Our architecture

A composable system of governance artifacts, supporting defensibility by construction.

The Voventis platform’s architecture follows directly from its understanding of the need for the governance function to secure not solely the correctness of institutional decisions, but also their defensibility – the capacity for those decisions to be revisited, inspected, explained, and justified over time.

Voventis supports this defensibility by construction. It is an architecture for producing, transforming, and committing to objects of institutional reasoning – and for preserving the interpretive pathways through which those objects come to justify decisions, disclosures, and outcomes.

Defensibility requires intelligibility

Voventis supports the defensibility of decisions by ensuring that they remain intelligible over time. This intelligibility requirement is not satisfied by preserving decisions and outcomes alone; rather, it implies an ability to examine the process by which those outcomes were reached, the reasoning which underpinned relevant decisions, the sources and information which informed them, the assumptions they embodied, the alternatives that were considered or rejected, and the substantive changes in decision-making context between the moment of institutional commitment and the moment of scrutiny.

Architectural conclusion

In light of this perspective, the Voventis platform becomes concerned with governance artifacts which are broader in scope than pure outcomes. These artifacts are produced as stable units of justification – which can persist through time, evolve, and remain faithfully interpretable.

The platform’s architecture is a system for creating these governance artifacts, for rigorously committing to them as stable units, and for assembling them into broader, coherent accounts of institutional justification.

The Voventis platform is built upon two core entities: governance artifacts and the mechanisms which transform them.

Artifacts are persistent, structured objects that represent the work products of governance and act as the platform’s units of justification.

Actions are the unique mechanism by which artifacts may be created, modified, interacted with, and published.

Artifacts

Artifacts are characterized by their stable identity, which ensures that they remain identifiable across time; by their explicit structure and content, which helps render them intelligible; by their revisability, which allows them to evolve and change without losing their history; by their modularity and composability, which permits them to connect with and depend upon other artifacts; and by their lifecycle, which governs the states they may occupy across their progression towards commitment.

Actions

An action is an explicit, accountable transformation of an artifact’s state. Actions are the only way for artifacts to come into being, change, or become committed.

Consequently, every meaningful change to artifacts occurs within an admissible, validated action context.

The platform retains an append-only narrative of all executed actions and changes.

This formal execution boundary enables actions to validate and preserve the integrity of work. Further, action execution history provides a record of the nature and circumstances of change over time, helping the platform support procedural defensibility, in addition to substantive defensibility.

The workspace

Artifacts and actions exist within a structured workspace.

The workspace does not impose predetermined workflows or provide judgments regarding how artifacts should be substantively defined, which actions should be taken, or under what interpretation institutional commitments should be justified. Rather, it provides a native structure within which those judgments may be formally expressed.

This workspace also provides a collaborative environment which institutions can safely govern. It supports membership roles, user permissions, collaboration, and project-based organization for work.

Within the Voventis architecture, individual artifacts do not exist in isolation. Instead, they are designed to be assembled into a web of justification, from which substantive accounts may be derived, described, and defended.

Composition is the core mechanism by which this assembly occurs. It reflects governance work as being subject to an interpretive discipline which produces, revises, iterates upon, and assembles justificatory material under evolving constraints, continually relying upon its own informed discretion, judgment, and expertise to re-direct itself and its work. Correspondingly, the composition mechanism captures the products of governance work as a system of connected objects, which evolves explicitly under the discretion of the governance function. Even as that discretion remains unconstrained and free to operate over its own work products, the formality of this underlying medium within which it is expressed protects its defensibility and its capacity to be directly inspected.

In practice, this mechanism allows artifacts to rely upon and inform one another, and for a coherent narrative to be assembled from multiple distinct artifacts without undermining the self-contained coherence of the individual artifacts comprising it. This property correspondingly supports artifact reuse, allowing institutions to explicitly reiterate justificatory reasoning, rather than reconstructing or paraphrasing similar rationales across multiple disclosures or artifacts.

The Voventis architecture organizes artifacts into four core classes, each of which plays a distinct role in supporting the production, articulation, and defense of institutional justification.

Source

A source represents stable material that the institution is willing to treat as a basis for reasoning. In practice, sources provide an evidentiary ground for governance work, and typically include documents, datasets, tables, and other types of information.

