Top 5 Positron Takeaways from posit::conf: AI Agents, Polyglot Workflows, and Quarto Notebooks

Positron is Posit’s modern IDE for data science. Since its general release, our team has been working hard to develop features that support modern, multi-language workflows, and we’ve been delighted to see the community adopt Positron and build amazing things with it.

And we got to hear all about it at our annual conference, which we just wrapped up! At posit::conf, speakers highlighted major highlights and product updates for the Positron IDE:

1. Polyglot runtime execution: Native, equal-citizen support for R, Python, and SQL in one editor.
2. Literate notebook evolution: Built-in Quarto integration with interactive data previews.
3. Context-aware AI agents: Transitioning from basic completion to environment-aware assistants.
4. Code-first workflows: Moving past static BI dashboard limits into scalable data products.
5. Flexible cloud deployment: Frictionless setup across local laptops, containers, and Amazon SageMaker.

A huge thank you to all our speakers for joining us to explore the future of Positron, polyglot data science, and AI-assisted workflows.

Missed a session? Registered attendees can access session replays on-demand by logging into pos.it/conf (select ‘needs based’ for a free ticket) and heading to the “Schedule” tab. Public YouTube releases will follow soon!

Curious what stood out to us from all these talks? Here are our five biggest takeaways:

1. In a multi-language stack, execution environments matter most

Modern data science has moved past rigid language boundaries; teams care far more about solving problems than debating syntax. But multi-language workflows still require a reliable execution environment to run smoothly.

Positron acts as a true polyglot engine, giving R, Python, and SQL first-class runtime citizenship in a single editor so you can focus on analysis instead of toolchain friction.

Watch these talks to learn more:

• SQL as a first-class language in your data science IDE by Julia Silge and Brian Lambert
• From Programming Alone to Programming and Prompting: Teaching Bilingual AI-Enabled Data Science with Positron and Quarto by Derrick Cogburn

2. Notebooks and literate programming are evolving

From Jonathan McPherson’s dive into Quarto Notebooks to sessions exploring notebook history and design, the talks demonstrated how modern exploratory tools should mirror how data scientists actually think and work.

Positron reimagines literate programming by integrating native SQL query execution, interactive data previews, and deep Quarto publishing tools directly into notebook environments.

Watch these talks to learn more:

• Quarto Notebooks and Literate Programming in Positron by Jonathan McPherson
• Positron Notebooks: Built Around How You Actually Work by Nick Strayer
• Where Do Notebooks Come From? by Wasim Lorgat

3. AI is moving from “Copilot” to “Context-Aware Teammate”

AI-assisted data analysis was a central theme across the conference, with sessions showing how AI tools are evolving far beyond basic auto-complete. Speakers tackled key questions around AI integration, featuring live demos of Posit Assistant, Commons, and custom AI agents.

AI integration in Positron focuses on context awareness: feeding agents relevant environment, database schema, and project context to produce reliable code and minimize LLM hallucinations. Speakers also covered prompt engineering versus traditional coding, error handling, and enterprise AI governance.

Watch these talks to learn more:

• What context do your AI agents actually need? by Sharon Wang
• Three Agents, Two Mistakes, or Designing for LLM-Assisted Data Analysis by Sam Clark
• AI Agents for Data Science: Posit Assistant, Commons, and Beyond by Joe Cheng
• AI Governance: Bringing Posit Assistant to the enterprise by Tom Mock

4. Positron in Practice: Workflows Beyond the “Tableau Plateau”

Several talks focused on how breaking out of traditional BI tools allows organizations to scale complex analysis. Positron empowers teams to scale beyond UI boundaries by giving creators the tools to transition from static dashboards into maintainable, code-first data products.

Watch these talks to learn more:

• The Tableau Plateau: From Dashboards to AI-Ready Data Products by Blake Abbenante
• Making the Impossible Possible: A Software-First Stack for Resource-Constrained Research by Ran Li

5. Lowering barriers to adoption through extensibility, usability, and deployment

A data science IDE is only as good as its deployment story. It needs to feel identical whether running on a local laptop, a departmental server, or a managed cloud container. However, deploying new tools across an organization involves administrative and adoption hurdles.

Positron provides a consistent operational environment across platforms. The conference talks tackled the realities of getting Positron into users’ hands, overcoming switching friction and enterprise administrative deployment, and what Positron provides to reduce administrative setup friction and simplifying organizational rollout on services like Amazon SageMaker.

Watch these talks to learn more:

• Positron Everywhere All At Once by Isabel Zimmerman
• Positron and the Three Bears: Teaching Python with Positron by Marc Dotson
• Can I offer you a nice IDE in this trying time? by Luis D. Verde Arregoitia
• Positron for Everyone: Organization-Wide Posit Tooling on Amazon SageMaker by James Blair and Subodh Kumar