This initiative uses retrospective facial recognition (RFR) alongside CCTV and local intelligence to identify repeat offenders in crime hotspots, enabling timely, targeted enforcement and improving the disruption of high-harm individuals.
| Stage of practice |
Untested
|
|---|---|
| Purpose |
Prevention
|
| Topic |
Intelligence
|
| Organisation | |
| Contact |
|
| Region |
London
|
| Partners |
Local authority
Private sector
|
| Stage of implementation |
The practice is implemented.
|
| Start date |
|
| Scale of initiative |
Local
|
| Target group |
Offenders
|
Aim
The initiative aims to strengthen the identification and disruption of high-harm, repeat offenders within crime hotspots by integrating Retrospective Facial Recognition (RFR) with CCTV and local intelligence.
It was introduced to address limitations in traditional, business-as-usual policing, where officers may use CCTV for evidential or investigative purposes but do not routinely have the capacity to systematically review footage for intelligence development.
By enhancing identification capability through RFR and building a more comprehensive operational picture, the initiative assists with creating frequent, up to date and actionable intelligence for frontline officers.
Intended outcome
- increased volume of intelligence reports identifying priority individuals and groups operating within hotspot areas
- increased use of RFR to support identification of unknown or repeat offenders
- increased use of Anti-Social Behaviour interventions, including community protection warnings (CPWs) and community protection notices (CPNs)
- increased utilisation of CCTV as a proactive intelligence development tool, beyond evidential use
- improved understanding of offender profiles and activity within hotspot locations, including identification of repeat and linked individuals
Description
This initiative began in October 2025 in response to persistently high levels of crime within the West End. Despite ongoing patrols, investigations and the use of CCTV, offending levels remained high, indicating that business-as-usual approaches were not sufficiently disrupting those responsible.
The initiative was based on the hypothesis that a disproportionate amount of crime was being committed by a number of repeat, high-harm individuals who were not consistently identified, and whose activity was not fully understood. The goal was to develop a clearer picture of those operating within hotspots, enabling more effective identification, mapping of behaviour, and targeted disruption.
To successfully deliver this initiative, four critical resources were required. The absence of any one of these would significantly limit its effectiveness:
- a RFR system capable of processing image-based identification queries
- access to high-quality CCTV covering identified hotspot areas
- engaged CCTV operators able to proactively identify and capture relevant footage
- motivated police officers or staff to review footage and develop actionable intelligence
The initial phase involved establishing strong working relationships with local CCTV control rooms covering the identified hotspot areas. This began with the Local Authority CCTV network, which provided extensive coverage supported by high-quality cameras. Engagement quickly expanded to include a local private security company operating approximately 60 cameras within the area.
Both control rooms were visited in person to build rapport with staff, enabling effective day-to-day collaboration and strengthening operational relationships. This face-to-face engagement was supported by the establishment of ongoing communication channels, including regular email contact, to ensure continuity and responsiveness.
In addition to this, the following activities were progressed as part of this phase:
- delivering clear and consistent briefings on hotspot locations and priority behaviours, ensuring CCTV operators were fully aware of current policing priorities and key issues affecting the area
- requesting that CCTV operators prioritise capturing clear facial imagery, clothing, and distinguishing features, improving the quality and usefulness of footage for identification and subsequent intelligence development
The project was created and managed by a police constable and an intelligence analyst. To make sure the initiative was successful, the small team undertook the following alongside their day-to-day duties:
- developing their knowledge and capability in submitting RFR requests within the force, ensuring requests were completed accurately and in line with force procedures
- carrying out additional research and verification to confirm identities, using a range of intelligence systems and data sources to corroborate RFR results
- Receiving and reviewing CCTV footage on a regular basis, identifying relevant clips and extracting key information to support further enquiries
- conducting follow-up checks to confirm identities, including cross-referencing intelligence systems and liaising with colleagues where necessary
- producing detailed intelligence reports, incorporating analysis of individuals and identifying links between offenders and emerging patterns of behaviour
- maintaining an up-to-date understanding of offender networks within hotspot areas, allowing for a more proactive and informed policing approach
- delivering regular, actionable intelligence briefings to frontline officers, ensuring they were equipped with current information to support targeted patrols and interventions
Firstly, CCTV operators were given specific tasking in identified hotspot areas to focus on specific behaviours. Following this a standard cycle of activity is as follows:
- CCTV operators identify suspicious behaviour within hotspot areas
- footage capturing clear facial images and distinguishing features is obtained
- CCTV operators securely share footage with designated police staff
- intelligence officers review the footage and submit RFR requests
- where identifications are returned, further checks are conducted to confirm identity
- intelligence reports are created, including linking individuals where appropriate
- at the end of the reporting period, a detailed briefing is produced and disseminated to officers deployed in the hotspot areas
This initiative does not incur any additional financial cost. While it would benefit from dedicated personnel, it can be successfully delivered by a small number of officers allocating a proportion of their time, as demonstrated in its current implementation. Similarly, although senior leadership support would be advantageous, it is not essential for the initiative to operate effectively. The model relies on motivated officers and proactive CCTV operators, with the strength of this partnership underpinning the development of timely and actionable intelligence.
