Quality policy

Venturalítica does assurance and regulatory governance of artificial intelligence systems: a product that evaluates AI systems and projects their conformity, and the professional services that accompany it (scope, clause 4.3).

That makes quality not a support function here. It is what is sold. A client buys from Venturalítica the confidence of being able to demonstrate before a regulator, an auditor or a client of their own that their AI system complies. An organisation that sells that and cannot demonstrate it of itself does not have a compliance problem: it has a product problem.

Hence the strategic direction this policy sustains, declared in clause 4.1: to sell to whoever has to SHOW conformity —the developer and the deployer of AI systems, obliged to demonstrate that their AI complies— and, in order to get there, to be the assurance infrastructure of the AI supply chain. This policy supports it because it makes quality the product and not an annex: whoever sells the proof has to be able to produce it of itself.

And this management system holds a specific place in that direction, which is worth stating without dressing it up: it is the first requirement and the first counterexample of the line for organisations. It was built from the company's own need and, in using it, it falsified the hypothesis that what serves to assure a SYSTEM serves to govern an ORGANISATION. It is not claimed that the company governs itself with the product it sells, and a counterexample of one's own backs a method better than a success story of one's own.

Hence too the value that orders decisions: proof, not promises — what is claimed is backed by evidence verifiable by a third party, or it is not claimed.

Venturalítica commits to:

1. Meeting applicable requirements — those of its clients, the legal and regulatory ones, and those it imposes on itself. Where a regulatory requirement affects the client and not Venturalítica, it is attended to just the same: that is the product's ground.
2. Continually improving the suitability, adequacy and effectiveness of the management system, with the cycle of findings and corrective actions as its engine and the management review as its channel.
3. Not over-claiming. No public statement about the product, about compliance or about the state of the management system goes beyond what the evidence sustains. "Aligned with" is not "certified".
4. Sustaining the conformity of what is delivered: nothing is released without its verification in green, and every AI-assisted output with relevant effect passes through human decision before taking effect.
5. Listening to the client: satisfaction is measured, not assumed, and complaints enter through the same channel as any nonconformity.

This policy is the framework from which the quality objectives derive: each objective must be traceable to one of the five commitments above, be measurable, have an owner, a baseline and a cadence, and be reviewed at the management review.

The objectives in force —with their metric, target, owner, baseline and linked risks— are not transcribed here: they are single-source data in objectives.yaml and are consulted in the objectives register. Copying them into this policy would only create a second version that would age worse than the first.

1. Customer focus. Understanding and satisfying the client's needs is the priority.
2. Process approach. Results are achieved by managing activities as interrelated processes, not as loose tasks.
3. Evidence-based quality. Decisions rest on verifiable data, and versioned change is the act of compliance.
4. Continual improvement. Quality is a permanent objective, not a state reached.
5. Compliance. Applicable legal, regulatory and contractual requirements are respected.

The policy is communicated to every person working under the organisation's control upon their onboarding, with record kept in their person file, and is made available to the relevant interested parties, through the channels the system keeps in force at any given time and with the record that §7.5.3 requires.

At least annually and at every management review; also upon significant changes of context, of scope or of strategic direction.