Skip to main content

AI how-to guides Risk, compliance and governance

How to run third-party due diligence with AI

Short answer

Connect an AI system to screen and monitor your third parties continuously against sanctions, PEP registers, cyber posture and ESG signals. The AI maps the ownership graph and triggers alerts when risks shift. People review the flagged risks, decide on the outcomes and keep a full audit trail.

Updated: · Thinkia

What AI does

  • Maps the beneficial ownership graph to uncover indirect exposure through subsidiaries or parent companies.
  • Screens third parties continuously against sanctions lists, PEP registers and adverse media.
  • Tracks external cyber posture and ESG signals, surfacing them as tracked attributes.
  • Triggers alerts and routes them to named owners with proposed actions when a risk shifts.
  • Logs every screening, decision and override to maintain a continuous, audit-ready evidence vault.

What stays with people

  • Reviewing the alerts the AI triggers and deciding whether to proceed with, suspend or terminate the relationship.
  • Defining the risk taxonomy, the screening sources and the rules for escalation.
  • Owning the final procurement and risk decisions for high-tier suppliers.

How to put it in place

  1. Map your third-party portfolio and risk taxonomy

    List all your current suppliers, partners and the risks you need to monitor, such as sanctions, cyber or ESG. Define the risk tiers and the decisions the system must support.

  2. Integrate screening sources and build the ownership graph

    Connect the AI to official registers, sanctions lists and media databases. Let the system map the corporate structure for each third party to make indirect exposure visible.

  3. Set up continuous monitoring and alert workflows

    Configure the system to refresh data constantly instead of at contract renewal. Define who receives alerts when risks shift, along with service level agreements and escalation paths.

  4. Test the review gates with a pilot group

    Run the system on a subset of suppliers. Have risk and procurement teams review the alerts, attach commentary and check that the evidence trail is complete before trusting the automated mapping.

  5. Scale to the full portfolio and expand signals

    Once the workflow is stable and audit sign-off is secured, roll it out to all third parties. Add new signal sources over time as regulatory requirements evolve.

What you need

  • Access to official APIs for sanctions lists, PEP registers and corporate registries.
  • Clear definitions of your risk tiers and escalation policies.
  • A designated team across procurement and risk to handle alerts.
  • Data privacy controls to manage the information collected on individuals, such as beneficial owners.

When it is not worth it

  • When you only have a handful of local suppliers that you know personally.
  • If your company lacks clear rules on what constitutes an unacceptable risk.
  • When third-party contracts are one-off purchases with no ongoing relationship.

Risks and how to control them

  • False positives from matching common names against sanctions lists: require human review for flagged entities and tune the matching logic.
  • Relying on outdated or unverified external databases: ensure the AI pulls from official, primary sources and timestamps every check.
  • Automated decisions leading to unfair contract termination: restrict the AI to triggering alerts, leaving the final decision to human reviewers.
  • Evaluating directors and owners: processing personal data and profiling may touch privacy laws and high-risk areas under regulations like the AI Act. Consult legal counsel, apply strict data minimisation, and ensure human oversight.

How Thinkia does it

We transition third-party diligence from one-off checks to a continuous flow. We map current processes, define risk tiers, and integrate screening providers to build ownership graphs and monitor cyber and ESG signals. Alerts are routed to named owners rather than shared inboxes, ensuring that procurement and risk teams review them where decisions are made.

Every screening, decision, and override is stored with its source, date, and reviewer on Thinkia Sentinel. This provides an exportable, audit-ready evidence vault for internal teams and supervisors from day one.

The Thinkia solution behind this guide

Thinkia products involved

Frequently asked questions

Can AI make the final decision to block a supplier?

No. The standard practice is for the AI to screen, map the corporate graph and trigger an alert if a risk shifts. A person from the risk or procurement team reviews the evidence and makes the final decision.

How is this different from a traditional background check?

Traditional checks are done once at contract signing and gather dust until renewal. AI enables continuous monitoring, refreshing data constantly so you know immediately if a supplier's status changes.

Does the system check for indirect exposure?

Yes. By mapping the beneficial ownership and corporate structure, the system highlights risks hidden in parent companies or subsidiaries that direct screening might miss.

How do we handle false positives in adverse media or sanctions?

The system provides explainable hits with the exact source and context. A human reviewer checks these hits in their workflow and can dismiss false matches, leaving a clear record of the decision.

What happens during an audit?

The system acts as an evidence vault. Every check, human review, and override is logged with timestamps and sources, allowing you to export a complete lineage of your diligence process.

Related guides

Related decisions

Key terms

Thinkia articles

Sources

Want to do this in your organisation? Let's talk.

Talk to an AI Expert