Analysis

An analysis artifact is used to produce findings. It describes a structured transformation; given defined inputs and sources, it applies a methodology to produce a result.

Analysis artifacts are the executable context for the methodology they define; particular analyses are immutable specifications of that context into a set of inputs to that methodology and its resultant finding.

Policy

A policy artifact is a normative expression. Policies may be used to articulate a range of normative concerns, such as institutional principles, standards, ethics, rules, or constraints.

Policy artifacts can be composed of both narrative and structured policies. Narrative policies enable the institution to explain rationale, principles, and intent in linguistic terms; structured policies support the explicit definition of constraints, conditions, thresholds, and other logical expressions. Correspondingly, structured policies may be used to formally derive compliance expectations and verify operational alignment.

Report

A report artifact represents a mode of representation. Its purpose is to communicate and substantiate particular accounts of justification – internally or externally. Reports are composite objects, supporting structured presentation, disclosure, and explanation.

Drafting, publication, and institutional commitment

The artifact lifecycle distinguishes between work that remains in formation and work that has reached a state of stable commitment.

Correspondingly, the Voventis platform recognizes working states and sealed states for artifacts. Drafts in working state remain within a space of iteration and continuing discretion. Within this space, governance work remains exploratory, open to revision, and often incomplete. By contrast, published artifacts enter a space of institutional commitment. Published artifacts become sealed as defensible units of justification – demonstrably stable, immutable, and fit to be cited or relied upon.

Publication as sealing

Within Voventis, publishing is not equivalent to saving. Saving remains normally applicable to work-in-progress; publishing is the act of sealing which declares an artifact fit to be relied upon, defines its committed state, and fixes it as a point of reference.

What publication guarantees

Publication turns working material into a committed object of record. This action of publishing an artifact carries architecturally-enforced guarantees.

Firstly, committed work cannot silently change. Once sealed, a published artifact remains what it was at the moment of commitment and becomes a stable point of reference.

Secondly, commitment is an additive rather than overwriting mechanism. Each act of publication creates a new committed version. Prior commitments remain available for comparison and scrutiny.

Thirdly, dependencies are sealed alongside the artifact. This implies that working references between draft artifacts are converted into committed citations alongside their publication. This mechanism structurally protects the institution from classes of risk created by shifting dependencies.

Analogously, generated content is committed alongside its exact basis. For example, where an artifact contains content that was produced through analysis or other forms of computation, publication commits not only the outputs themselves, but also the inputs and methods used to produce them.

Fifthly, lineage is made explicit as a navigable map of dependencies. Once sealed, an artifact is not only preserved in isolation, but as part of a broader map of what the artifact relies upon – and, as the map grows, of what relies upon it.

Finally, publication maintains continuity between the artifact and future work. After publication, the working copy in the platform’s workspace begins from the newly sealed commitment, so further edits are maintained as deliberate revisions of a commitment – not as mutations of it.

Versioning and comparison

Each publication creates a new version; the artifact’s identity persists across versions, but its commitments are discrete and enumerated. Comparisons help explain the progression of commitments over time – making it possible to examine what changed, when, and for what reasons.

Judgment, interpretation, and discretion are not only drivers of the interpretive discipline which animates and coheres governance work; they are also exercised directly in the production of specific work artifacts under that discipline.

Consequently, Voventis respects the role of this latter form of discretion within the identity of an artifact itself, by distinguishing between what is authored and what is computed – that is, between content that is written, chosen, specified, or asserted by an institutional actor, and content that is derived, generated, validated, or calculated as a finding.

The architecture explicitly separates these concerns without erasing the connection between them. Authored material remains the primary definition of an artifact – specifying what it is, what it claims, and what it commits to. Computed material remains derivative and anchored to definite inputs.

Alignment and drift

In light of the iterative and often provisional nature of the draft work underpinning governance artifacts, computed content often arises within an evolving context and with respect to mutable references. In other words, given that drafts create live, working references, Voventis must be able to accommodate their change without silently corrupting the defensibility of computed outputs.

Given this, computed content presents a distinctive risk: it can become epistemically invalid or outdated as a consequence of its dependencies changing, without itself visibly changing.

