Direct Answer: Kernaro Assist is an AI productivity tool embedded in the Sparx EA client, currently in Beta. It can accelerate element creation, suggest diagram structures, run governance checks, and answer natural language queries about the repository: without leaving the EA environment. What it cannot do is replace architectural judgment, make governance decisions that require business context, or produce reliable output from a poorly-governed repository. The MDG dependency is the most important thing to understand about Kernaro Assist: its output quality is a direct function of your repository’s MDG governance quality. An architect with a well-governed repository will find Assist genuinely time-saving. An architect working in an ungoverned repository will find Assist’s suggestions incomplete and inconsistent. Assist is a multiplier. It multiplies whatever quality already exists in the model underneath it.
Key Takeaways
- Kernaro Assist is embedded in the Sparx EA client: architects interact with it via a natural language panel without switching tools.
- Its core value is reducing the mechanical overhead of modeling: element creation, diagram scaffolding, tagged value population, governance conformance checking.
- MDG governance is the quality gate. Well-governed repository = useful Assist output. Poorly-governed repository = unreliable Assist output.
- It is in Beta. Treat it as a capable productivity tool, not a production-grade automation system. Review everything it produces.
- What it frees up time for: higher-judgment architecture work: stakeholder engagement, architectural decision-making, strategic analysis.
What Kernaro Assist Actually Does
Kernaro Assist operates through a natural language interaction panel within the Sparx EA client. Architects type requests or questions; Assist interprets them and acts on the repository or returns information. The capabilities that are functional in the current Beta:
Element creation from natural language. An architect can type “Create a Capability element called Customer Onboarding under the Customer Engagement domain package, with MaturityLevel tagged as 2 and InvestmentCategory tagged as Transform.” Assist creates the element, places it in the correct package, and populates the tagged values: faster than navigating the UI manually, especially for bulk creation tasks.
Diagram scaffolding. Given a description of a scenario: “Create an ArchiMate diagram showing the Customer Onboarding capability, the three applications that realize it, and the data entities they consume”: Assist can generate a first-draft diagram structure. The architect reviews, adjusts relationships, and corrects any misinterpretations. The value is getting to a starting point in minutes rather than constructing from blank canvas.
Governance conformance checking. Assist can be prompted to check a package or a set of elements against the MDG profile: identifying elements missing mandatory tagged values, elements created with base metaclasses where stereotypes should be applied, or naming convention violations. This is a time-consuming task if done manually across large packages; Assist accelerates it significantly.
Natural language repository queries. An architect can ask Assist “Which applications in the portfolio have a lifecycle status of Under Review and have at least two capability realisation relationships?” without writing a query or building a matrix manually. Assist traverses the repository and returns structured results. The accuracy depends on how consistently the relevant tagged values and connectors are populated.
MDG-guided tagging. When creating new elements, Assist can suggest tagged value completions based on the MDG profile and the values already used on similar elements in the repository. This reduces the blank-field problem that leads to inconsistent population.
The MDG Dependency: The Honest Version
This point is important enough to state plainly. Kernaro Assist reads the repository to generate its suggestions and answers. The suggestions it makes about element types are based on the stereotypes defined in your MDG profile. The answers it gives to repository queries are based on the tagged values and connectors that are actually populated.
If your MDG profile is undefined or inconsistently applied, Assist cannot infer what you meant by “application”: it sees a mixture of ArchiMate Application Components, UML Class elements used as application placeholders, and free-text notes. It cannot produce a coherent “list of all applications” because the repository does not have a coherent “application” concept.
If your tagged values are sparsely populated, Assist’s answers to questions like “which applications are at end of life?” return whatever is there: which may be 20% of the actual answer and misleading as a result.
This is not a criticism of Kernaro Assist. It is a description of how data quality constraints propagate through any system that reads and interprets data. The same constraint applies to EA GraphLink, to Power BI dashboards built on EA data, and to any other downstream consumer of repository content.
The implication for Assist adoption: if your repository governance is weak, the highest-value Assist use case is governance improvement: using Assist to accelerate the conformance checking, tagging backfill, and element correction work that improves the repository quality. Not generating new content from a broken foundation.
What Kernaro Assist Does Not Do
It does not make architectural decisions. Whether an application should be retired, whether a capability is strategically important, whether a particular integration pattern is appropriate: these require judgment that Assist does not provide. It can surface information relevant to those decisions (what depends on this application? what is its maturity rating?), but the decision itself remains with the architect.
It does not replace stakeholder engagement. Architecture work is fundamentally a communication discipline. Understanding what business stakeholders need, translating that into architecture, and presenting findings in ways stakeholders can engage with: none of that is automated by Assist. It affects the time available for that work (by reducing mechanical overhead), not the work itself.
It does not fix poor governance retrospectively. Assist can identify governance problems and accelerate remediation. It cannot determine the correct answer to a governance question that requires business context: which of these three identically-named applications is the canonical one? What should the lifecycle status be for a system the original owner has left the organization? Those answers require human inquiry.
