This is the subhead for the blog post

Starting September 27 and completing October 17, Google will be rolling out their Exact Match Close Variant Expansion. The aim of this expansion is to help advertisers keep up with the estimated 15% of brand new searches that occur daily. Using machine learning, exact match keywords will now match using the intent of a search, instead of the specific words.

Exact match keywords will show ads on searches that include implied words, paraphrases, and other terms with the same meaning. For example, the keyword [auto insurance] will match with the user query for car insurance. These keywords are clearly different, but the meanings are the same.
You may remember a similar change in March 2017: Exact Match Close Variants, which allowed exact match keywords to show on typos, plurals, and other variants as long as the meaning was similar. Our agency-level analysis of CPA across accounts shows that in a majority of cases, that change was generally neutral to positive. In the few cases where there were negative effects, we were able to control them through targeted negatives.

Is it something worth watching? One hundred percent yes, but we’re not saying it’s cause for alarm. Other automated strategies like dynamic ads and machine learning algorithms work very well when managed properly.

What does this mean for account management? What steps to take?

I recommend closely monitoring search queries to ensure these new queries are helpful and not driving up costs. You will be able to see these variations in the queries report as Exact match (close variants), making it relatively easy to compare performance. 3Q accounts are all built on the Alpha-Beta campaign structure, and we’ll be keeping a close watch on budgets given that this will likely increase traffic to Alpha campaigns. 

If you do see degradation in performance, you can add specific negatives (negatives will not be affected by this change) at the ad group level to better control that matching. But again, the automation that Google has proposed in the past has worked well, when properly managed, and I anticipate that being the case here.
In short, will this increase the need for closer query scrutiny? Yes. Will it make Google more money? Also yes. But we expect the fallout to be relatively minimal, performance-wise, and it may open up new revenue sources along the way.