Pardot Einstein: What We Know So Far (and some speculation!)

By Lucy Mazalon

Last week Salesforce announced Pardot Einstein. This new Pardot technology will enhance Salesforce’s B2B Marketing tool by injecting AI into its core, therefore earning its place in the ‘Einstein’ brand of ‘intelligent’ platform products.

Pardot has stepped up the speed of innovation this year, with new developments such as the Pardot Lightning app that sits within Salesforce, Connected Campaigns, and soon-to-be Connector 2.0. In true Salesforce fashion, the announcement was made only days before Dreamforce – no doubt there will be flashy, forward-looking demos throughout the week-long conference (I will be keeping my eyes peeled for coverage of the Pardot keynote for sure!).

This post will cover the two Pardot Einstein tools and their functionality. Then, I will draw upon past previews and parallel features in other platform products (eg. Sales Cloud, Marketing Cloud), to visualise what form an AI-driven Pardot could take*.

*NOTE: these are my own opinions based on past previews and similar features in other Salesforce products. I might be wrong, but think making feature comparisons is exciting!

Pardot Einstein summary, shown in the B2B Marketing Keynote at Dreamforce ’18.


Einstein Pardot: A Two-prong Fork

Pardot Einstein will work using machine learning, ie. what Pardot (the machine) has learned, based on your Prospects’ all-time engagement and surfaces recommendations and applies indicators to Prospects coming through the marketing funnel. Pardot Einstein takes form as two tools – I have grouped their functionality into ‘Identify’, ‘Optimise’ & ‘Discover’:

Einstein Campaign Insights

  • Identify most engaged audiences eg. by persona, or by geographical region;
  • Identify underperforming campaign assets, eg. emails, landing pages;
  • Optimise email content with recommendations for subject lines and body copy;
  • Optimise audiences for campaigns (by learning from above points);
  • Discover new audiences, predicted to engage well with your campaign content.

Einstein Behavior Scoring

  • Goes far beyond the Pardot Score you’re used to now…
    Identify key touchpoints that have driven past Prospect purchase cycles (using Pardot Engagement History & Opportunity data)
  • Identify a prospect’s probability of converting, eventually into won revenue. Takes into account behaviour patterns beyond their own activity trail.
  • Optimise Prospect score changes – both increases and decay, intelligently (ie. less need for user-created score adjustments)
  • Discover new target Prospects to accelerate through the marketing sales funnel based on behaviour score.
  • Image captured from Pardot Dreamforce Keynote 2018, titled “B2B Marketing on the World’s #1 CRM”

A Little Speculation…The Crystal Ball

A little speculation over what form an AI-driven Pardot could take won’t be any harm. I should emphasise speculation once again – these are my own opinions based on past previews and similar features in other Salesforce products.

Funny, there was something ever-so-familiar about these two Pardot Einstein marketing features…

Predictive Scoring (Marketing Cloud & Sales Cloud)

Salesforce did, in fact, release Einstein Lead Scoring some time ago. Einstein Score was developed for sales reps, to prevent them “sifting through a sea of leads making decisions based on simplistic criteria like…’that one is in our target industry,’ or even ‘that company is big.’”. The score calculations are tailored based on your own business’ lead conversion patterns. Beyond just the numeric score, ‘Einstein Score’ comes as sidebar component on lead record pages that explains why a lead has received their designated score.

Now Predictive Scoring is opened up to marketers, will the Prospect’s score be able to explain itself? I’m hoping so!

Predictive Engagement & Audiences (Marketing Cloud)

The Marketing Cloud ‘Predictive Engagement Dashboard’ is a treasure chest of marketing metrics. On here, marketers find multiple predicted metrics such as: Email open / click prediction, Subscriber retention, web conversion.

Images from Audience Analytics Builder Demo Video.

However, the crown jewel is the ‘Email Engagement Prediction’ component on the dashboard. This categorises your subscriber database (equivalent of Pardot Prospects) into 4 categories: ‘Win-back’, Window Shoppers, Selective Subscribers’, and ‘Loyalists’ – each with a different combination of high/low email opens/clicks.

These findings (and more) are channeled into Audience Builder, to build optimal audiences and find additional ‘lookalike’ subscribers, ie. “people showing multiple predicted behaviors in common”.

Campaign Content Insights

There’s mention on the news stream that Pardot Einstein Campaign Insights will be able to optimise email content with recommendations for subject lines and body copy. Pinpointing potential email content tweaks could go a long way in boosting opens & clicks.

I dug up an image from Dreamforce 2016*, which hints at content insights right within Engagement Studio. It reads: “…unusually low click-thru rate. Try changing the CTA to a demo video. Demo videos increase the click-thru by 12%”.

Image courtesy of Cindy Zhou – a great article, clearly a legend!

Could this be the product that comes to fruition at general release?

*yes, two years ago, an eternity in ‘Salesforce time’!

Also…Einstein Engagement Programs?

Could this be next? When I first heard Pardot Einstein was on its way last week, the first vision that came into my head was predictive routing through Engagement Studio programs.

My initial vision followed something like Marketing Cloud’s ‘Persona Journey Split’, intelligent journey routing of subscribers (ie. Prospects). Below is an image of this Marketing Cloud feature in last year’s Dreamforce Keynote Demo (read this article for more info on the full demos).

Yes, it’s hard not to be jealous of Marketing Cloud marketers sometimes!


Pardot Einstein is an exciting prospect. By joining the Einstein brand, one of the most popular, heavily adopted marketing automation tools in the world is impacting the overall democratisation and accessibility of AI-driven marketing.

I’ve given you a bullet-pointed overview of the 2 Pardot Einstein products – Einstein Campaign Insights & Behaviour Scoring. My curiouslty took me further, to dig around for other Einstein enabled products on the Salesforce platform, such as Predictive Scoring, Predictive Engagement Dashboard & Audiences, and Content Insights.

Pardot Einstein is kicking off in beta to those customers registered on the pilot program. General release is scheduled for February 2019. As we’ve seen with some of the throwbacks from 2016, these products leveraging cutting-edge technology take time to get into users’ hands.

I hope this post has given you a clearer understanding of what Pardot Einstein could do for your marketing, and help you visualise the product within your own Pardot account/app.

The Author

Lucy Mazalon

Lucy is the Operations Director at Salesforce Ben. She is a 10x certified Marketing Champion and founder of The DRIP.

Leave a Reply