Overall impact
It is difficult to attribute impact solely to this initiative due to the absence of formal evaluation and the presence of multiple concurrent activity within the same area. However, there has been a notable increase in arrests for breaches of Community Protection Notices (CPNs). Between October 2024 and October 2025, there were seven arrests for CPN breaches. In comparison, between October 2025 (date of implementation) and May 2026, there have been 12 arrests, despite the shorter time period.
In addition to quantitative indicators, there has been a notably positive response from both frontline officers and CCTV operators. Local authority CCTV staff have provided feedback on the collaborative nature of the initiative, highlighting increased engagement and a more purposeful use of their capability. Operators expressed a clear desire to replicate this approach in other areas and expressed a desire for similar models to be adopted in other areas. Feedback provided includes:
- “We would love to see this strong collaborative work take place in other wards.”
- “I firmly believe this approach to build a collaborative and shared initiative has resulted in both Police and CCTV being able to focus their resources efficiently and effectively, resulting in these offenders being disrupted and reprimanded for their offences.”
Frontline officers have also reported that more frequent, up-to-date intelligence briefings have substantially improved their ability to operate proactively. Body-worn video (BWV) footage has captured officers actively using these briefings on mobile devices while deployed, searching for individuals recently identified through the initiative. This has directly contributed to arrests of wanted individuals, enforcement against those breaching conditions, and increased use of CPWs and CPNs.
In May 2026, approximately 100 submissions had been made to RFR, the majority of which resulted in confirmed identifications following additional checks. This has enabled the development of a significant volume of intelligence, linking over 100 individuals, five of whom have since been formally recognised as an organised crime group (OCG). When considered alongside the increased use of CPWs and CPNs, and the positive feedback from both officers and CCTV operators, this demonstrates that the initiative has, to date, largely achieved its intended outcomes.
Learning
Several learnings have arisen from implementing this initiative:
- Time and resource – RFR processes can be time-consuming, which can delay the development of timely intelligence. Limited officer capacity may constrain the ability to review all available CCTV footage, requiring prioritisation and potentially leaving some material unanalysed. In practice, the volume and speed at which CCTV can be provided may exceed the capacity available to review and process it.
- Structural provision - the structure of local CCTV provision should also be considered. If provision is fragmented across multiple control rooms without central coordination, maintaining consistent engagement can be challenging. If formal agreements or structured processes are not present, the type, quality and frequency of CCTV submissions may vary, with material received on an ad hoc basis. This variability can make it more difficult to sustain partner buy‑in and to establish standardised approaches to information sharing and tasking.
- Existing awareness - forces should consider levels of awareness and understanding of RFR across policing teams. Where awareness is limited, there may be reduced use of the capability, for example through fewer officer‑initiated RFR requests, which may impact overall effectiveness. Securing internal buy‑in across relevant teams may be more challenging where the approach differs from established working practices or where resource pressures limit capacity to engage.
- Actionable intelligence - forces should ensure there is a clear and shared understanding of how intelligence products are actioned. Without consistent processes and expectations, opportunities to act on intelligence may not be fully realised, and responses may be less coordinated or effective than intended.
If implementing this initiative in a new area, the following recommendations are made:
- establish an agreed schedule for CCTV submissions with all relevant control rooms to ensure consistency and timeliness
- allocating multiple dedicated officers or staff to review CCTV footage, produce intelligence, reduce the backlog and improve output quality would be a significant benefit
- provide training and awareness to local officers on RFR, including how to submit requests and utilise results
- hold regular engagement meetings with senior leaders overseeing teams in the area, to ensure clear direction and agreement on how intelligence products should be actioned