Quarterly Briefings Digest  |  Q1 2026

Three Briefings  |  January – March 2026

This digest summarises the three Clustre Briefings held in Q1 2026, covering digital twin applications, the Expected Value (xV) methodology for innovation investment, and the evolving convergence of cyber and physical security threats.

Each session was hosted by Robert Baldock and featured specialist practitioners from Clustre’s curated partner ecosystem.

Briefing 1

Digital Twins: Why do critical systems collapse under complexity?

Context

The session examined digital twins as a mature, deployable technology that enables organisations to anticipate the consequences of decisions, plan more effectively, and improve asset availability. The discussion contrasted two distinct approaches -custom-built solutions versus purpose-designed, off-the-shelf platforms.

Custom Development: The NHS COVID-19 App

Zühlke presented the NHS COVID-19 contact-tracing app as a case study in bespoke digital twin development. Built under extreme time pressure -six weeks to develop, six weeks to test -the app used Apple and Google’s Exposure Notification Framework to create anonymous digital representations of individuals, tracking potential exposures through rotating Bluetooth keys.

Key capabilities included live multi-source data collection, risk scoring across more than 1,000 scenarios, venue check-ins via QR code, and population-level analytics capable of forecasting infection rates two to three days ahead of laboratory results. The system enabled policy simulation and informed variant tracking but required 40–50% population adoption to achieve meaningful effectiveness.

Lessons learnt centred on the underestimated complexity of model design, the time required to build stakeholder trust in outputs, and the necessity of balancing rigid determinism (regulatory compliance) with contextual flexibility.

Off-the-Shelf: Aerogility

Malcolm Bridgeford presented Aerogility, an enterprise digital twin application developed over 15 years, built on agent-based systems. Software agents represent assets, customers, resources, and skills; the platform sits atop existing systems of record (SAP, IFS, Maximo) and plays out scenarios over time to inform decision-makers rather than prescribe answers.

Two client relationships illustrated the platform’s depth. Rolls-Royce Defence (10+ years) uses Aerogility to plan multi-year bids, manage programme disruption, model carbon and environmental metrics, and extend coverage from aerospace into maritime  work recognised with the Sir Frank Whittle Medal. EasyJet (9 years) has achieved an 8% improvement in its maintenance shield against an annual spend exceeding £100 million, with expanded use cases in fleet optimisation and procurement planning.

Aerogility is best suited to complex assets with multi-year planning horizons where optimisation delivers significant financial benefit. The session also addressed how generative AI can enhance agent-based digital twins, with speakers agreeing that transparency and auditability remain non-negotiable for high-value decisions.

Key Takeaways

• Digital twins are production-ready, the choice between custom and off-the-shelf turns on timescale, complexity, and the availability of domain expertise.
• Both approaches require iteration on data granularity and active collaboration with domain experts to build confidence in model outputs.
• AI -particularly agentic AI – will expand digital twin capabilities, but safe, trusted, and auditable AI is essential where decisions carry significant consequences.
• Many organisations still rely on Excel and MS Project for complex forecasting: a significant opportunity exists to demonstrate the step-change that digital twins provide.

If you wish to discuss this subject further please email innovation@clustre.net and put Digital Twins in your subject line

Briefing 2

Innovation Value: Is innovation a budget-burning vanity?

Context

Simon Hill addressed one of innovation management’s most persistent failures: the inability to determine, with any rigour, whether an idea is genuinely worth pursuing. Organisations routinely over-fund weak projects, kill promising initiatives too slowly, and lack the common language needed to engage finance teams who control budgets. The Expected Value (xV) methodology, developed by Wazoku over three years, provides a practical, financially grounded framework to address this.

The xV Algorithm

Inspired by the Expected Goals (xG) algorithm used in football analytics, xV quantifies the expected economic value of an innovation at any point in its development. The calculation draws on four components:

• Confidence (0–100%): the degree of empirical evidence supporting the idea. Early-stage concepts typically score below 40% until proof of customer willingness to pay is established.
• Predicted Value: a realistic financial projection of the idea’s outcome -not an optimistic pitch-deck figure, but a stage-appropriate honest estimate.
• Time Sensitivity: an urgency multiplier (up to 1.5×) reflecting whether competitive dynamics demand acceleration or whether intentional deceleration is warranted.
• Strategic Fit: a composite assessment of whether this is the right idea for this organisation at this time, encompassing long-term alignment, available capability, and market positioning.

The output – the xV figure – can be compared against investment requests to produce ‘xV efficiency’: the cost of generating each pound of expected value.

Case Study: The Loyalty App

Hill illustrated the methodology through a retail company scenario. An innovation team proposed transitioning a £40 million loyalty app from static blanket offers to a dynamic, AI-driven behavioural engine. The pitch projected £10 million in additional value over three years and requested £4.1 million for a pilot.

xV analysis returned a decisive verdict: Kill. Despite strong strategic fit, the idea generated only £1.05 million in expected value -implying a spend of £3.90 to create £1.00 of value. Crucially, the system did not declare the idea fundamentally flawed; it prescribed specific validation steps and capped initial investment at approximately £100,000 to gather the evidence needed before larger commitments.

Dynamic Recalibration and Kill Credits

Unlike static financial models, xV operates as a living system, recalibrating as pilots run, customer evidence accumulates, market conditions shift, and regulations change. Hill recommended automated recalibration triggers at a 20% xV movement threshold. He also introduced the concept of ‘kill credits’: the validated knowledge generated by a terminated project can improve confidence scores on related future ideas, converting failure into organisational capital.

