Salesforce Einstein Marketing Intelligence is a central place where marketers can go to optimize spend or improve campaign performance in order to, ultimately, prove marketing impact and ROI. The main selling point is that it guides marketers in creating an end-to-end campaign in just a few minutes. This brings together multiple ways that Agentforce can be used for marketers, including insights on analytics and automating data preparation.
Remember, Agentforce AI is effective because it’s grounded in your organization’s own data – contrast this to using a consumer AI, like ChatGPT, you won’t get an informative answer (because ChatGPT doesn’t know anything about your audience data and past campaign performance).
Announced at the Dreamforce ‘24 Marketing keynote, Einstein Marketing Intelligence will be generally available in February 2025. For now, let’s recap what we know.
Note: Don’t get Einstein Marketing Intelligence confused with Marketing Cloud Intelligence (formerly Datorama) – the reporting and dashboarding tool that enables marketers to access and leverage their own first-party marketing data.
Intelligence Home
The home tab is the first place we see the centralized place, or ‘bird’s eye view’, for marketers. The overview shows the channels you’re working with (toggle between total revenue generated vs total spend to get a sense of your most fruitful campaigns) and your campaigns (toggle between top performed vs lowest performed).
A Campaign Summary, delivered by Agentforce, gives you an immediate sense of how campaigns are performing against your goals (such as a revenue target), with a breakdown and explanation as to why Agentforce has come to that conclusion.
Aside from summarizing and exploring your data, this prompts you to take actions, such as creating a data pipeline or a goal.
Goals Overview
Agentforce in Einstein Marketing Intelligence works heavily on a goal basis. You can view all goals, active goals, goals not started, and expired goals – as well as the trends over time for each.
KPIs inevitably change over time, so you can add or update a goal. In the example below, we’re editing the revenue target that I have for the quarter (left-hand section). The goal updates in real time, showing that based on this information, it’s on track (right-hand section).
Note that you can set goals on a number of parameters including some or all of the following: channels, certain campaigns, goal value, goal type (e.g. at least), and start and end dates.
Agentforce Campaign Briefs
The Campaign Summary on the Marketing Intelligence home page gives an on-track/off-track status towards your goals. If off-track, the agent provides a paragraph with its recommendations, including “as a result” revenue-generating, business benefits. In the example above, the agent’s recommendation rounds off with “which can drive towards installation appointments”.
Kickstart by using the “Explore more” action or launching Agentforce from the Campaigns dashboard. Generate a campaign brief with Campaign Agent (Agentforce) using natural language prompts (i.e. how humans have conversations). As agents are autonomous, after describing what you want, the agent returns a thorough brief, including a description of the target segment and key messaging. This process is backed up by the Atlas reasoning engine and leverages Data Cloud data to make judgments on segments.
Audience Segments
Time now to create the target segment. There are four options:
- Create with Agentforce (user gives natural language prompts).
- Create List-Based Segment.
- Create from Segment Builder.
- Select an Existing Audience Segment.
The images below show the first option listed above. From natural language prompts, the agent translates your description into the correct segment attributes – a huge time saver, especially for marketers who don’t want to dabble in SQL.
Content (Email)
Having generated the brief and created the campaign record in the background, the agent is now recommending a promotion be offered to customers to make strides towards hitting the revenue target we input as a goal.
In the email builder, a draft email is generated based on the campaign brief. Ask the agent to make changes that I consider pretty sophisticated – such as “turn the second paragraph into a three-column section. Each section should be equally sized with a heading and a one-sentence value prop” – to which the agent creates a layout block and populates it with content.
Summary
Thanks to Kelly Eliyahu, Senior Director of Product Marketing, for a great demo, told through a story that links together how a marketer would actually use Einstein Marketing Intelligence in real life.
Upon reflection, I do see value in Einstein Marketing Intelligence, stitching together the jobs to be done and decisions to make as a marketer throughout the end-to-end campaign process.
As always, who can use it and how much does it cost? We can make two assumptions here: firstly, that this is a feature for those on Marketing Cloud Growth Edition and Advanced Edition, and that this will be priced on a consumption basis (Agentforce is $2 per conversation*). Then, there’s adding yet another analytics offering. There are clearly differences between this and B2BMA (CRM Analytics) or Datorama. There’s a company-wide analytics consolidation that’s in progress.
*What counts as a conversation, I can’t say for certain at this point.
