UX patterns for the Context Plane

Ways that the Context needs to be accessed and by whom

“One size doesn’t fit all”.

Different personas and use cases demand different ways of interacting with Context.

“Context” of this post

I often find writing helps me coalesce and refine my thoughts when new patterns start to emerge, but aren’t very clear yet.

So this article is a brain dump / train of thought continuation of the architecture needed to have one Context Plane to rule them all, as part of a proposed “AI Data Stack”.

This article provides an overview of the specific persona types / use cases I have identified so far that need to access the Context Plane and the typical UX patterns for some of those.

Updated Context Plane Architecture

This iteration in my thinking has a resulted in an update to the architecture diagram for the Context Plane:

The four personas / use cases

Here are the 4 persona types / use cases I have discovered so far:

• Human, GUI centric
  A person who wants to use a web based App, chat or Graphical User Interface to access/discover/explore/update the Context.

• Human, Code Centric
  Wants to use a Command Line Interface (CLI) or Code based App to access/discover/explore/update the Context.

• System
  Systems can access the Context directly, either querying the Context or creating to it programatically.

• Agent to Agent
  AI Agents can access/discover/explore/update Context autonomously, collaborating with other agents without a human in the loop.

The UX patterns

Here are the UX patterns I have experimented with that seem to make sense:

• Human, GUI centric

  • GUI centric Data Catalog

  • GUI centric Chatbot

• Human, Code Centric

  • GenAI App

  • CLI tool

• System

  • API’s

GUI centric Data Catalog

Your typical browser based Data Catalog interface.

Prefer a manual click through version?

https://guides.agiledata.io/demo/cmflomypg048l170irpj5zf6h

GUI centric Chatbot

Your typical browser Chatbot and Text to SQL interface.

GenAI App

Typical Claude “AI Agent” app tool interface.

CLI Tool

Typical Gemini CLI command line interface.

[Screenshot TBA]

API’s

Typical API endpoints.

Missing UX patterns

The key UX pattern I have yet to discover is how the Agent to Agent UX works.

I think we will need to do a McSpikey with the Google A2A to understand the options in that space a little more.

The Technology patterns

We already had a lot of the technology patterns in place before we started experimenting with the Context Plane.

Things like the browser based Data Catalog capability.

We also had a lot of ADI already built based on “AI Assisted” features we have experimented with over the last 6 odd years.

Our App and Platform architecture has always been based on API’s in the middle:

App > API > Context > Code / Data

The main iteration technology wise has been the addition of a MCP server. This has allowed the use of tools like Claude and Gemini CLI.

We have iterated ADI to use the MCP server to access the Context (well we actually use a hybrid access model but ill leave the diagram as simple as this for now.)

So many new questions

User?

UX stands for User Experience, but some of these persona types and use cases are machines not humans, should they still be referred to as Users?

BI Semantic Layer?

Where does the typical BI Tools and the “BI Semantic Layer” pattern fit into this?

For Context Plane we are only holding an Organisations Context not their Data, so we can’t execute any queries, like we can in the AgileData App and Platform.

Or do we want to look at generating the query the human can cut and paste into the data platform.

Or do we want a Context Agent to push the Query to a BI agent inside the Organisations agent ecosystem?

We don’t provide a caching layer or query rewrite patterns which is what the BI Semantic Layers / Metric Layers are doing these days. Im pretty sure we don’t want to go there.

When will BI Tools move to using MCP servers as a way of querying the data?

When will they all put Agents in from of their BI Semantic Layers"?

One step forward, but a raft of new uncertainties

So many new questions, so few answers.

Looks like there even more McSpikeys to add to the list!

Wood from the Trees

Still a way to go before I have a coherent set of Patterns that I can Coach / Mentor / Teach somebody else for the “Context Plane”, and the “AI Data Stack” or present as a robust Architecture map.

But as I have already said, writing my half formed ideas helps me think.

An incoherent stream of Context

You can find all the previous articles with my train of thought listed in this thread:

https://agiledata.substack.com/t/context-plane

We are building the Context Plane while flying it, so always looking for early adopters to help us decide the final destination

If you want a virtual chat grab a slot here:

https://contextplane.ai/contact-us/#bookemdanno