Useful knowledge, carefully handled.
A clear handover explains what was completed, the evidence and checks behind it, its limits and any next actions retained by you.
A short, non-confidential first message is enough. We agree handling before you send sensitive documents.
Work you can inspect.
Important facts are traced to supplied material or the actual external sources used. Calculations and assumptions are distinguishable. Code is supported by documented tests and limits. Proposals are explained against useful criteria.
We make clear what each source or check establishes and what remains uncertain. Baseline completion checks are included even if you choose not to inspect the output. Extra scrutiny addresses a defined additional concern.
Information protection and confidence in a conclusion are separate questions. The result should show its reasoning, evidence, objections, limits and what could change it.
Inspect an illustrative delivery or the software demonstration.
When experts contribute, the delivery includes an agreed expert-contributions appendix or companion document showing each contribution's scope, substantive input and limits. Entries use assignment-specific labels and broad capability descriptions to reduce identification risk. AI-only services do not include specialist input. See how an expert contribution is represented.
What is shared, and what stays separate.
We agree handling, including any AI processing, before you send sensitive documents. IQ256 keeps original confidential material separate and prepares a purpose-specific packet: the part of the brief needed for a particular contribution. It preserves relevant context while withholding unnecessary identifying or sensitive details. These paired details are illustrative examples.
Exact figures, identifying details and confidential source material.
Last year's revenue: €1,827,164
Employee count: 112
A private staff discussion unrelated to the decision.
Original confidential material is not placed in purpose-specific packets for AI models or human contributors.
Broader ranges and only the context needed for a focused contribution.
Last year's revenue: €1.5 million to €2 million
Employee count: 100 to 140
Omitted. It is not needed for the review.
This material is shared only after you approve the full packet, its purpose and the models, providers or human contributors that may receive it.
A combination of details can still identify your company, even when names are removed or language is translated. You check the full packet before external sharing.
IQ256 brings the contributions together into your confidential final delivery.
Approve external sharing.
- You name a privacy approver. It can be you.
- Before external work begins, we send each purpose-specific packet for approval: what may be shared, with whom and for what purpose.
- Only the approved purpose-specific material is shared with AI models, human contributors or external tools, including during web research. New people, models, providers or materials require approval before sharing.
- You can correct or narrow the packet, or stop the work before external review begins.
The same information-handling baseline applies to every assignment, free or paid.
IQ256 selects, briefs and coordinates suitable contributors. Your approval here concerns information sharing. You do not recruit or direct contributors, or approve each internal step within the agreed boundaries.