Multi-party email scams expose the limits of email security

At a glance

What this is: This analysis shows how multi-party email scams combine lookalike domains, AI-written persuasion, and fabricated trust chains to defeat single-message email defences.

Why it matters: It matters because IAM, accounts payable, sales, and security teams need controls that verify relationship context, not just sender reputation or message content.

Context

Multi-party email scams are a form of relationship-driven fraud in which attackers impersonate several people inside the same business conversation to create credibility. The core problem is not just spoofed email, but the abuse of trust across a transaction chain, which makes isolated message inspection a weak defence.

For identity and access teams, this is an adjacent but important governance problem: business email workflows, approval paths, and payment verification processes all depend on assumptions about who is already known, who can vouch for whom, and what normal correspondence looks like. Once attackers can manufacture that context, traditional email security and human vigilance both degrade.

The article's examples show that these campaigns can span legal, sales, and operations workflows without requiring a sophisticated technical intrusion. That makes the issue relevant across human IAM, fraud prevention, and security operations rather than only email security tooling.

Key questions

Q: How should security teams stop multi-party email scams in business workflows?

A: Focus on the business process, not just the message. Require independent verification for payment changes, vendor onboarding, and order fulfilment, then add behavioural detection that scores conversation patterns, domain age, and relationship history. The goal is to make fraud harder to advance through normal approvals, even when the email content looks polished and plausible.

Q: Why do lookalike domains and AI-written emails still fool staff?

A: Because people validate a request using context, not just syntax. A polished email from a convincing domain can fit the expected role in a transaction, especially when it arrives in a thread that already appears to include real participants. When the control depends on noticing a bad-looking message, attacker-crafted credibility often wins.

Q: What are the warning signs that a transaction thread is fake?

A: Watch for same-day domain registration, unfamiliar senders entering a known relationship late, unusual email applications, urgent payment pressure, and requests that expose missing process details. One weak signal may be normal, but several of these together indicate that the conversation may be fabricated rather than merely suspicious.

Q: What should organisations do when a fraudulent payment request reaches finance or sales?

A: Escalate through a separate validation path before any money moves or goods ship. Confirm the requester through a trusted number or internal directory, compare the domain and account history with known records, and pause the workflow until the approval owner has reviewed the anomaly. That containment step matters more than trying to judge intent from the email alone.

Technical breakdown

How fabricated multi-party threads gain credibility

Attackers do not need to compromise every mailbox to create a convincing conversation. They can register lookalike domains for the supposed participants, seed the thread with plausible roles, and use prior transaction details to make each reply appear consistent. The deception works because recipients validate the conversation socially, not cryptographically. Once one person engages, each follow-up message adds legitimacy for the next participant.

Practical implication: treat multi-party thread validity as a workflow control problem, not only a mailbox filtering problem.

Why generative AI and phishing-as-a-service change the scale of fraud

Generative AI lowers the effort required to produce polished, personalised messages, while phishing-as-a-service packages the infrastructure for mass deployment. Together they turn social engineering into repeatable operations, with template quality, persona creation, and campaign management separated from attacker skill. That shifts the bottleneck from writing believable messages to exploiting the right business process at the right moment.

Practical implication: assume that message quality alone is no longer a useful fraud indicator.

Why legacy secure email gateways miss multi-party scams

Legacy secure email gateways are designed to detect bad links, known malicious senders, or obvious impersonation patterns. Multi-party scams often use clean domains, normal language, and legitimate-looking threads, so the malicious signal sits in relationship history and conversation behavior rather than in a single email body. Behavioural baselining is therefore more relevant than static sender reputation when the attack is built around social context.

Practical implication: supplement gateway controls with behavioural detection that can spot out-of-pattern conversations.

Threat narrative

Attacker objective: The attacker aims to divert money, extract goods on trade credit, or collect payment details by convincing the victim that a fake multi-party transaction is real.

1. Entry occurs when the attacker registers lookalike domains for each party in the transaction and inserts them into an apparently routine business thread.
2. Credential or access abuse is replaced by social trust abuse as the attacker uses the fake identities to request payment instructions, quotes, or shipping fulfilment.
3. Escalation happens as the attacker expands the thread with more fabricated participants, using each reply to increase credibility and extract missing details.
4. Impact is achieved when the victim authorises payment, ships goods on credit, or exposes sensitive payment data under the assumption that the conversation is legitimate.

Breaches seen in the wild

• CoPhish OAuth phishing via Copilot Studio: Datadog showed Copilot Studio agents on a Microsoft domain can front OAuth consent phishing and forward stolen tokens; no victims reported.
• Mailchimp breach 2022: Attackers socially engineered Mailchimp staff, used a support tool to export 102 customer lists and exposed customer API keys for phishing.

