Scaling Open-Source
documentation with AI
Bethany Jepchumba | @bethanyjep
Slide 2
Slide 2 text
Would you trust an AI to …?
Slide 3
Slide 3 text
Pick a movie or book you’d enjoy?
Slide 4
Slide 4 text
Drive you in busy Thika Road Traffic?
Slide 5
Slide 5 text
Recommend your next career move?
Slide 6
Slide 6 text
Assist you in coding?
Slide 7
Slide 7 text
Write and send emails?
Slide 8
Slide 8 text
Write documentation?
Slide 9
Slide 9 text
Documentation in OSS
Slide 10
Slide 10 text
- Documentation as a Service
- Agents as a Service
- Translation as a Service
- Questions?
Slide 11
Slide 11 text
What we want
- Clear, accurate, and up-to-date documentation
The struggle?
- maintaining robust documentation at
scale is often resource-intensive, manual,
and challenging
Slide 12
Slide 12 text
Automating and streamlining workflows…
How can we integrate
AI to how we create,
maintain & translate
OSS documentation?
Slide 13
Slide 13 text
Documentation as a Service
Slide 14
Slide 14 text
Prompty
https://github.com/microsoft
/prompty
Slide 15
Slide 15 text
What we want
- Clear, accurate, and up-to-date documentation
The struggle?
- One maintainer
- Widely used across presentations
- Not enough technical depth on the project
Slide 16
Slide 16 text
Prompty as a language-agnostic prompt asset
Specification
Prompty is intended to be a language
agnostic asset class for creating prompts and
managing the responses.
The goal is to simplify your workflow by
creating a standard that can be used by any
language, any framework, and any tool to
create a prompt and manage the response.
Prompty can be imported to/exported from
Azure AI Studio for smooth transition
between local and cloud. Prompty assets can
be also shared across organizations.
Slide 17
Slide 17 text
Prompty Tooling in Visual Studio Code
Tooling
Visual Studio Code Extension
• Frontmatter Autocompletion
• Colorization / Syntax
Highlighting
• Validation
• Quick Run
• Code Generation
• Evaluations – Future
• ASK US
Slide 18
Slide 18 text
Prompty Runtime Library
Runtime
Prompty runtime is the whatever engine
that understands and can execute the
format.
• Azure AI Foundry
• Python – LangChain
• C# – Semantic Kernel
• TypeScript/JavaScript – Extension
• Your Language – Implement your
own!
Slide 19
Slide 19 text
Scenario
A ton of empty doc pages with
some understanding of the
codebase, how could we make it
to 100%
Slide 20
Slide 20 text
Agents as a Service
Slide 21
Slide 21 text
What we want
- Streamlined and well-maintained docs
The struggle?
- One maintainer
- Lots of issues & discussions
Slide 22
Slide 22 text
AI Agents
Unlike existing automation,
AI Agents can automate requests
as soon as they come in by…
Reasoning over the request
Proactively retrieving more context
Performing required actions with APIs
Slide 23
Slide 23 text
What is an AI agent?
LLM
Instructions
Tools
Agent
+
+
An AI agent is a micro-service that
takes unstructured messages,
optionally invokes other APIs
and returns messages
1
2
3
Input
System
events
User
messages
Agent
messages
1
Tool calls
Knowledge Actions Memory
2
Output
Agent
messages
Tool
results
3
Slide 24
Slide 24 text
Unlike previous tools, AI Agent can…
reason, act and learn
Slide 25
Slide 25 text
Agentic patterns
Acting
aka tool calling
Learning
aka memory
Reasoning
aka planning
I’ve captured the underlying
issue in the details
{
"name" : "John Clar
"issue": ”Washing m
Create a JSON describing
the user issues
Check logs
Machine is past due for maintenance
Go beyond transcriptions and slot
filling by allowing agents to deduce
the why
I’ll look at the logs to see what
happened
Transcript
Fill out and submit the workorder
Populate field: name
The form has been successfully
Agents can take a simple request and
chain multiple tool calls to complete the
ask
Only make a mistake once. With memory,
agents can recall prior experiences
Populate field: issue
Name field populated
Issue field populated
Submit form
Update work order
Let me try again with an
available technician.
Update work order
Alex is already booked at this same time
Successfully assigned technician
Assign technician
Saved memory: Alex is OOF next week
Slide 26
Slide 26 text
No content
Slide 27
Slide 27 text
Scenario
An open-source maintainer for
Prompty looking for a way to
simplify tagging issues.
Slide 28
Slide 28 text
Auto-tagging
issues & discussions
Slide 29
Slide 29 text
Translation as a Service
Slide 30
Slide 30 text
What we want
- Localized docs
The struggle?
- One maintainer
- Lots of issues & discussions
Slide 31
Slide 31 text
https://github.com/Azure/co-op-translator
Slide 32
Slide 32 text
From en to xx in a single click
Slide 33
Slide 33 text
No content
Slide 34
Slide 34 text
Scenario
How can a global educational
curriculum with >2M views be
made accessible?
Slide 35
Slide 35 text
Auto-translation for ml for
beginners curriculum
https://github.com/microsoft/ML-For-Beginners
Slide 36
Slide 36 text
https://aka.ms/ai-agents-prompty
Resources and code!
Slide 37
Slide 37 text
https://aka.ms/azureaifoundry/discord
Resources and code!