VisionX AI has unveiled AEGIS, a proprietary Large Graph Model (LGM) designed to identify relationships and suspicious patterns across financial transaction networks, alongside TraceX, an AI-powered blockchain investigation platform built on the model. The launch signals the company’s move toward graph-based financial crime intelligence, combining transaction-network analysis with large language models to help financial institutions and investigators trace illicit fund flows, connect potentially related cases, and generate investigative leads. The company is also establishing itself as a standalone AI-native RegTech business within Alpha Ladder Group.
VisionX AI Builds Financial Crime Detection Around Transaction Networks
Financial crime investigations often begin with an individual suspicious transaction, but the activity behind it can span multiple wallets, accounts, exchanges, intermediaries, and jurisdictions. Identifying a suspicious payment is only one part of the challenge. Investigators also need to establish how funds moved, which entities may be connected, and whether apparently separate incidents are part of a coordinated operation.
VisionX AI is targeting that problem with AEGIS, a proprietary Large Graph Model (LGM) designed to analyse relationships across financial transaction networks. The company has introduced the model alongside TraceX, an AI-powered blockchain investigation platform that uses AEGIS to help investigators trace fund flows and identify potential links between cases.
The launch marks a new stage for VisionX, which was originally developed within MetaComp and is now operating as a standalone AI-native RegTech company under Alpha Ladder Group. Its target customers include financial institutions, digital asset businesses, and public-sector organisations.
AEGIS combines graph intelligence with language models
AEGIS stands for Attention-based Entity and Graph Intelligence System. VisionX says the model uses a Graph Transformer architecture and is pretrained on financial transaction graph data.
Rather than evaluating transactions only as individual records, graph-based systems represent activity through connected entities and relationships. In a financial crime context, these connections can include wallets, accounts, individuals, organisations, and the movement of funds between them.
This network-level view can help identify patterns that are difficult to detect through isolated transaction screening. For example, multiple wallets may appear unrelated when reviewed separately but may share connections to the same intermediary or follow similar fund-transfer paths.
VisionX is combining AEGIS with large language models from established providers in a hybrid architecture. AEGIS supplies the graph-based analysis, while the LLM layer supports natural-language interaction, reasoning, and AI-assisted investigative workflows.
The division of responsibilities is significant. Language models can make investigative tools easier to query and use, but they are not inherently specialised in reconstructing complex financial relationships. A graph model is intended to provide that structural analysis, while the language model helps investigators interact with the results.
Tin Pei Ling, Executive Chairwoman of VisionX AI and Co-President of Alpha Ladder Group and MetaComp, said the model is designed to recognise relationships between transactions and identify patterns that may connect activity that initially appears unrelated.
TraceX brings graph analysis into blockchain investigations
The first product built on AEGIS is TraceX, a blockchain investigation and tracing platform aimed at financial crime investigators and law enforcement agencies.
The platform is designed to follow fund movements across interconnected blockchain addresses and transactions, identify suspicious patterns, and surface potential relationships between separate investigations. Its purpose is to turn network analysis into investigative leads that human investigators can examine.
VisionX has outlined several planned capabilities:
- Reverse tracing: Following fund movements backwards through transaction histories to investigate their origins.
- MegaMap network analysis: Visualising broader connections across addresses, transactions, and entities.
- Pattern recognition: Identifying suspicious activity and recurring fund-flow behaviour.
- Case association: Highlighting potential links between investigations that may otherwise be handled separately.
These capabilities could help investigators move beyond address-level screening and examine the wider network surrounding a transaction. However, the company describes several of these functions as planned capabilities. Their availability, performance, and effectiveness in live investigations have not been independently established in the announcement.
VisionX says investigators retain decision-making authority. TraceX provides recommendations and investigative leads, while human judgement and supporting evidence determine how a case proceeds. Decisions are recorded, creating an accountability layer around AI-assisted analysis.
Summer Yu, CEO of VisionX AI, said the platform is intended to help investigators connect apparently unrelated cases and identify the criminal networks potentially linking them.
Why transaction-network intelligence is becoming more important
The launch comes as regulators continue to highlight the complexity of illicit activity involving digital assets. In July 2026, the Financial Action Task Force (FATF) reported that criminals exploit decentralised finance through techniques including chain-hopping, cross-chain bridges, decentralised exchanges, and mixers. These methods can make it harder to follow funds across different services and networks.
