Good suggestion. Meeting notes should work as normal input to the agent; the team will test this scenario and follow up.

Generic AI assistants require exhaustive prompt engineering to structure complex diagrams. In contrast, AI in Prolaborate is built with specialized domain tools specifically designed for BPMN 2.0 process modeling.

When you describe a business workflow, our agent analyzes actor responsibilities, handoffs, and process boundaries to infer logical pools and lanes. It automatically groups tasks under the appropriate participants while applying standard BPMN layout rules, removing the need for tedious formatting instructions.

Every process diagram generated by AI in Prolaborate or authored through our web editor is saved directly to your Enterprise Architect repository in real time.

Because changes commit straight to the underlying repository database, desktop Enterprise Architect users can open and view the updated diagrams immediately. Within Prolaborate web views, such as the My Diagrams page, the new diagram appears as soon as the background data cache refreshes to render the web preview.

Yes. Set up the process in a branch or dev repository and migrate it to prod, the same way change management works in standard Sparx EA practice.

Use the regular change management approach: make changes after a baseline or in a branch, then merge after review/approval (see Hack 4).

They’re starting with standard diagrams, BPMN first, then Class (blocks) and Sequence. Custom languages need more modeling support; for those, use Enterprise Architect.

Relationships can be created and edited in the matrix.

For specific use cases like BPMN, offline LLMs are available. Most cases use public or private LLMs; some clients host private LLMs inside their own network.

In this initial preview, layout adjustments and path highlighting are performed manually within the interactive analyzer canvas.

Our current development focus is targeted on high-friction, time-consuming tasks, such as dynamically querying multi-level dependencies, creating dashboards and discovering connected artifacts across repositories. Because automated layout arrangement is a clear driver of architect productivity, expanding conversational controls to include automated graph formatting and visual path styling is on our plan.

Most LLMs already know BPMN rules, and such rules can be incorporated.

Yes. It works in both On-prem and Cloud versions; you connect your own AI subscription to Prolaborate to enable it.

Matrices created natively in Prolaborate are specialized web views designed for property-level and artifact-level analysis, and they are saved directly within Prolaborate DB.

However, Prolaborate supports bidirectional matrix visibility of basic type matrix. So any relationship matrix configured inside Sparx EA using the package selection method can be opened, and rendered seamlessly within Prolaborate matricies.

Visio process diagrams can be imported into Enterprise Architect with this feature, which can also help refine them and fix their layout.

It supports BPMN today. Based on customer feedback, Class and Sequence diagrams are planned.

Architecture governance reviews are mostly expected to be manual, but the team will analyze this feedback further.

Not at this time. Our initial release focuses purely on accelerating day-to-day modeling, analysis, and governance activities, such as creating process diagrams, creating dashboards, and configuring reviews. Repository lifecycle management tasks, such as automated cross-repository migration are broader operations that will be evaluated for future plan.

we recommend using either Claude Opus or Claude Sonnet 4.5 and 4.6 with Kernaro.

While Kernaro supports major enterprise providers, including Azure OpenAI and Anthropic, our benchmarking shows that Anthropic Claude delivers the highest quality results across complex architectural workflows. In internal evaluations—such as automated dashboard creation, structural layout composition, and diagram synthesis, Claude consistently demonstrated closer alignment with user expectations and produced more refined, context-aware outputs.

We are targeting our next major Prolaborate release between mid-December and early January.

This upcoming release will incorporate the foundational AI capabilities previewed during the webinar alongside diagramming and matrix tools. Customers interested in previewing these hacks before general availability can reach out directly to support@prolaborate.com to join our early access program.

Yes, that seamless continuity is the core design objective behind AI in Prolaborate. In our webinar, we deliberately demonstrated each capability independently so attendees could grasp the specific mechanics behind diagrams, matrices, analyzers, reviews, and dashboards.

Our conversational agent is designed to execute this complete lifecycle sequentially, from extracting requirements and generating process models to building impact analyses, configuring compliance reviews, and summarizing closeouts. We are preparing a complete end-to-end video walkthrough illustrating this full workflow, which will be released on our official YouTube channel shortly.

Yes, conversational prompts can define both the widgets and the overall dashboard structure.

In your prompt, you can specify the target widget types along with their positioning sequence. If widget types are not stated, the agent intelligently selects the most effective visual formats for your data query. While users do not need to calculate pixel measurements, our layout engine automatically manages sizing and spatial margins, eliminating dead blank space and avoiding cramped grids to deliver a clean, business-ready executive view.