Compuvi
Sustainability & SDG alignment

Where compliance infrastructure meets the Global Goals.

How Compuvi's work on compliance infrastructure maps to the UN Sustainable Development Goals — and what we are committing to next.

Talk to usSee our compliance posture
Why this page exists

Governance is not a side effect of what we build. It is what we build.

Compuvi builds Confinaid, a preventive compliance layer for regulated enterprises. The Global Goals ask institutions to be accountable, transparent, and evidenced — which is the same thing an examiner asks of our customers, and the same problem our architecture was built to solve.

That overlap is real, and it is narrow. It sits almost entirely inside Goal 16. Where it thins out, we say so on the same line rather than on a different page.

Every claim below is split in two: what Confinaid does today, and what we are working towards. Nothing in the second column is a current feature.

The four goals we report against

Four goals, stated honestly.

Peace, Justice & Strong Institutions

Targets 16.5, 16.6

Target 16.6 asks for effective, accountable, and transparent institutions. Target 16.5 asks for a substantial reduction in corruption and bribery in all their forms.

This is the goal our architecture was designed against — not as an ESG narrative, but as the functional requirement our customers' regulators already impose.

What Confinaid does today
  • Confinaid supports transparent governance by making every policy decision, review, and escalation timestamped and cryptographically logged, and not editable after the fact.
  • Confinaid's policy library lets a compliance team encode its own rules — including anti-bribery and anti-corruption policies — and applies them inline to outbound communications, documents, and workflows before content leaves the organization.
  • Granular role-based access control keeps reviewers, admins, auditors, and investigators strictly scoped, so the record of who decided what holds up under examination.
  • Policy answers from our AI assistant are grounded in the organization's own documentation and traceable to a source, to minimize hallucination risk.
What we are working towards
  • Publishing how we evaluate the quality of a policy decision, so a control's effectiveness can be assessed by someone outside this company.
  • Extending the same evidence standard to AI-generated content as the frameworks that govern it mature.
Read the goal on un.org

Decent Work & Economic Growth

Targets 8.8

Target 8.8 asks for the protection of labor rights and safe working environments for all workers.

Confinaid is a policy engine, not a labor-law service. We are precise about the difference between what a compliance team can configure in it today and what we have not built.

What Confinaid does today
  • Workplace-conduct and labor policies can be encoded in Confinaid's policy library and enforced inline, alongside any other rule set.
  • Context-aware compliance training is triggered by employee risk events, so a conduct breach becomes a teaching moment before it becomes a disciplinary one.
  • Monitoring software can itself harm the people it watches. Review decisions are made exclusively by the customer's own legal and compliance staff, Confinaid staff have zero access to message content, and retention stays under the customer's control.
What we are working towards
  • Automated monitoring of labor-law change, and screening beyond the organization's own boundary, are roadmap items — not current features. We will describe them as features when they are features.
  • Internally: the hiring and workplace practices set out on our Careers page, held to the standard we ask of the organizations that deploy us.
Read the goal on un.org

Responsible Consumption & Production

Targets 12.6

Target 12.6 asks companies to adopt sustainable practices and report on them. Our commitment is to publish our own practices and progress, starting with the disclosures below.

We do not sell an environmental compliance product, and nothing here should be read as one.

What Confinaid does today
  • Our reporting module produces an organizational risk score, trend analysis, and compliance reporting for communication and document risk. It does not process environmental data.
What we are working towards
  • A first disclosure of our own practices, written the way we write our compliance posture: what is measured, what is estimated, and what is still unknown.
  • Reporting on our own footprint before we ever describe it as a strength.
Read the goal on un.org

Partnerships for the Goals

Targets 17.16

Target 17.16 asks for multi-stakeholder partnerships that share knowledge and expertise.

For a compliance company, this is where a claim can quietly contradict the product. So, plainly: we do not publish or pool customer compliance data. Our partnerships are infrastructure, research, and ecosystem relationships. Customer data stays inside the customer's environment.

What Confinaid does today
  • Compuvi participates in NVIDIA Inception, Cloudflare for Startups, and Google for Startups, and engages with academic and regulatory communities on evidence standards for AI in compliance.
  • Our privacy posture is the boundary of every partnership: data minimization, scoped access, and customer-controlled retention are architectural, not contractual afterthoughts.
What we are working towards
  • Contributing to public consultations and published research on how AI decisions in regulated workflows should be evidenced.
  • Sharing methodology, not data: our evaluation approach, written up for people who have to review it.
Read the goal on un.org
ESG

The same page, read the way an investor reads it.

Governance is where our product operates. Social is where it touches people. Environmental is where our own conduct is the only thing we can honestly report on.

Governance

Confinaid is designed to be an operational layer for corporate governance: the place where written policy becomes an enforced, evidenced control. It leaves an auditable record of every decision — so a policy breach is less likely to leave the organization unnoticed, and easier to evidence when it does.

Governance — measurement

We are defining, with pilot participants, how to measure policy-breach rates before and after deployment. We will publish the methodology before we publish any number.

Social

Internally, Compuvi commits to the hiring and workplace practices set out on our Careers page. For customers, conduct and workplace policies can be encoded in Confinaid's policy library alongside any other rule set.

Social — boundary

We treat the privacy of the people whose communications pass through the system as an obligation, not a setting. Content is not stored beyond policy requirements, and encryption keys, retention, and deletion rights stay with the customer.

Environmental

Where we are, and what we are committing to.

We do not operate our own data centers. Confinaid runs on managed cloud and edge infrastructure, and our model training uses GPU capacity provided through partner programs — so most of our environmental footprint sits with our infrastructure providers rather than with facilities we control.

We have not yet measured our own emissions, and we are not making a net-zero or carbon-neutral claim.

What we are committing to
  1. Building a first emissions estimate covering our cloud, edge, and model-training usage.
  2. Including our providers' published environmental commitments in our vendor-selection criteria.
  3. Treating inference efficiency as an engineering goal, not only a cost goal.
  4. Publishing what we measure — including where the number is an estimate — in our first impact disclosure.
How we will measure and report

Three steps, in order. None of them finished.

An SDG mapping is worth nothing if it is written once and never checked. These are the steps we are working through, stated as plans. The output gets published when it exists, not before.

01

Feature-to-target mapping

An internal audit tying each shipped module to the specific SDG targets it genuinely supports, and recording the ones it does not. This page is the first pass, and it will be corrected as that audit continues.

02

Materiality matrix

A published view of where legal and compliance risk overlaps with the social and environmental issues our customers are accountable for. Planned, not yet published.

03

Impact disclosure

An annual report covering what we can actually count: compliance tasks automated, policy decisions logged, commitments met and commitments missed. We will publish the methodology before any number, and we will not count incidents that never happened.

Our security and certification statuses are reported on the platform page and kept current there, rather than restated here.

Hold us to this page.

If you are an enterprise buyer, an investor, or a regulator reading this as a representation rather than as marketing — that is the correct way to read it.

Ask us what is measured, what is estimated, and what is not built yet. We will answer in writing.

Talk to usSee our compliance posture