AI-built software. Human accountability.
Generated by AI.
Reviewed by people.
Built with Lovable, Cursor, Claude Code or another AI agent? We assess what you have, fix what fails and explain the risks before your next decision.
First conversation: 30 minutes, free.
01 / Services
Where do you need clarity?
Five distinct services. We start with the situation your application or company is in.
AI Software Audit
You have an AI-built application and need a technical assessment before investing further or launching.
See what's includedAI Software Rescue
Errors, vulnerabilities or scaling problems. We work on your existing application and prepare a version for production.
See what's includedAI Agent Security Assessment
We assess AI agents that read data, send messages or change systems, focusing on permissions and action controls.
See what's includedContinuous Software Assurance
After audit or remediation, we periodically review significant changes, application security and dependencies.
See what's includedSoftware Risk Assessment
A documented technical assessment for companies buying software and insurers evaluating a software supplier.
See what's included02 / Evidence, then decisions
A diagnosis you can act on.
Clear conclusions, technical evidence and next steps. A report you can discuss with management, developers or a business partner.
- 01
What we found
Findings tied to the components and scenarios reviewed, with evidence and assessment limitations.
- 02
What matters first
Priorities based on impact, exposure and how you use the application.
- 03
What to do next
Practical recommendations, remediation steps and checks for the changes made.
Illustrative finding
Data access is too broad
- Scenario reviewed
- A user can request another organisation's data through an API.
- Recommended next step
- Server-side permission checks and tests for isolation between organisations.
Presentation example. This is not a client assessment result.
03 / How we work
From a first conversation to a practical plan.
- 01
Define the assessment
We clarify the application, objective and available access. Scope, deliverables, timeline and cost are agreed upfront.
- 02
Work from evidence
We examine code and configuration, then test agreed scenarios. Testing stays within authorised environments and boundaries.
- 03
Explain the next step
We present findings and priorities. Remediation and recurring reviews are agreed separately when needed.
UP2DATE / AI TRUST
AI speed. Engineering judgement.
AI can accelerate building an application. Its quality shows in the code, how it protects data and how it behaves in real situations. That is where we come in.
- Agreed scope of work
- Evidence-backed conclusions
- Clear risks and limitations
04 / FAQ
Before we start
Does the entire application need to be AI-built?
No. We can assess applications with some AI-generated code, as well as conventionally developed systems. The scope starts with your risks and the decision you need to make.
How much does it cost and how long does it take?
The proposal depends on application size, available access, integrations and assessment depth. We agree on scope, timing and cost before starting. Continuous Software Assurance is a recurring engagement with defined activities and capacity.
Do you need production access?
Not by default. We agree on the minimum access needed, test environments and working rules. Access to code, configuration or sensitive data is arranged after scope and confidentiality are agreed.
Is the report a certification or a guarantee?
No. It documents findings within the scope and period of the assessment. It does not certify the absence of all vulnerabilities or guarantee acceptance by an insurer.
05 / Discuss your project
Let's look at what you've built.
Tell us about the application, how it was built and what concerns you. We will work out which service fits.
First conversation: 30 minutes, free.
There is no need to send source code, passwords or customer data with your first enquiry.