Rather than programming AI agents with a company's formal policies, a more powerful approach is to let them observe thousands of actual 'decision traces.' This allows the AI to discover the organization's emergent, de facto rules—how work *actually* gets done—creating a more accurate and effective world model for automation.

To personalize his email-sorting agent, Notion's co-founder didn't manually label data. Instead, he prompted the agent to ask him questions about which emails to archive. This interactive 'interview' process allowed the agent to learn his preferences and generate its own rules from the conversation.

Vercel's CTO Malte Ubl suggests a simple method for finding valuable internal automation tasks: ask people, "What do you hate most about your job?" This uncovers tedious work that requires some human judgment, making it a perfect sweet spot for the capabilities of current-generation AI agents.

Vercel builds internal AI agents and tools, like an Open Graph image generator, to automate tasks that were previously bottlenecks. This not only increases efficiency but also serves as a critical dogfooding process, allowing them to innovate on their core platform by building the tools their own teams need.

Sendbird created an internal platform where employees post 'quests' for AI tools. This marketplace connects needs with builders (engineers or AI-enabled staff) and even AI agents, bypassing slow prioritization processes and fostering a building culture.

You don't need to be a developer to build custom marketing automation. By describing your workflow, providing screenshots of errors, and having a back-and-forth conversation, you can guide an AI like Claude to build a tailored software agent for your specific needs.

To build its internal AI system, Cloudflare set up an email address that employees thought was an advanced AI. A human team fulfilled the requests behind the scenes, allowing them to precisely map the company's key 'jobs to be done' before building the actual automation.

Instead of manually crafting complex "mega prompts" or training rules for AI assistants, ask the AI to generate them for you. You can have a dialogue with the AI to refine its suggestions, dramatically speeding up the process of creating sophisticated workflows.

The barrier to creating AI-powered solutions has dropped dramatically. An HR team member with no AI expertise built a Slack bot trained on the employee handbook to answer common questions, saving hours of repetitive work. Every department should be empowered to identify and automate its own low-value, repetitive tasks using accessible AI tools.