GATEKEEPER AI AGENTS
Vendor Code of Conduct Review Agent.
A vendor’s code of conduct sets out how it does business, but the topics that matter are easy to skim past.
The Vendor Code of Conduct Review Agent checks every code against 12 topics, so your team can see what's missing.
See it in action.
HOW IT WORKS
How the Vendor Code of Conduct Review Agent works.
Code submitted.
A vendor attaches its code of conduct or ethics policy to the card in Gatekeeper’s workflow.
Topics identified.
The agent reads the code against 12 responsible business topics, counting a topic as addressed under any wording, so “kickbacks” and “facilitation payments” both count as anti-bribery.
Coverage rated.
Each topic is rated Covered, Partly covered or Missing. Covered means a clear commitment and how it is applied through a rule, process or responsibility.
Governance checked.
The agent notes whom the code applies to, whether it has an owner, date or version, and whether breaches have consequences.
Structured output.
Findings are summarised with a verdict, the number of topics covered and a coverage table with supporting evidence.
Follow-up focused.
Your team reviews the gaps that matter most and raises them with the vendor.
FAQ
Frequently asked questions.
What does the Vendor Code of Conduct Review Agent review?
The Vendor Code of Conduct Review Agent reviews a vendor’s code of conduct or ethics policy against 12 topics expected of a responsible vendor:
-
Anti-bribery and corruption.
-
Anti-money laundering, fraud and sanctions.
-
Human rights, modern slavery and child labour.
-
Health and safety.
-
Diversity, inclusion, equal opportunity and anti-harassment.
-
Environment.
-
Data protection and confidentiality.
-
Conflicts of interest.
-
Gifts and hospitality.
-
Whistleblowing and non-retaliation.
-
Fair competition.
-
Expectations of the vendor’s own suppliers.
What is the difference between Covered and Partly covered?
The Vendor Code of Conduct Review Agent rates a topic Covered when the code makes a clear commitment and explains how it is applied, through a rule, process or responsibility.
Partly covered means the topic is mentioned in passing or stated as a value with no rule or process. Missing means the code does not refer to the topic anywhere.
When does a code need attention?
The Vendor Code of Conduct Review Agent returns Needs attention when anti-bribery and corruption, human rights and modern slavery, or whistleblowing and non-retaliation are missing, or when three or more topics are missing.
Every other code returns Looks fine, with the coverage table showing any partial gaps.
Does the wording have to match?
No. The Vendor Code of Conduct Review Agent counts a topic as addressed under any wording. A code that refers to corruption, facilitation payments or kickbacks is treated as addressing anti-bribery, so vendors are not marked down for using their own terms.
Does it check what the code expects of the vendor’s own suppliers?
Yes. Expectations of the vendor’s own suppliers is one of the 12 topics the Vendor Code of Conduct Reviewer Agent checks. The agent also notes who the code applies to, where the code states it, and whether breaches carry consequences.
Does it replace the compliance team’s review?
No. The Vendor Code of Conduct Review Agent handles the first-pass review and structures the findings. Your compliance team reviews the output and decides whether the vendor’s standards meet yours as part of your vendor risk and compliance process.
What are the configuration options?
Tailor the Agent to your organisation’s resilience requirements rather than a generic checklist, including:
-
Topics required.
-
Priority topicsRating definitions.
-
Verdict rules.
-
Review outputs.
How does Agent automation compare with a manual process?
Without Agent automation, teams read each code of conduct and check it topic by topic against their own expectations.
-
Codes vary in structure and wording, which slows comparison.
-
Partial commitments are hard to tell apart from real ones.
-
Missing topics such as whistleblowing can be overlooked.
With Agent automation, codes of conduct are checked automatically as soon as they reach the card.
-
Every code is rated against the same 12 topics.
-
Partial and missing topics are flagged with evidence.
-
Teams follow up only where coverage falls short.
With Gatekeeper’s Agentic AI, routine tasks are completed up to 95% faster.
Is there an audit trail?
Every Agent action is recorded in Gatekeeper with a timestamp, including the criteria applied, decision made, supporting evidence and reasoning where applicable.
This creates a searchable record of how each outcome was reached, supporting compliance verification, internal review and audit requirements.
How is my data kept secure and compliant?
Your data stays within Gatekeeper while the Agent performs its work. Actions, decisions and supporting evidence are logged, creating a clear audit trail.
The Agent operates within the permissions, playbooks and authority controls you configure.
Data handling follows Gatekeeper’s enterprise security standards, including SOC 2 Type II compliance and GDPR requirements.
one unified record
Every capability, unified.
One system of truth.
Customer and vendor contracts, obligations, and renewals in one record.
See spend against contract.
Contract value sits next to what is actually being spent.
No single point of failure.
Contract knowledge lives in the platform, so nothing is lost when someone changes role or leaves.
Draft from approved templates.
Every agreement starts from language Legal already trusts.
Keep control of the clause library.
Guard-railed templates keep off-policy wording from going out.
Nothing renews unnoticed.
Every renewal date tracked, with owners and reminders set automatically.
Expert insight & opinion
Related AI agents.
Gatekeeper's agents work as a single, connected team, handing off to one another across the lifecycle rather than running in isolation like bolted-on point tools.
