Agents in production. Auditors asking questions.
Half of deployed agents run in silos. Tenders and insurers want ISO/IEC 42001 and SOC 2 evidence.
Suite pricing · audit £19,500, credited · AWS Marketplace private offers available
of boards have limited or no AI knowledge (Deloitte, 2025)
of UK organisations lack AI access controls (IBM, 2025)
of deployed agents run in silos (Salesforce/MuleSoft, 2026, vendor survey)
First findings within 48 hours of your health check.
Issues detected · needs attention first
| Issue | Rule | Severity | Owner | Status |
|---|---|---|---|---|
| Payroll file uploaded to an external AI tool | R-15 | High | HR Director | Blocked · reviewed |
| 31 personal AI accounts found in Sales | R-02 | High | Head of Sales | Migrating · 27 done |
| Customer data in a prompt without redaction | R-01 | High | Ops lead | Redacted · 0 left |
| Premium model used for internal drafts · 412 requests | R-07 | Medium | Engineering lead | Rerouted |
| CRM "AI assist" running without a DPIA | Policy | Medium | Head of Sales | DPIA in progress |
| Agent attempted an external email | R-12 | Low | S. Reid | Held · approved |
| Job | Owner | RAG | Effort | Cost to date | Forecast | Actual | Variance | Return / yr | Status |
|---|---|---|---|---|---|---|---|---|---|
| Migrate personal accounts to managed | IT Director | 12 d | £4,200 | £6,000 · 20 d | £4,200 · 12 d | −£1,800 · −8 d | risk | Done | |
| Route everyday tasks to standard tier | FD | 3 d | £900 | £1,200 · 4 d | £900 · 3 d | −£300 · −1 d | £58,400 | Done | |
| Payroll upload block (R-15) | HR Director | 1 d | £300 | £300 · 1 d | £300 · 1 d | 0 | risk | Done | |
| Collections letters via Comms agent | Head of Collections | 18 d | £11,600 | £14,000 · 25 d | Pending | on track | £96,000 | In progress | |
| KYC document checks | Head of Onboarding | 9 d | £5,100 | £9,000 · 15 d | Pending | on track | £61,000 | In progress | |
| Contact-centre adviser assist | Head of CS | 26 d | £19,800 | £16,000 · 20 d | Pending | +£3,800 · +6 d | £142,000 | At risk · vendor delay |
Effort, cost and return are recorded against each job as it happens. Done means the job is solved in production with its owner’s sign-off, not that a licence was bought.
We test it before you use it.
New models arrive most weeks from a handful of companies. Deciding which to use, for what, and when not to, is not a decision your employees should be making. So we make it first, in a controlled sandbox, with a dedicated research team.
Ask us to test a specific frontier model against your workload. You get a written result and, if it passes, a staged switch with a way back.
| Release | Behaviour | Stability | Cost vs current | Data | Verdict |
|---|---|---|---|---|---|
| Standard-tier model, 9 Sep | Pass | Pass · 0.4% drift | −11% | Pass | Approved |
| Premium model, 2 Sep | Pass | Fail · 6% drift | +38% | Pass | Declined |
| Small model, 11 Sep | Testing | Day 3 of 7 | −54% | Pass | In progress |
| Client request · voice model | Testing | Day 1 of 7 | Pending | Review | In progress |
Example data. Results are written up monthly in the Change Control briefing.
From businesses your size.
What is an AI governance layer for an enterprise?
Use a governance layer: a register, enforced rules, a log and Evidence Packs control-mapped to the standards your buyers ask for. Guard covers public vendors and private models in your environment. Usage Insight serves as the governance function's console.
We already have a governance function. What does GenWA add?
Usage Insight becomes their console. Evidence Packs and Compliance Mapping are generated from the log, so the function spends its time on judgement, not assembly. We support your function; we do not replace it. They keep the decisions, and we supply the data.
Can GenWA sit alongside our model platform and procurement?
Yes. GenWA is vendor-neutral and covers private models in your environment as well as public vendors. GenWA is an AWS Marketplace seller, so an existing agreement can be used through private offers. Contract and procurement terms are agreed with your team, and we confirm all of this in writing before you commit to a platform plan.
What does the enterprise AI audit cost?
The indicative audit fee for enterprise is 19,500 pounds, excluding VAT, confirmed in writing before you commit and credited in full against year one. Platform pricing and discounts for enterprise are to be confirmed. Ask for the written quote and credit terms. The audit also gives your board a baseline for the platform decision, so you can compare it with other options on equal terms.
What evidence can we give to tenders, insurers and auditors?
Evidence Packs control-mapped to ISO/IEC 42001, UK GDPR, SOC 2 and Cyber Essentials, drawn from the log. Buyers decide what is sufficient. We prepare evidence and make no certification claim. Packs are produced from the log, so they reflect what actually happened, not what was planned. Evidence Packs and Compliance Mapping are available now; other components may be in early access.
How many deployed AI agents run in silos?
Salesforce and MuleSoft's 2026 survey found 50% of deployed agents run in silos (a vendor survey). Deloitte's 2025 research found 66% of boards have limited or no AI knowledge. Both support a single standard across teams. Treat both figures as context, not a diagnosis of your estate.
How do we start in a large organisation?
Begin with an AI health check scoped to one or two units. First findings arrive within 48 hours and the audit and 90-day plan within ten working days. Then agree a security review, hosting in your AWS or Azure account (or GenWA AWS as agreed) and a phased rollout.
Find out where your AI stands. Then decide.
Under a minute, an upload, or a visit. Something for the board either way.