# Collty Intelligence Core

Canonical HTML: [https://collty.com/intelligence-core](https://collty.com/intelligence-core)
Site: Collty

Collty Intelligence Core is the private orchestration and learning layer connecting permission-aware evidence, domain algorithms, replaceable AI models, evaluation, provenance and governed product decisions.

## Architecture

- A permission-aware work graph connects relevant professional, team, project, task, collaboration and outcome evidence.
- Domain algorithms structure retrieval, ranking, coverage, availability, compatibility, workload, delivery and risk before model reasoning.
- External models are replaceable reasoning components; the Collty evidence model, ranking logic and governance remain inside Collty.

## Controlled learning

- Versioned private evaluation datasets and task-specific graders measure team assembly, planning and recommendation quality.
- Human feedback and eligible project outcomes can become controlled learning signals.
- Candidate calibration is checked on holdout data, observed in shadow mode and activated only after human approval.

## Privacy and accountability

- Provider context is minimized, structured and filtered; secrets, payment details and unrelated workspace data are excluded.
- Customer workspace content is not sent to model providers to train shared foundation models.
- Provenance connects high-value runs to task, evidence, algorithm, grader and provider versions without publishing proprietary implementation details.
- Quality, latency, errors, token usage, cost and drift are monitored as one operational system.

## Standards and guidance

- [NIST AI Risk Management Framework](https://www.nist.gov/itl/ai-risk-management-framework)
- [NIST Generative AI Profile](https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-generative-artificial-intelligence)
- [ISO/IEC 42001:2023](https://www.iso.org/standard/42001) and [ISO/IEC 23894:2023](https://www.iso.org/standard/77304.html)
- [European Commission guidance on AI Act transparency](https://digital-strategy.ec.europa.eu/en/library/guidelines-transparency-obligations-providers-and-deployers-ai-systems)
- [OpenAI evaluation best practices](https://developers.openai.com/api/docs/guides/evaluation-best-practices)
- These sources inform Collty's design approach; they are not a claim of certification or endorsement.

## Default description

Collty helps experts, remote teams and companies assemble AI-matched teams, plan projects, create proposals and manage delivery.