It is not fully production-stable in Beta. Assist is in Beta. This means some interactions produce unexpected outputs, some requests are misinterpreted, and some features are incomplete. Every output should be reviewed before use. Build into your team workflow: Assist produces first drafts; architects review and finalize.
What Assist Frees Up Time For
The practical benefit of Kernaro Assist is not automation of architecture work: it is reduction of the mechanical overhead that currently consumes architect time, freeing that time for higher-judgment work.
In typical EA teams, a significant portion of architect time goes to:
- Creating elements manually and populating tagged values one at a time
- Building routine diagram scaffolds from blank canvas
- Checking large packages for governance conformance
- Compiling answers to ad hoc data questions (which applications connect to this service? which capabilities are owned by this division?)
When Assist handles these tasks, architects have more time for the work that actually requires their expertise: interpreting what the business needs, making architectural trade-off decisions, engaging with senior stakeholders, designing patterns that will hold up over time.
The productivity multiplier framing is accurate. If an architect currently spends 40% of their time on mechanical tasks and Assist reduces that to 15%, the time freed is significant. What the team does with that time determines whether Assist delivers business value. Assist time saved + more mechanical modeling = marginal benefit. Assist time saved + more strategic architecture engagement = genuine value shift.
Deploying Kernaro Assist Effectively
Effective Assist adoption requires three things before deployment:
1. MDG governance baseline. Assess and strengthen the MDG profile before enabling Assist. Define stereotypes, establish tagged value schemas, run a governance conformance pass on existing content. This is the prerequisite that determines Assist quality.
2. Team workflow integration. Define how Assist fits into your team’s modeling workflow. Which tasks are Assist-first (element creation, routine diagram scaffolding)? Which are always human-led (architectural decisions, stakeholder deliverables)? Which require mandatory human review before publishing (governance check results, complex queries)?
3. Feedback and iteration. As a Beta tool, Assist improves through use and feedback. Establish a simple mechanism for architects to flag when Assist output is incorrect or unhelpful. This feeds improvement both to the Kernaro Assist development roadmap and to your own MDG profile (if the errors reveal governance gaps).
FAQ
What is Kernaro Assist? Kernaro Assist is an AI productivity tool embedded in the Sparx EA client. It allows architects to interact with the EA repository via natural language: creating elements, generating diagram scaffolds, running governance checks, and querying repository data: without leaving the EA environment. It is developed by Sparx Services and is currently in Beta.
Does Kernaro Assist work without EA GraphLink? Kernaro Assist operates within the EA client and reads directly from the EA repository for in-client functionality. EA GraphLink is the connectivity layer that enables external AI tools and BI platforms to query the EA repository. The two are complementary: Assist is the in-EA architect tool; EA GraphLink enables external connectivity. Some Kernaro Assist features that involve cross-system queries may require EA GraphLink to be configured.
What is the MDG dependency for Kernaro Assist? Kernaro Assist’s output quality is directly determined by the MDG governance quality of the repository it reads. Well-governed repositories: with defined stereotypes, consistent tagged values, and populated properties: produce accurate, useful Assist responses. Poorly governed repositories produce incomplete or inconsistent responses. Improving MDG governance before deploying Assist is the most effective way to maximize its value.
Can Kernaro Assist create full ArchiMate diagrams from a description? Assist can generate diagram scaffolds: first drafts of ArchiMate diagrams: from natural language descriptions. It creates elements, places them in the correct diagram, and draws suggested connectors. The output is a starting point, not a finished diagram. Architects should expect to review and refine connector types, validate element placement, and check that the diagram accurately represents the architecture. Assist reduces the time to first draft; it does not replace the review step.
Is Kernaro Assist a replacement for architectural judgment? No. Kernaro Assist automates mechanical modeling tasks: element creation, diagram scaffolding, governance checking, repository queries. It does not make architectural decisions: whether to retire an application, which integration pattern to apply, whether a strategic investment is justified. Those decisions require context, business knowledge, and judgment that Assist does not provide. Assist is a productivity multiplier for architects, not a substitute for them.
How does the Beta status affect how I should use Kernaro Assist? Beta status means Assist is capable and valuable but not fully production-stable. Some interactions may produce unexpected or incorrect outputs. Some features are incomplete. The right approach is: use Assist for efficiency on mechanical tasks, treat all output as a first draft requiring review, build review steps into your workflow for any Assist-generated content that will be shared with stakeholders or used in governance decisions. Report unexpected outputs to Sparx Services to contribute to the improvement roadmap.
Build the Governance That Makes Kernaro Assist Work
Kernaro Assist is most effective on a well-governed repository. Sparx Services’ Amplify offering builds the MDG governance discipline and deploys Kernaro Assist as part of a structured program: ensuring the tool has the foundation it needs to deliver real productivity gains.
If you are considering Assist adoption, start with an honest assessment of your repository governance. Amplify addresses both in a single engagement.