Read and download The State of NHI & AI Agent Breach Report 2026, covering 150+ breaches impacting Non-Human Identities including AI Agents.

NHI Mgmt Group analysis

Multi-party email fraud is now a relationship-governance problem, not just a message-security problem. The attack surface is the business conversation itself, where trust is accumulated across replies, copied participants, and expected process steps. That means the control failure sits above the inbox and inside the workflow. Practitioners should stop treating email as the unit of defence and start treating the transaction context as the control boundary.

Identity assurance for communication chains has become the missing control plane. The article shows that attackers can fabricate enough of the surrounding business context to make a single request look routine. Traditional sender reputation checks do not answer whether the party, the domain, and the transaction history actually belong together. The practical implication is that identity programmes must cover who can validate a request, not only who can send it.

Context-aware detection is the named concept this category now needs. Multi-party scams succeed because the fraudulent message is often individually plausible even when the conversation is structurally wrong. Behavioural AI that baselines normal communication patterns is therefore more relevant than content scanning alone. For identity and security leaders, the governance question becomes how to encode relationship history into fraud control.

Email security controls that stop at the gateway leave finance and operations exposed. The article makes clear that the scam can proceed through ordinary human approval paths even after suspicious signals appear. That is why secure email gateways need to be paired with payment verification, vendor-change controls, and out-of-band confirmation rules. The practitioner conclusion is that fraud resistance must be designed into the process, not bolted onto the inbox.

Human judgement remains part of the control stack, but it is no longer sufficient as a primary defence. Attackers are using professionally written, personalised campaigns at a scale that exceeds manual review. The governance task is to reduce reliance on memory, intuition, and email tone as trust signals. Teams should assume that any process depending on a person to notice a slightly wrong-looking thread will eventually fail.

What this signals

Context-aware fraud controls are becoming a core identity-security requirement. Multi-party scams exploit the gap between what a message says and what a business relationship should look like. Teams that rely on single-message indicators will continue to miss attacks that are credible only when viewed as a conversation.

Trust signals now need to be governed like access signals. When a reply chain can be manufactured from lookalike domains and AI-generated text, the question is not whether the email is readable but whether the request belongs in the current workflow. Security leaders should design controls that verify relationship legitimacy before approvals, payments, or shipments proceed.

For practitioners

• Redesign payment verification around out-of-band confirmation Require a separate trusted channel to confirm banking changes, payment instructions, and release of funds before any transfer or fulfilment step.
• Baseline normal communication patterns for high-risk workflows Use behavioural signals such as sender history, domain age, reply structure, and conversation timing to flag transaction threads that do not match expected relationships.
• Lock down vendor and customer change requests Treat new payee details, new shipping contacts, and sudden quote changes as identity events that require stronger approval than ordinary email acknowledgement.
• Train frontline teams on multi-party fraud cues Teach accounts, sales, and operations staff to recognise transaction threads that add urgency, duplicate roles, or introduce unfamiliar advisers late in the exchange.

Key takeaways

• Multi-party email scams exploit transaction trust, not just email content, so inbox filtering alone cannot stop them.
• The article shows that attackers can register lookalike domains, use AI-written messages, and escalate credibility across a fake conversation before anyone notices.
• The practical defence is workflow-level verification, behavioural detection, and stricter approval rules for payment and fulfilment changes.

Key terms

• Multi-Party Email Scam: A fraud technique in which attackers fabricate several participants in the same email conversation to create trust and urgency. The goal is to make a request appear routine because it seems to come from a network of related parties rather than a single suspicious sender.
• Lookalike Domain: A lookalike domain is a web address designed to resemble a trusted brand closely enough to trick users into believing it is legitimate. In identity attacks, the domain becomes part of the deception layer, letting attackers capture credentials, identity details, or payments through a counterfeit flow.
• Behavioural email detection: A detection approach that looks for patterns in sender behaviour, message timing, language change, and downstream user interaction rather than relying only on signatures. It is designed to catch attacks that mutate quickly. For identity programmes, its value is in finding the moment an email becomes an access risk.
• RFQ Fraud: RFQ fraud is a social engineering pattern where an attacker poses as a prospective customer requesting a quote, pricing, or product details. It exploits teams that are incentivised to respond quickly to new inbound business, turning sales responsiveness into a control weakness.

Deepen your knowledge

NHI governance, agentic AI identity, and machine identity lifecycle are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If you are building or maturing an IAM programme, it is worth exploring.