OpenAI showcased how its teams use AI agents for competitive research, launch monitoring, and internal coaching, freeing up time for high-value work.
OpenAI hosts Nick and Danielle surveyed the audience, highlighting the common challenge of starting a workday amidst an influx of emails, Slack messages, and various dashboards.
This initial management of digital clutter often consumes a significant portion of the morning, delaying engagement with more substantive tasks and reducing overall productivity.
OpenAI introduced key internal figures — Dani, Adrian, and Casey — who serve as 'outliers' within their departments, leveraging AI to achieve previously unmanageable tasks.
Dani leads social media for OpenAI developers, Adrian spearheads competitive research, and Casey functions as the Chief People Officer.
This capability allows users to offload repetitive responsibilities, such as building a custom event site for DevDay, thereby freeing up time for more high-value strategic work.
Dani, who manages social for developer outreach, monitors product launches like GPT-6 Astra and provides performance metrics to executives.
She uses a 'dot' (a ChatGPT agent) to automatically track developer feedback and engagement across social feeds, integrating this data into a content sheet.
This automation shifts her role from manual data collection to strategic analysis and allows the agent to monitor Slack channels for urgent matters, delivering timely alerts.
Adrian and his two-person team utilize a custom plugin called DOT to investigate patterns and vet competitive inquiries, ensuring that research outputs remain accurate under scrutiny.
This AI agent investigates, identifies patterns, and critically evaluates its findings to ensure the research holds up to scrutiny.
The AI agent analyzed 90 days of sales call transcripts, identifying reasons for lost deals that standard CRM summaries failed to capture.
This deep-dive methodology incorporates diverse sources, including newsletters, podcasts, and call transcripts, providing a comprehensive view of competitor performance.
OpenAI developed 'Compete Corner' as a centralized site to organize competitive research around frequently asked questions, allowing colleagues to self-serve information.
The AI agent automatically monitors Slack for new inquiries and updates the knowledge base, with Adrian reviewing and approving AI-generated conclusions before publication.
Utilizing GPT-6 Astra capabilities, the new Model Galaxy tool offers an interactive interface to track the performance, cost, and latency metrics of frontier models across various benchmarks.
Casey, OpenAI’s Chief People Officer, utilizes a shared knowledge base to help teams adapt to AI integration, developing a coaching site for leaders.
This site, powered by GPT Live, suggests improved workflows and team enhancements, enabling leaders to discuss team dynamics and identify bottlenecks.
The coaching platform is trained using compiled research from meeting notes, industry papers, and articles, ensuring relevant and practical advice.
While individual chats remain private, edits to the shared page are visible to the entire team, and page instructions, similar to an agents.md file, ensure consistent agent output.
Users can invite teammates into a shared space to collaborate on the same thread within ChatGPT, converting their work into a persistent 'page' for shared editing.
AI agents remain active in these spaces, continuously updating content as users refine their work, streamlining collaborative workflows.
These Dots integrate with Slack, sending pings only when specific user attention is required, and operate in the background to maintain context and update shared documents automatically.
Dots are accessible via Slack or voice calls, ensuring their constant availability regardless of the user's location, further enhancing seamless task automation.
Users can personalize their Dot by defining a name, visual accessories, and core responsibilities, allowing the Dot to learn individual preferences and working styles over time for improved efficiency.
Dots can integrate with external communication platforms like Slack and text messaging, expanding their utility and reach.
The next thing you're going to do is give accessories. I like having mine kind of look like the frog guy with the fun eyes and the monocle. And after that, you're going to give it a responsibility.
An interactive tool helps users identify tasks suitable for automation by inferring their work habits and suggesting initial tasks that can be handed off to Dots, streamlining the transition to AI assistance.
Automation enables professionals to dedicate more time to high-value activities such as community engagement and strategic insight delivery.
The speaker shared a personal anecdote about automating technical onboarding tasks, which facilitated a transition into a developer experience role, emphasizing that automation provides a choice in how professionals invest their 'currency' of attention.
AI integration allows Dani to increase her engagement with the developer community, Adrian to disseminate competitive insights more broadly, and Casey to share his expertise without needing to be present in every conversation.
This strategic reallocation of time enhances operational efficiency across departments.
This shift was made possible by automating significant portions of the onboarding process using ChatGPT, illustrating how delegating routine tasks creates opportunities for engaging in higher-value work.
The speaker emphasized that attention is a finite 'currency,' and individuals have the agency to decide where to invest it.
This philosophy encourages professionals to identify tasks they consistently delay or projects they wish to pursue, prompting a reconsideration of how they balance routine labor with desired, impactful goals.
Attention is currency and you get to decide where to spend it. So, let me ask you. What will you hand off? And what could you finally take on?
Answers come from the transcript, with the exact spot cited.
Want the next article from OpenAI?
When OpenAI publishes, we'll write it up like the one you just read and email it to you.