The system’s map of dependencies enables Voventis to transform this dependency drift within the drafting stage of the artifact lifecycle from a risk into a visible, actionable condition.

If and as the working references upon which draft artifacts depend evolve, Voventis surfaces this drift as an explicit alignment signal and supports regeneration – thereby preserving the integrity of exploratory work without collapsing either its flexibility or its distinction from committed work.

Published commitments, by design, do not drift; the action of publishing converts working references into committed citations and seals dependencies.

Composition and the ability of artifacts to meaningfully relate to and rely upon one another in the Voventis platform is broadly supported by its referencing system.

Most of the persistent work or source material produced within or admitted into the platform becomes available as referenceable material. This referenceable material acts as a substrate for further work. Referenceable material takes on many forms, including artifacts, computed findings, structured policy outcomes, source materials, and other items within a project or workspace.

Across different contexts, this material may be reused, cited, or built upon by continued work. Particular instances of this further reliance are specified through the combination of a reference and a mode of reliance.

The reference provides the structured mechanism for addressing or pointing to different types of referenceable material.

The mode of reliance ascribes contextual meaning to that reference. A mode of reliance is given by the substantive or causal context within which the reference is created, not by the nature of the referenced material itself. Through diverging modes of reliance, the same references can carry different meanings across different contexts.

Each instance of reliance becomes part of a project or workspace’s overall map of relationships. In aggregate, this map ensures that the broader system of artifacts is navigable as a chain of reasoning, thereby allowing each incremental change, publication event, and proposed action to be understood as a part of – and to be validated against – a broader body of work.

The lineage map broadly distinguishes between core lineage and projection lineage. Core lineage encompasses instances of reliance between artifacts, while projection lineage includes reliance between artifacts and artifact-derivative materials, such as immutable computed findings or action execution records.

Finally, a web of justification is a causal interpretation of the lineage map with respect to a given artifact. It is an ordered view containing the exact subset of lineage relevant to that artifact.

The consequence of this referencing system is that Voventis explicitly distinguishes between what is being referenced, how it is being referenced, why it is being referenced, the durable relationship which arises from that reference, and how that relationship supports broader justification.

Actions are the platform's unit of accountable change.

This action framework ensures that change does not occur through implicit edits, side effects, or informal application logic, but instead through an explicit sequence of accountable transformations with definite characteristics. Actions are therefore not primarily records of change after the fact; they are the mechanism by which change is validated and becomes admissible.

Firstly, actions are contextual. State changes within the platform are rejected unless they occur within a valid action context. This action context establishes who is acting, which organisation they are acting within, what operation is being attempted, what target it applies to, what permissions and capabilities are required, and whether the change satisfies the architecture’s integrity requirements.

Secondly, actions are explicit. They act only as formal operations over artifacts and platform objects, not through side effects.

Thirdly, actions are attributable. They are tied to an actor, a time, and an execution context.

Fourthly, actions are substantively auditable. They preserve the inputs used and the outcomes produced.

Fifthly, actions are idempotent and concurrency-safe. They remain protected from silent duplication, stale draft changes, and other forms of technical corruption.

This execution framework for actions supports procedural defensibility – enabling the institution to retain a comprehensive understanding of which actors made which changes, at what time, on what basis, and to what effect. In this way, actions help ensure that not only the substance of governance work, but the procedure by which it was produced, remains transparent and defensible by construction.

Voventis supports extensibility by accepting domain packages in an architecturally-native manner. Domain packages provide a mechanism for specifying and adding bespoke capabilities to the platform, such as analytical methodologies, structured policy baselines, and reporting patterns, which organizations may adopt and adapt within their own workspace and governance domains.

Domain packages seamlessly extend the expressive capacity of the Voventis platform within the same architecture, ensuring that specialized platform use does not come at the expense of defensibility.

Voventis is designed for institutional use. It approaches security, provenance, and access control as native parts of the platform’s architecture, not as peripheral features. Work products are consistently treated as structured institutional material, with database-level enforcement of the platform’s architectural guarantees.

The platform is built around tenant isolation, role-based capabilities, project-matter boundaries, controlled file handling, immutable published versions, action-based mutation, provenance-bearing computation, and explicit auditability.