The FATF’s July 2026 update on virtual assets also found that 83% of surveyed jurisdictions had passed legislation implementing the Travel Rule, which is intended to support the sharing of originator and beneficiary information in qualifying virtual asset transfers. The figure indicates regulatory progress, but implementation and effective supervision remain uneven.
For compliance teams, this creates a need to combine transaction monitoring with broader intelligence about entities, wallet relationships, and fund movements. Conventional rules-based systems remain useful for flagging known risk indicators, but they can struggle when illicit activity is distributed across multiple transactions or obscured through complex networks.
Graph-based machine learning is one approach being explored to address that challenge. Research published in Scientific Reports in June 2026 examined graph neural networks for cyber financial fraud detection, highlighting the use of network structure and temporal transaction patterns in risk analysis. That research supports the broader technical rationale for graph-based approaches, but it does not validate AEGIS or TraceX specifically.
From compliance tools to a standalone RegTech business
VisionX began as an internal initiative within MetaComp, focused on financial crime and compliance risks arising from the intersection of traditional finance and digital assets.
Its existing platform includes VisionX KYT, which consolidates blockchain risk intelligence from multiple analytics vendors. The company says the service draws on data covering one billion wallets across 25 blockchain networks. Another component, Web2.5 TxM, combines identity, behavioural, and network intelligence for transaction monitoring across traditional finance and digital assets.
VisionX also supports the compliance intelligence infrastructure of MetaComp’s StableX Agentic Network. Becoming a standalone company gives it a route to offer its products beyond the group’s own financial operations.
The expansion could position VisionX across several overlapping RegTech requirements: blockchain analytics, know-your-transaction screening, transaction monitoring, fraud and scam prevention, and investigative case analysis.
The competitive landscape already includes established blockchain analytics and financial crime technology providers. VisionX’s proposed distinction is its proprietary graph model combined with third-party LLMs, rather than relying solely on conventional blockchain risk data or general-purpose language models.
The next test will be operational. Financial institutions and public agencies will need evidence of detection accuracy, false-positive rates, investigation time savings, explainability, and performance across different chains and transaction types. They will also need clarity on how the platform handles incomplete attribution data and cross-chain activity.
AEGIS and TraceX give VisionX a defined technical proposition: use graph intelligence to uncover relationships that individual transaction checks may miss, then use AI-assisted workflows to help investigators examine those connections. Whether that translates into stronger financial crime outcomes will depend on independent validation and deployment results.
Market Landscape
Financial crime technology is moving toward systems that combine transaction monitoring, entity resolution, blockchain analytics, and network intelligence. This shift reflects the growing complexity of investigations involving digital assets and traditional financial accounts.
Key market developments include:
- Graph-based financial crime detection: Analysing relationships among accounts, wallets, entities, and transactions to identify potentially coordinated activity.
- Blockchain analytics: Tracing digital assets across addresses, decentralised exchanges, bridges, and other blockchain infrastructure.
- Hybrid AI architectures: Combining specialised models with LLMs to support structured analysis and natural-language investigative workflows.
- Cross-channel transaction monitoring: Connecting identity, behavioural, and transaction information across traditional finance and digital assets.
- Human-supervised investigations: Keeping investigators responsible for decisions while using AI to prioritise leads and surface potential connections.
VisionX is entering this market with a model-led approach. AEGIS is positioned as its proprietary graph intelligence layer, while TraceX applies that capability to blockchain investigations. Its ability to compete will depend on the quality of its underlying data, detection performance, explainability, and integration into existing compliance operations.
Top Insights
- VisionX AI has introduced AEGIS, a proprietary Graph Transformer-based model designed to identify relationships and suspicious patterns across financial transaction networks.
- TraceX applies AEGIS to blockchain investigations, helping investigators trace fund movements and identify potential connections between separate cases.
- The hybrid architecture combines graph-native intelligence with large language models for natural-language interaction and AI-assisted investigative workflows.
- Planned TraceX capabilities include reverse tracing, MegaMap network analysis, pattern recognition, and case association.
- VisionX’s transition into a standalone RegTech company expands its potential market beyond MetaComp to financial institutions, digital asset businesses, and public-sector organisations.
Get in touch with our fintech expert

