When a regulator comes knocking, the request is never neatly sorted by communication channel. It names an individual, sets a date range and asks what that person said. Surveillance data, however, is structured in the opposite direction.
According to Wordwatch, every channel runs on its own identifier, whether a mailbox, chat handle, mobile number, turret extension or device ID, and none of these is the person. Linking them together is a task someone must complete, and according to 1LoD’s 2026 Surveillance Benchmarking Survey and Report, most banks have yet to finish it.
The research features two surveillance heads at US banks who describe the same shortfall. One notes that their organisation has still not rolled out a unique identifier across all voice channels that definitively rules out duplication or misidentification. The other goes further, pointing to the challenge of recognising an individual across every platform so that voice and e-comms are consistently assigned to the correct person.
The scale of the challenge is growing. Roughly 22% of banks now monitor more than 30 electronic communications channels, up from 14% in 2024, while a third monitor more than 20. An audit at one US bank uncovered around 70 further channels that could require monitoring or restriction.
Each new channel introduces another identifier that must be reconciled with a person, and that reconciliation rarely arrives with the channel itself. At the same time, 24% of respondents cited a lack of standardised data formats or identifiers as the data issue most hampering their surveillance, placing it level with fragmented data across silos. In effect, this is an identity problem expressed in the language of data architecture.
The consequences are significant, even if they never appear on a dashboard. Detection is the first casualty. Channel hopping can only be spotted when activity is tied to one individual, meaning a conversation that begins on Teams and ends on WhatsApp appears as two disconnected fragments.
Review workloads follow, as one person appearing under four identities triggers duplicate alerts that analysts must triage separately. When an investigation begins, rebuilding an individual’s activity becomes a manual exercise across multiple consoles, slowing retrieval and weakening chain of custody.
Identity resolution is often treated as a surveillance feature, but it sits poorly there. A surveillance system only sees the data it ingests, so mappings must be rebuilt whenever a tool is replaced or a channel added, and legacy estates remain out of reach. The record layer offers a stronger foundation, since every channel passes through it on entry and the mapping endures regardless of what sits above it.
A simple internal test is worth running: choose an employee with five years’ tenure and measure how long it takes to produce everything they said last quarter across every channel. If the answer involves several consoles and heavy manual assembly, identifiers were never linked to the person before records were stored.
Read the full Wordwatch post here.
Copyright © 2026 RegTech Analyst
Copyright © 2026 RegTech Analyst
