← All sectors / The AI transformation
098 · Payments fraud & chargeback management
The cost of saying yes
Curve position
Binding constraint
Merchant willingness to share data with a network they do not control.
Why this sector sits at Launch Pad
Adoption here is rising steadily rather than inflecting. The binding constraint is trust and workflow integration, and neither has been broken by a specific event. Buyers are running pilots and expanding them at their own pace.
We place this at Launch Pad because nothing dated has changed. There is no regulation with a compliance deadline, no certification, no cost threshold crossed. Gradual improvement is not an inflection, and calling it one would make the position meaningless.
It moves to Liftoff when we can name the event that changed the rate of adoption, not merely observe that adoption grew.
Reviewed on a two month cycle. The position moves only when a dated, verifiable change in the binding constraint justifies it.
Merchants lose more revenue to declining good customers than to actual fraud, and most of them cannot tell you the ratio. That measurement gap is the whole opportunity: a system that approves more genuine orders while catching more fraud pays twice.
Historically fraud tools optimised for one number, the fraud rate, because that was the number the processor reported. Optimising it alone produces aggressive declines and quiet revenue loss.
The structural driver is margin. Commerce margins are thin enough that a few percentage points of falsely declined orders is the difference between a good year and a bad one, and finance teams have started asking for the figure.
The technology layer spans device and behavioural signals, network level data shared across merchants, machine learning scoring, chargeback representment automation, and the guarantee products that shift liability from merchant to vendor.
Adoption economics are unusually testable. Merchants can run a challenger model against their incumbent on live traffic and measure both approval and fraud, which makes this one of the few enterprise purchases with a clean experiment.
The beneficiaries include fraud platform vendors with guarantee products, chargeback management specialists, device intelligence providers, and the payment processors bundling fraud into their stack.
The value chain runs from checkout signals through scoring to authorisation and dispute resolution. Network data is the moat, since a vendor seeing more merchants detects better.
The overlooked layer includes chargeback representment firms, identity verification vendors for high risk categories, subscription billing recovery specialists, and the analytics tools that measure false declines at all.
Competitive dynamics reward scale in transaction volume, which creates a reinforcing advantage that is difficult for new entrants to overcome.
Risks: processors bundling fraud into their base offering compresses the standalone market, guarantee products carry real balance sheet risk, and a large fraud event can impair a vendor severely.
What to watch: approval rate improvements disclosed by merchants, chargeback ratios, guarantee product loss rates, and processor bundling announcements.