Salesforce has been rapid to roll-out multiple generative AI capabilities that enhance their existing technologies. We first saw Einstein GPT (with marketing use cases that lacked details), then at Connections ‘23, Marketing GPT was showcased (since then, the ‘GPT’ has been dropped). Finally, at Dreamforce ‘23, the Einstein 1 platform came onto the scene, which is essentially Einstein GPT on steroids – with Copilot and the Studios for admins, plus Data Cloud underpinning it all.
So, with the context of where we’ve come from, now join me into the depths of my imagination. These are a number of interesting ways that generative AI, powered by the Einstein 1 Platform, could enhance our usage of Account Engagement (formerly Pardot).
Note: This is purely speculative, I have no ‘insider information’; I’m just considering the ways the wave of generative AI, and more specifically the Einstein 1 Platform, could help workflows.
1. Account Engagement Optimizer
The Account Engagement Optimizer, which became generally available in the Summer ‘23 release, gives you the big-picture view for each of your business units. If items need your attention, they’ll be flagged, plus you’ll be presented with proactive recommendations to take action quickly.
As of the Winter ‘24 release, you can now see prospect data changes in a ‘histogram’ – a graph that shows trends on specific feature areas, over time, by business unit.
Wouldn’t it be great if we could have an injection of predictive insights into this?
Not only how your account is performing now/has been recently, but also running “what-if” scenarios (e.g. connector sync capacity if you were to add more marketing assets or prospect records).
2. Sync Error Diagnosis
Sync errors between Account Engagement and Salesforce can occur for a number of reasons, for example, your Salesforce Admin changes a picklist field’s allowed values and the Account Engagement field is using the old value/s.
In organizations where the Salesforce side is in flux, with multiple configuration changes ongoing, you could log on in the morning to find thousands of errors.
Not only could we be alerted to sync errors and have AI analyze the root cause, imagine if AI could actually fix or recommend fixes?
3. Segmentation (Dynamic Lists)
Could be a bugbear of mine only, but how time consuming can segmentation lists be to put together? Not only do you have to type out the values you want included as criteria, figuring out matching rules can be time-consuming for some users – especially trying to troubleshoot why you don’t have the results you’re expecting.
Could conversational prompts allow us to tell Account Engagement what segment we want the list to pull? It would be a huge time saver.
4. Operational Engagement Studios
Aside from the obvious use cases (i.e. those for marketing automation), there are a number of other use cases that admins can take advantage of Engagement Studio for.
Have you ever considered using Engagement Studio for prospect data management? We call these use cases ‘operational’ because they assist with the heavy lifting in the back-end.
Imagine if we could take this one step further. Again, could conversational prompts allow us to tell Account Engagement what we need to do, or, using insights from other places – like Optimizer or sync error diagnosis – and it generates a program for us?
5. Testing Marketing Assets
How do you know that the assets and automation in your Account Engagement instance are working correctly, and working together? This is where testing for your tech stack comes in, including forms, ebooks, your preference center, and more.
There are different types of testing, and test automation is the category of technology that goes systematically through each possible scenario (if this happens, then that happens), to the extent that no human would feasibly be able to do. An injection of AI here would make the process more robust.
6. B2BMA Guidance/Report Performance
Starting out with B2B Marketing Analytics involves a learning curve, with new graph/chart concepts, and working with data sets.
There are different ways that AI could be blended into this analytics tool:
- Guidance on which visualization types to use, what datasets would be suitable, and which you may need to create. This could also include a report analyzer, which would be helpful to keep the datasets, lenses (etc.) optimal for the benefit of loading reports faster.
- Perform data exploration more effectively, in natural language – i.e. the way we ask questions when we speak. Take Tableau Pulse, as an example – the reimagined Tableau user experience which enables users to surface insights in natural language, by simply asking questions within the console. You can either take guidance from Tableau’s prompts or ask your own questions.
Bonus: Predictive Data Import (Analyze My File)
In many organizations, the marketing team plays a role in importing data into Salesforce/Account Engagement. They are either proficient overall, or trained up for certain import use cases – all to decrease their reliance on their Salesforce Admin.
I’m sure many of you will be familiar with the nature of data imports. you won’t often get it right and error-free on the first attempt. Wouldn’t it be (truly) awesome to have an analyzer that would look at the file, compare it to your org’s configuration, and detect potential errors before you make the first attempt? This would be a great productivity boost, for example:
- INVALID_OR_NULL_FOR_RESTRICTED_PICKLIST: You’re aware that picklists are intended to maintain data quality, and a restricted picklist means that you cannot use any value that is not exactly the same as the option listed in the picklist. This also reflects back to the sync error diagnosis (point #2), as these values can often be changed without considering that the values in Account Engagement need re-syncing. The point is, you might see these previous values, and be none the wiser until you attempt an import into Salesforce.
- Picklist Syntax: If you outline multiple values for a picklist or multi-select picklist using a comma (“,”), and you attempt to import into Account Engagement, you will receive an error. On the other hand, use a semicolon (“;”) separator with Salesforce, and you will also be unsuccessful. With different syntax, it’s easy to get caught out (I’ve been there when manipulating data on auto-pilot). It seems like a simple check for formatting!
- Import Template Generation: Imagine if marketers gave the requirements for what they want to achieve – what the campaign/marketing activity is, what data points they plan to collect – and the system produces an import template that adheres to how their org is set up. This would be next-level! It could even highlight the changes they, or their admin, would need to make to accommodate a successful import. Too often, what some people would think of as ‘small’ changes, in reality have greater impacts on the org – I think of this as an ‘iceberg request’, that what you believe from your point of view is a small part of the whole.
Summary
We’ve come a long way over the past year, with Salesforce releasing multiple generative AI capabilities, and a re-architected platform to support it all, at speed.
There are a number of interesting ways that generative AI, powered by the Einstein 1 Platform, could enhance Account Engagement (formerly Pardot) – which is when I started imagining what I would find useful.
Now that you’ve joined me into the depths of my imagination – what do you think? Is there anything that you would love to see in the future?
