Solutions · Deploying AI

Decide whether your AI deployment is ready to go live.

AIQURIS works out what your deployment needs to be safe, reliable and compliant, shows which controls are in place and which are still open, and gives you a clear go-live recommendation. In days, not months.

Not sure how deep you need to go? Your first impact screening is free.

Go-live readiness
Customer service assistant Generative AI · customer-facing · vendor model
Impact: Moderate
EXPOSURE AREAS
Customer data in chat logsHigh
Wrong or invented answers to customersHigh
Accountability between you and the model vendorHigh
Commitments the assistant must not makeModerate
REQUIRED CONTROLS64
31 in place 9 partly 24 open
RECOMMENDATION Proceed with conditions Launch to a limited customer group. Close data and vendor controls before full roll-out.
The go-live question

Everyone asks whether it is ready. Nobody can say what "ready" means.

01

The pilot works. Then it stalls.

The business wants to launch. Without agreed criteria, every review raises new questions and the decision keeps moving.

02

Each function sees one part

Legal, cybersecurity, risk and the AI team each check their own aspect. Nobody holds the full picture of what this deployment needs.

03

Or it goes live anyway

And when the board, a customer or a regulator asks why, there is no documented basis to point to.

The go-live decision needs a basis, not a guess.

How it works

From a description of your deployment to a go-live decision.

1

Describe it

Purpose, users, data, vendor and jurisdiction, in one guided intake. No access to the model or your data needed.

2

Work out what it needs

The risks across all six risk domains: safety, security, performance, legal compliance, ethics and sustainability; the requirements that apply; the controls that meet them.

3

Check what is in place

Each control marked in place, partly in place or open, with the evidence it requires and a named owner.

4

Decide

Proceed, proceed with conditions, or not yet. Your decision, on a documented basis.

Depth follows the risk: an impact screening (IMPACT+), a risk assessment with the controls you need (RISK+), or a full assessment against standards, regulations and your policies (CONTROL+).

A go-live basis in days, not months.Manual expert assessment: 3 to 9 months. With AIQURIS: hours to days.
Testing

Testing is necessary. It is not sufficient.

There is no single test or test suite for AI, and the testing standards are still being harmonised. Tests show that a system meets basic requirements and passes its benchmarks. They cannot show how it behaves across everything it may face: the possible inputs and outputs are too vast, and any test sample is too thin.

AIQURIS works out which tests each risk calls for and sets the conditions of acceptability with your accountable team before testing starts. It also specifies the controls that cover what tests cannot.

A passed test is evidence, not assurance.

1
Select the testsAIQURIS
AIQURIS derives the tests each risk calls for, and identifies the risks that need none.
2
Set the conditions of acceptabilityAIQURIS, with your team
AIQURIS proposes them. Your accountable team agrees them before testing starts.
3
Run the testsYour team or a test lab
Your AI team, a test lab or a specialist tool runs them.
4
Cover what tests cannotAIQURIS
AIQURIS specifies the controls for the gaps: human oversight, operating limits, and monitoring once live.
5
DecideYou
You decide, on the test results and the status of every other control.
Three deployments

Same method. Three different answers.

READY TO GO LIVE Baseline controls in place. Go live.
Internal knowledge assistant Staff ask questions of internal documents.
Impact: Minor Impact screening

Low impact, so the assessment stops early. Reassess if the scope widens.

PROCEED WITH CONDITIONS Limited launch now. Full roll-out once the gaps are closed.
Customer service assistant Customer-facing generative AI on a vendor model.
Impact: Moderate Risk assessment

Data and vendor controls are conditions for the full roll-out, each with a named owner.

NOT YET Not ready until the critical controls are in place.
Claims assessment assistant Recommends whether to pay insurance claims.
Impact: Major Full assessment

Fairness testing and genuine human review must be in place and evidenced first.

AIQURIS assessments in practice: pharma, healthcare, public transport, hospitality, maritime, telco.

The answer depends on the deployment.

What you get

Everything the go-live decision needs, in one record.

FOR SMEs No AI governance team needed. Start with one deployment.
FOR ENTERPRISES One method across business units, every decision in one AI register.
  • ✓Impact tier, affected stakeholders and potential harms
  • ✓Risk profile across safety, security, performance, legal compliance, ethics and sustainability
  • ✓Applicable requirements, each linked to its source
  • ✓Controls as a checklist: in place, partly in place, open
  • ✓The tests each risk calls for, with conditions of acceptability agreed before testing
  • ✓What to monitor once the deployment is live
  • ✓A recommendation: proceed, proceed with conditions, or not yet
Start free

Your first impact screening (IMPACT+) is free.

Describe one AI deployment. See who could be affected, how badly, and how deep the assessment needs to go next. No sales call needed.

  • ✓Impact tier for your deployment
  • ✓Affected stakeholders and potential harms
  • ✓Where to go next, and how deep
  • ✓The deployment registered in your AI register
Start your free screening
Get started

Bring one deployment. Leave with a go-live decision.

The AI Deployment Assessment Sprint takes one real deployment to a decision within three weeks, with five hours of your team's time.

Briefing1 hour

Choose a deployment. AIQURIS prepares the risk profile, requirements and controls.

Workshop3 hours

The accountable stakeholders validate the risks, challenge the controls and agree what closes the gaps.

Readout1 hour

An assessment record, a recommendation and an action list with named owners.

Privacy Overview

We use cookies to operate our website, ensure its proper functioning, improve performance, analyse traffic, and support our marketing activities. Some cookies are strictly necessary and cannot be disabled. Others can be enabled or disabled below according to your preferences. For full details, please see our Privacy Policy.