Getting work done with AI has never been faster. Give it a clear task, and it can complete in minutes what once took hours.
Databox’s AI Analyst is a good example. Ask it to analyze the performance of your team or channels, and it can take the work from your initial question to a finished report in one conversation.
The result, though, depends on the instructions you give it. If you want your AI Analyst to follow your usual process, your prompt needs to explain which metrics to review, what goals to compare them against, which time period to cover, and how to structure the final report.
That setup is manageable for a one-off task, but for recurring work, it means providing that same level of detail every week or month.
What if you could teach your AI Analyst that process once, and have it follow it every time?
Meet Skills
A Skill captures the steps, standards, and context behind a task, so your AI Analyst can follow the same process whenever the work comes up. Define it once, and get consistent output without rewriting the prompt each time.
Build your process into your AI Analyst
Get analysis and reporting done faster, and have it done the way you want.
Create a Skill from an existing process or a conversation
Write the instructions yourself, upload a file you already use, or describe the process to your AI Analyst in chat. Your AI Analyst asks follow-up questions and drafts the Skill for you to review, helping you turn what you know into a reusable process.
Start recurring work with the right process in place
Run a Skill directly with a slash command or ask your AI Analyst as you normally would. When your request matches an enabled Skill, your AI Analyst applies it automatically.
Give everyone a consistent way to work
Keep a Skill personal when it reflects how you work. When the process is ready for wider use, promote it to a Company Skill so everyone in your account can follow the same steps and standards.
Get a head start with the Marketplace
Browse Skills built for common reporting and analysis tasks, then install one directly from the Marketplace. It will be ready to use with your AI Analyst right away, saving you the work of building the process from scratch.
Keep every process up to date
Manage your Personal and Company Skills from one place. See the slash command, owner, and status for each Skill, then edit, disable, or delete it as your process changes.
How teams put Skills to work
Any reporting or analysis process your team or agency repeats can become a Skill.
- VP of Marketing: Each month, the marketing team reviews channel performance against pipeline goals and prepares an update for leadership. The VP saves the process as a Skill, including which channels to cover, how to compare results with goals and the previous quarter, and how to structure the takeaways. Anyone on the team can run it and produce the same type of update, leaving the VP to focus on where to adjust the budget.
- Agency account manager: An agency turns its proven client reporting process into a Company Skill. Account managers run it for each client, and their AI Analyst applies the same steps and standards to that account’s data. Reports come back consistent regardless of who prepares them, helping new hires produce work to the agency’s standard sooner.
- CEO: The CEO wants every performance review to answer the same questions about revenue progress, changes in key metrics, and areas that need attention. They save those questions and their preferred format as a Skill for whoever prepares the update. Each review arrives concise and ready for discussion, without the CEO repeating the same brief every cycle.
From explaining the work to running it
Skills move your process out of prompts and documents and into your AI Analyst. Once saved, your AI Analyst can apply the same steps and standards every time the task runs, regardless of who runs it. That keeps recurring work consistent while reducing the time your team spends setting it up and reviewing it.
Save one task you repeat often as a Skill, so your AI Analyst can run it whenever the work comes up.
