, a manager change.

Treating a model's flag as a conclusion is how a franchisor damages a franchisee relationship over what turns out to be roadworks.

On the 95% figure

We checked the widely-repeated claim that AI raises royalty collection to 95% and found no disclosed sample, baseline or method. It is unverified.

The defensible position

Anomaly detection reduces the cost of finding cases worth investigating. That is worth having and is a much smaller claim than the ones being sold.

Have a different question?

Ask Vida — answered only from verified data.