Key Takeaways

• Innovation investment decisions can -and should -be made with the same financial rigour applied to sales forecasting and financial planning.
• CFO engagement becomes achievable when innovation teams speak in P&L terms rather than qualitative conviction.
• The system encourages small, rapid, low-cost pilots (many under £10,000) rather than large early-stage bets on unvalidated ideas.
• xV is most powerful as a portfolio tool: comparing ideas systematically creates genuine prioritisation discipline.
• Transparency across all assumptions and decision criteria improves forecast accuracy and organisational accountability over time.

If you wish to discuss this subject further please email innovation@clustre.net and put Innovation Value in your subject line

Briefing 3

Threat Landscape: Are hostile states attacking your infrastructure?

Context

This session addressed a structural weakness in how organisations approach security. Boards have rightly elevated cybersecurity -the NCSC reported 204 severe-consequence cyber incidents in the year to August 2025, up from 84 the prior year -but this focus has created a blind spot for physical and hybrid threats. State actors and sophisticated criminal groups routinely conduct physical hostile reconnaissance ahead of both cyber and kinetic attacks. Recent incidents, including DHL depot fires across Europe and infrastructure disruptions near Heathrow, illustrate the continuing potency of physical and hybrid vectors.

Robert Baldock introduced two complementary specialist firms: Open Horizon (Norway), providing anticipatory threat intelligence, and Vantiq (USA), providing real-time autonomous response. Together, the two platforms address the full threat cycle from forewarning to action.

Open Horizon -Threat Intelligence

Founded by three Norwegians with backgrounds in military and police intelligence, including operational experience in active conflict zones, Open Horizon was born of frustration at the inability to share unclassified but highly relevant threat information with the organisations that needed it.

At the core of the platform is a graph-technology database that automates the NATO intelligence collection and dissemination cycle using AI and large language models applied to more than 260 open-source feeds, all assessed to Admiralty grading standards. Over 8,000 active threat actor groups are tracked -from lone individuals to state-affiliated networks -mapped against geography, tactics, funding, motivation, and prior targets.

Client organisations are ‘dropped into’ the graph, generating a tailored view of which threat actors are most likely to target them and why, together with ranked attack scenarios. A change in geopolitical circumstances -such as the elimination of a state actor’s leadership -propagates instantly across all related nodes, eliminating weeks of manual analyst effort.

Open Horizon also possesses a signals intelligence capability enabling the detection of physical signals associated with state-affiliated threat actors in the vicinity of client infrastructure. The practical implication is significant: rather than knowing abstractly that a state-level threat exists, an organisation can be alerted that a threat actor is present in a specific physical location -providing time to act before an incident occurs. This capability is particularly relevant to Critical National Infrastructure operators.

Vantiq – Agentic AI Orchestration

Nick Bray, Global Head of Defence, Aerospace and Security at Vantiq (former head of the RAF Regiment), presented a platform that has operated for approximately eight years, having evolved from real-time edge compute into an agentic AI orchestrator before the term became widely used.

Vantiq does not manufacture AI products; it orchestrates, integrates, and automates third-party sensing and effector systems. Processing occurs at the edge -at or near the device -enabling real-time responses at very high speed and very low latency, without routing data to a centralised cloud. The platform connects to 157 sensor and camera types, requires no replacement of existing infrastructure, and operates on an open architecture with no sealed black boxes.

To reduce hallucination risk, Vantiq employs a multi-LLM voting mechanism in which three large language models cross-check each other’s outputs, with results categorised by confidence level. Operational decisions are further informed by a graph database encompassing doctrine, tactics, intelligence feeds, and historical data -directly complementary to the Open Horizon graph described by Polly Copeman.

Notable deployments include: Japan’s national disaster management system (the largest known agentic AI deployment in the world); a SoftBank/Toshiba campus in Tokyo with real-time facial recognition and identity management; an 11,000-camera integration across 95 buildings in Saudi Arabia’s financial district; judicial department security in Delhi; 330 US hospital facilities; and defence applications including submarine systems and drone platforms.

Nick attributed slower European adoption to cultural risk aversion, constrained budgets, and -critically -the absence of a sufficiently acute burning platform. Nations in the Middle East and Asia Pacific, facing more immediate existential threats, have moved faster. Vantiq is actively building European partnerships to close this gap.

The Combined Value Proposition

The two platforms are explicitly complementary. Open Horizon provides forewarning -identifying who is likely to attack, why, and potentially detecting their physical presence near a target. Vantiq provides the real-time autonomous response -orchestrating sensors, access systems, cameras, and human alerts to respond at machine speed without depending on an operator to spot the threat. Together, they address the full threat cycle from anticipation to action.

Key Takeaways

• The cyber-physical gap is a material and growing vulnerability: physical hostile reconnaissance precedes many cyber and kinetic attacks, yet most security postures treat these as separate domains.
• Graph-technology threat intelligence, applied at scale with AI-driven dynamic updating, offers a qualitative step-change over static threat assessments.
• Agentic AI at the edge enables response at machine speed -essential when threat actors are already physically proximate to infrastructure.
• The multi-LLM voting approach is a pragmatic safeguard for high-stakes autonomous decisions where hallucination would have serious consequences.
• European organisations are materially behind the Middle East and Asia Pacific in adopting these capabilities; the competitive and security disadvantage is compounding.

If you wish to discuss this subject further please email innovation@clustre.net and put Threat Landscape in your subject line

About Clustre Briefings

Clustre Briefings are curated executive sessions that connect senior leaders with specialist technology practitioners. Each session is hosted by Robert Baldock, Managing Director of Clustre Ltd, and features firms from Clustre’s pre-vetted partner ecosystem. Upcoming Briefings will address autonomy, using AI within mobile apps and the issue of digital sovereignty.

For introductions to any of the firms featured in this digest, contact Robert Baldock at Clustre Ltd  |  robert.baldock@clustre.net