AI Excels at Transformation Verbs
~630 Words | ~2.5min Read
I recently shared a mental model for thinking about systemic improvement with AI. AI is an adapter of time from one knowledge domain to another. It’s a kind of alchemy. You can now transmute your specialty time into entry-level output in any other domain. But here’s the trick: you have to know where to apply it. And that starts with recognizing the verbs AI is good at.
The key question to ask about any task in your process: “Is this a data transformation activity?” AI excels at transformation tasks—changing data from one form to another. Translation. Condensing. Reformatting. These are some of the verbs that reveal where AI can help.
Take meeting notes as an example. It’s a quintessential transformation task! Take notes during the meeting, send them out in an email. If you can get access to the transcript from the meeting, why spend human time pulling things out? Send the transcript through an AI to pull out action items. Then confirm them against your own understanding. We still need the human in the loop. With confirmed action items, recasting those into an email is another transformation task! You’ve transformed data (the transcript) into information (action items in context)
Or consider spinning up on a new project. You need to quickly gather, interpret, and embed a field of new knowledge quickly. That’s a condensing task! You are taking information and transforming it by saying the same thing, but shorter. And you do this by building on context you already have to connect to the things you don’t! Perplexity is an AI-search engine, and excels as a research tool. “Give me a quick base-level understanding of this topic.” Tools like Perplexity have you covered. Or you can just write your own research workflow!
But not all transformations are about words. Consider Graphs in Excel. They represent the same data (tables and cells) in a contextualize way (visualize graph). The Copilot Analyst agent is built for that. Not only graphing but light programmatic data analysis. Want to find a trend, and then break it down by market? Ask, then ask again, refine and adjust. AI puts the abilities of an entry-level data-analyst in your hands!
But as Drucker explained something important in the Effective Executive. Every knowledge worker must know something of who will consume his output to make it useful! “The task is not to breed generalists. It is to enable a specialist to make himself and his specialty effective. This means he must think through who is to use his output.” The knowledge worker specialist needs a service mindset—thinking about who consumes your output.
This is where your process improvement skills become critical. You need to start by mapping your processes and identify inefficiencies. Where are you converting one form of data into another? Where are you condensing information? Where are you reformatting outputs for different consumers? Those are opportunities.
Theory of Constraints teaches us that every process has one real constraint. Improvements anywhere else won’t improve throughput. But within each individual process, you can improve the bottleneck within that process. 1% better everyday can double your flow in 40 days! Just don’t expect it to revolutionize things just yet. It’ll just reveal the next bottleneck.
The new core skill isn’t coding or analysis. It’s clearly articulating your intent. But to do that, you first have to identify the places where AI excels. Know what AI is good at, e.g. transformation tasks, and then find those in your process.
Start by mapping one process you work with daily. Identify the places where data changes form. Ask: “What is that transformation?” Then try to leverage AI with one of those tasks. Confirm the output. Adjust your description. You’ll quickly learn where AI fits and where it doesn’t. And you’ll accelerate along the way!