Prolaborate or Kernaro AI Hub: Choose by the Job Your Stakeholders Need Done

Both of these products help people who will never open Sparx EA, so the useful question is what job your stakeholders are trying to get done. Prolaborate is Sparx Systems' web portal for browsing and reviewing curated architecture, and Prolaborate 6 adds AI that generates descriptions, summarizes discussions and diagrams, rephrases text and answers questions about the model. Kernaro AI Hub is a web application that lets stakeholders ask the live repository questions in plain language and get answers back. Since Prolaborate gained AI of its own, the old shorthand that one publishes and the other answers no longer holds. The choice now follows the people and the task.

The problem behind both is the same one: the people who most need architecture answers are the ones least able to open Sparx EA. The sections below describe the jobs, how each is done today and done AI-augmented, the skill each one asks of the architect, and which tool helps.

Three jobs stakeholders need done

Review and sign off a fixed set of views

Today. The architect exports diagrams, sends them round, and collects comments in email threads that nobody can reassemble.

AI-augmented. The architect publishes the views to a portal stakeholders use, and reviewers comment in place. With Prolaborate 6 with AI switched on, a reviewer can also ask for a summary of a diagram or a long discussion, or for a description rephrased in plain language.

The skill. Choosing which views a given audience should review, and writing the request so the AI's summary covers what the reviewer has to decide. The architect still checks the summary against the diagram.

Tool that helps. Prolaborate, with its AI for stakeholders already working there. Inactive by default, it needs a Kernaro AI Hub license and your own LLM.

Ask an open question across the repository

Today. A project manager emails an architect: which systems touch customer data, what supports onboarding, what has no owner. The architect searches, builds a list by hand, and replies days later.

AI-augmented. The stakeholder asks in plain language and the AI answers from the repository as it stands today.

• “Which applications support the Customer Onboarding capability?”
• “What is the current lifecycle status of our CRM applications?”
• “Which technology components have no documented owner?”
• “Which capabilities have no application coverage?”

The skill. Asking a question the model can answer, and knowing whether the answer is complete. An answer built from a repository where a fifth of the values are filled in looks complete and is not. Architects who watch what stakeholders ask also learn where the model needs work.

Tool that helps. Kernaro AI Hub, which Sparx Systems licenses and which is bring-your-own LLM: your organization supplies the keys for its chosen provider and pays that provider for usage. It connects to tools such as Jira, Confluence, Azure DevOps and SharePoint, and includes a read-only MCP server that MCP-compatible assistants can query. Hub needs an MDG Technology defined for the repository, so repository readiness comes first. Kernaro AI Hub launched in January 2026.

Get answers where stakeholders already work

Today. Architecture lives in one tool and the people who need it live in Teams, Word and Outlook.

AI-augmented. Stakeholders ask Microsoft 365 Copilot and it answers, read-only, from a Power BI model built on your EA data. With Kernaro AI Hub in place, the Kernaro Connector for Power BI is the preferred route.

The skill. Deciding which questions the published model should be able to answer, which fields it needs, and who may see them.

Tool that helps. Microsoft 365 Copilot for the reading, and we can help you set up the route behind it. Our reporting and Power BI page covers what that involves.

Side by side

Neither one builds the model

These two serve the people who read the architecture. Building and updating the model is a different job: turning design documents into elements, producing diagrams for every application, keeping mappings current as projects deliver. That job has its own helpers, chosen by how the architect works. Kernaro Assist sits inside the EA client for the architect at the desk. Claude or GitHub Copilot with AI Power Tools for Sparx EA, which is free and published by NovoCircle, lets an architect state what is needed and have the AI do it on the repository, including bulk work. AI tools for Sparx EA sets all six options side by side.

The model is what both stakeholder tools read from. A well-built, well-governed model improves what Prolaborate shows and what Hub answers, and a thin one makes both look unreliable. Architects who can state clearly what the model should contain, and review what the AI builds, are the ones whose stakeholders trust the answers.

When to run both

Many stakeholder groups need both jobs. The technology portal, the standards register and the published capability map suit Prolaborate, where reviewers study and comment. Executive questions, governance checks and ad-hoc analysis suit Kernaro AI Hub, where someone just wants an answer. Sometimes you want to study a map and sometimes you want to ask a question. The wider picture is laid out in AI-Augmented Architecture.

Frequently asked questions

AI is only as good as the person prompting it, and the same is true of the model it reads from. We can teach your architects to work this way on your own repository, or build and update the model for you so both stakeholder tools have something solid to work from. For the options side by side, see the full comparison of AI options for Sparx EA.