Services
What ISWT offers
Each engagement is scoped in writing before work begins.
The pilot: checking an agent's completion reports
For teams whose pipelines, dashboards or people act on an AI agent's status. The pilot checks what your agent says it did against records it cannot change.
What it measures
- How often your agent says "done" when the final check failed.
- How often it says "done" when the final check never ran.
- Whether the line it cites really shows what it claims.
- Which fix suits your setup, and what each costs on work that really passed.
What it does not do
It measures how your agent reports on evidence it is given. It does not test what your agent does on live systems, and it does not certify your agent as safe.
How it works
- A short call: which agent, which status words your team acts on, and where the truth is recorded.
- A test built from about 16 tasks your agent really does.
- Expectations are written down before your agent sees the test, so they cannot be adjusted afterwards.
- A report you can recheck: the counts, the cases behind them, and the change I would make to start.
Price and time
A fixed price and a timeline, agreed in writing before any work starts. About two hours of your team's time. One pilot at a time.
Review of AI-assisted work
A second look at work an AI helped produce: code, analyses, reports. What the evidence shows, what it does not, and what to check next.
Building with AI under written rules
Websites, tools and automations, made with AI under rules that I write down and you can read.
AI explained in plain words
For teams and people who do not work in AI: how it works, why it gets things wrong, and how to catch it.
Ground rules for any work
- Your work is yours. I publish nothing about your systems without your written yes.
- We can sign a mutual confidentiality agreement before you share anything.
- AI models are tools in my work. Your material goes only to the tools you approve. If it must not reach a model provider, I keep it to local models on my own machine.
- I say what I cannot do, and I report what the record shows, including where my own checks fall short.
- I develop these methods and tools, and I may offer paid services built on them. Evaluations of my own methods are developer-led, not independent validation. I hold no investments in AI companies. Outside review and replication are welcome.
- Paid work happens only under a separate written agreement. See Terms.