Copenhagen based spektr has added real-time transaction monitoring, policy management and AI-assisted investigations to its Know Your Customer (KYC) and Know Your Business (KYB) platform. spektr 3.0 converts written policies into operational rules, monitors customer activity against them and uses Case AI to gather evidence and recommend an outcome for analyst approval.
- Transaction Monitoring checks customer activity in real time against detection rules compliance teams build and control, with no engineering support required. Alerts are routed into the same customer record, risk profile and investigation workflow used for onboarding.
- Policy Engine turns an institution’s written policies into the rules the platform runs on. Compliance teams approve every rule before it takes effect, while versioning records which rules applied when so decisions can be explained to regulators.
- Case AI uses AI agents to run investigation tasks, gather evidence and recommend an outcome with citations and a confidence level, allowing analysts to focus on cases requiring judgement. Compliance teams remain responsible for the final decision.
The launch comes as banks reassess their anti-money laundering controls. spektr’s analysis identifies 13 FCA fines worth more than £300 million since 2021. PwC’s 2026 EMEA AML Survey found that 61% of banks plan to invest in transaction-monitoring technology before July 2027. Confidence in existing systems had fallen below 30%.
RegTech Insight spoke with spektr chief executive and co-founder Mikkel Skarnager who summed up what regulators are demanding: “Show us how you reached this decision,” he says.
From Onboarding to Compliance Operations
Skarnager’s background includes PayTech firm Nexi and Saxo Bank, where he worked on customer onboarding across markets and client types. Skarnager and co-founder Ciprian Florescu went on to create HelloFlow, a no-code onboarding platform that was acquired by Identity verification provider Trulioo in 2022. In 2023 Skarnager, Florescu along with Jérémy Joly and Jan-Erik Wagner went on to found spektr. The company raised €5 million in seed funding in 2024 followed by a New Enterprise Associates led $20 million Series A in April 2026, taking total funding to $26 million.The team has expanded its focus from onboarding into customer monitoring, risk assessment, remediation, transaction monitoring and investigations.
Skarnager traces spektr’s design to a common compliance problem. Analysts repeat defined tasks across large case volumes, drawing information from separate systems and tracing ownership through layers of legal entities. Each task may be manageable on its own, but the combined workload is not.
“At their core, these tasks are not that difficult, there’s just a lot of them,” he says.
spektr breaks that work into individual steps that its process engine can assign to an analyst, a deterministic tool or an AI agent.
Policies Set the Boundaries
The spektr platform is built around a process engine and a policy engine. The process engine structures the steps within onboarding, monitoring and investigation workflows. Agents perform assigned tasks within those processes.
The policy engine extracts rules from internal policies, standard operating procedures and working instructions. Those rules determine what an agent can do, what evidence it must collect, and which cases require human review.
“You need to have complete insight,” Skarnager says. “You need to have guardrails and frameworks and policies and so on, to stitch it all together.”
AI Agents draw on the customer record and the wider case context. The platform records each action, evidence item and approval. Compliance, operations and product teams work from the same decision record.
Skarnager draws a line between tasks that require interpretation and checks that demand a consistent, predictable method. An agent can identify the need for a sanctions check, but a deterministic tools handle screening and calculator functions.
Connecting Monitoring, Investigations and Outcomes
spektr 3.0 carries this structure across the customer lifecycle. KYC and KYB processes establish the customer profile and risk assessment. Customer monitoring detects changes in ownership, status or risk. Transaction Monitoring checks activity against the firm’s detection rules.
Case AI collects customer information, transaction data and screening results for the investigation. The workflow routes alert cases to analysts when the evidence calls for judgement or approval.
The Policy Engine also supports regulatory change management. Skarnager said the platform can compare a regulatory change with the firm’s policies and processes, identify the affected controls and propose amendments for approval.
The company aims to shorten the route from regulatory interpretation to policy, workflow and control changes. Firms remain responsible for deciding which tasks agents can perform, and which decisions require human judgement. Each case must show which policy applied, what the system did and who approved the outcome.
Subscribe to our newsletter
