For risk and compliance teams
The agents are running.
Sign off on evidence.
Get your risk and compliance team a record of how each AI actually runs, so a sign-off rests on evidence rather than on a form someone filled in once. Each agent you connect can ask before it acts on money, customers or anything that cannot be undone, a named person answers, the agent acts on the answer, and Deliverd keeps the question, the answer and a signed receipt on the AI register record it belongs to. Your platform team keeps running the agents; you read what they did.
No card required · Approvals on every plan · Secure by default
What you get
What Deliverd changes for risk and compliance teams.
The register shows how the AI ran, not how someone says it runs
Each record in the AI register carries the checks its agents made, what was put to a person and what they said, spend against budgets, the models named, and the latest signed receipts.
How: Governance recorded as the work happens ↓A named person answers before the agent acts
When an agent wants to move money, contact a customer or do something that cannot be undone, the question goes to the person who owns it, with what the agent wants to do and why.
How: Your AI asks before it acts ↓Rules your committee can read, and a second yes to change them
What is settled by a rule and what goes to a person is written as reviewable policy. Where you ask for it, a change to those rules waits for someone other than its author to agree.
How: A check before the agent acts ↓The situation
What this looks like today
The register is a form, and the agents moved on
An AI system is assessed once, at sign-off. By the next review the agent has a new model, new tools and new people relying on it, and nothing on the record says so.
Indicators filled in by hand
The monthly risk indicators are typed in from screenshots and recollection, so the committee reads a figure nobody can trace back to what happened.
Nobody can show who said yes
When an agent refunds a customer or changes a supplier's bank details, the question afterwards is who allowed it and what they were shown. Chat logs and tickets do not answer it.
Platform and risk
Two teams, one record
Your platform team runs the agents. Your risk team signs them off. Each thing the platform team does in Deliverd becomes evidence the risk team can read, without anyone retyping it.
The platform teamAgents check before they act, and can name the model doing the work.
The risk and compliance teamEach AI register record shows how its agents actually ran: checks, decisions, spend and signed receipts, and a model the record does not name.
The platform teamAI usage comes in from the products people use, and agents and gateways are set up in Deliverd.
The risk and compliance teamDiscovered AI lists what is in use but not yet in the register, when an administrator switches it on.
The platform teamRule changes can run beside the rules in force first, recording what they would have decided.
The risk and compliance teamWhere the organisation asks for it, a rule change waits for a second person's yes before it takes effect.
The platform teamDecisions can go to your own OpenTelemetry collector as spans in the agent's trace.
The risk and compliance teamSigned receipts travel in the evidence pack and in every register record export, checkable offline.
The platform teamEach gate call chooses what happens if Deliverd cannot be reached: refuse, allow or raise.
The risk and compliance teamA monthly committee pack: the risk indicators, register changes, incidents, decisions and AI spend, with an accountable executive's attestation on each record.
How it works
Deliverd for risk and compliance teams
Everything below ships today. Where something needs a paid plan, it says so.
Your AI asks before it acts
An agent that needs a decision puts it to a named person and waits: what it wants to do, why, how risky it is, and what to read first. They approve, approve with changes, reject with a reason, or ask the agent a question — on the page, from an email link, in Slack or Teams, or from a push notification on a phone — and a colleague can cover their approvals while they are away. It can also ask for a review, or for information it is missing, the same way.
Governance recorded as the work happens
Compliance Mode adds a register of the AI systems you use and, on governed work, the owners and the risk and data classification. That profile decides the controls: evidence or a person's approval meets each one, and an exception, with an owner and an expiry, shows as waived rather than met. The whole record exports as one evidence pack. It records how the work was governed — it does not certify compliance with any law or standard.
Included in Business and Enterprise
One pack for the auditor
Every step of a piece of work — who was asked, what they said, what was decided and when — exported as a single timeline in JSON, CSV or PDF. It carries a digest and the head of the hash-chained audit trail, signed with a published key, so anyone can check a copy offline, and the export itself is on the record.
A check before the agent acts
Before a consequential action — a payment, a deployment, a message to a customer — the agent declares what it is about to do and waits for the answer. A named person decides, or your organisation's rules settle the routine cases and send the rest to a person; with no rules, every one goes to a person. Afterwards the agent records what ran, and the record is checked against what was allowed.
On every plan; policy rules included in Business Starter, Business and Enterprise
A record of who opened it
Views, versions, access changes and share links are written to an audit trail you can export — which is the difference between believing a client saw something and being able to show it.
When deletion has to stop, it stops
A matter opens and routine destruction becomes the wrong thing to do. Place a hold — on the whole organisation, one project, one task, one person’s material or one report, over a date range if you need one — and the retention schedule suspends, the organisation cannot be closed, and what the hold covers cannot be deleted: a report’s data, a task, a retracted comment’s attachments. Refused in the database, not by a setting somebody has to remember. Released with a reason, and both ends are on the record.
Included in Enterprise
Questions
Risk and compliance teams and Deliverd.
Does this replace our GRC platform?
No. A GRC platform holds your obligations, policies and controls across the business. Deliverd is where AI agents ask before they act and where the answer is recorded, and its register, indicators and evidence packs export to whatever your GRC platform reads.
Will a gate catch an agent that never asks?
A gate is advisory when the agent calls it directly: the agent reports what it intends to do and performs the action itself, so it catches only an agent that asks. Put an MCP server or an HTTP API behind a Deliverd gateway and keep its credential there, and a call that changes something does not reach it without a yes. Discovered AI lists AI in use that is not in the register, so an agent that never asks is less likely to go unseen.
Who answers for each AI system?
Each register record can name one accountable executive, with a regime label if your regulator uses one (for example SMF24 under the UK Senior Managers and Certification Regime), and that person attests in their own words. The attestation is dated, belongs to the person who gave it, and travels with the record's exports.
What does the committee get each month?
A pack in your own branding: each risk indicator against the limits your appetite sets, with its trend, owner and action, then the register changes, incidents, what people decided and how fast, and what AI cost. Indicators Deliverd measures are read from what happened; for those an owner reports, Deliverd shows what it saw beside the figure.
Does it make us compliant with the EU AI Act or DORA?
No product can. Deliverd records how AI work was governed, as it happened, and maps that record to the frameworks you use. Whether that meets an obligation is your judgement and your regulator's.
Sign it off on what happened.
Every agent asks, a named person decides, and the record of who allowed what is on the AI register, in the evidence pack and in the committee's monthly pack.