Which model providers does Vaival use inside engagements?
Enterprise-tier AI providers with training-on-customer-data disabled. Anthropic and OpenAI are primary. Sub-processor list is disclosed in the engagement letter at scope confirmation, with right of refusal on changes. We do not maintain a public list because the right disclosure surface is the engagement letter, not the marketing page.
Are models trained on our data?
No. Vaival uses enterprise-tier provider configurations with training-on-customer-data disabled. Where a workflow benefits from custom evaluation, the evaluation runs inside Vaival's controlled environment, not on the provider's training pipeline. The sub-processor list and the contract clauses around training are in the engagement letter.
How does model revocation work?
Revocation is by written notice to the named Vaival owner. Within three business days, model access is disabled, processed-data artifacts (embeddings, fine-tuned weights, cached outputs) are deleted, deletion is certified to you in writing, and the model provider sub-processor is notified. The engagement continues; the workflow runs human-only until you and the named owner agree on a replacement approach. Trust carries the full procedure.
Is CostSage available as a standalone product?
No. CostSage is the AI product in the scope of the ISO/IEC 42001:2023 certificate. AI Labs builds it; Vaival's own teams use it every day. The services wing uses CostSage inside select client engagements where the workflow shape fits. It is not sold as a standalone SaaS product, and AI Labs is not commercialized separately either.
How is Operational Intelligence delivered, and what does it cost?
Operational Intelligence is productized inside AI Labs and delivered through the Operational Intelligence Pod, a three-month variant of Managed Operating Pod for single-business multi-unit operators. The engagement carries the Vaival Unit Scorecard, six operating dimensions, gates, evidence grades, and decision support across scale, stabilize, protect, restructure, incubate, and pause. Same governance, same engagement letter, and same nine-element AI Go-Live Gate as the standard Managed Operating Pod. Engagements start from $75,000 per quarter. The methodology is not licensed as software.
What is AI 360, and when can we buy it?
AI 360 is a portfolio management system AI Labs is composing for investors and operators of multi-business portfolios. It coordinates multiple layers and frameworks to produce a true picture of each business in the portfolio and recommend direction across grow, invest, risk, reposition, and expand. Operating Intelligence is one of the frameworks inside AI 360. AI 360 is not generally available today. Architecture is available under NDA in a real procurement conversation. Engagements start from $150,000 per year, delivered through the Portfolio Operations Pod. If you operate a single business in the $15M to $100M revenue band, the four direct offers are the right entry point; AI 360 is not for you yet. If you operate or invest across a multi-business portfolio and want to be among the early-access conversations, mention AI 360 in your first call.