speak to potential customers and distill what they say into a few problems to solve with statistics.

This is a centrifugal process: spinning around lots of data about what people are buying (supply) before settling on what customers need (demand). Many consumer products are designed this way because the potential customer pool is so large.

AI-First companies put capital, time, and effort behind getting data, ahead of other strategic priorities.

hiring AI-First teams, placing them across the organization, setting up data infrastructure, managing data, and deciding on projects to pursue, but ultimately deferring to business unit managers

appropriate customers from incumbent suppliers by selling a specialized product to a niche segment of customers at a fraction of the cost of the incumbent’s product.

capture the data “exhaust” (anything recorded when customers perform operations in an application, such as clicking buttons and changing values) from existing systems of record (the “single source of truth” for a business function), and build AI on top.

Here’s the formula: data learning effects = economies of scale to data + data processing capabilities + data network effects. Put in a sentence: get lots of data, process it into something useful in terms of making a decision, and create a system that automatically generates more useful data.

random forest algorithm

This table summarizes who to hire and when to hire them. role background cost outsourceable? hiring sequence Data analyst Business Low Yes First Data scientist Statistics Low Partially Second Data engineer Databases Medium Yes Third Machine learning engineer Computer science Medium No Fourth Data product manager Product management Medium No Fifth