AI tech is already changing a lot of professions, not least of which is CRM and lifecycle marketing. Like so many other software companies, the marketing automation platforms that brands use to send email, SMS, and in-app messages are in a heated race to roll out AI tools and features for their users. 

Such a wide variety of options on so many platforms can be more overwhelming than a Cheesecake Factory menu. And similarly, the results can be a mixed bag. So let's break them down, explainer style. (Yes, I’m a sucker for Vox’s “—Explained” articles.) 

Iterable

Iterable has been hard at work rounding out its self-titled "AI Suite" of tools available to its users, and the results have the potential to be highly impactful for marketing teams. Let’s take a look: 

Journey Assist 

‍What it is: Journey Assist is probably our favorite one of Iterable’s out-of-the-box generative AI features. Think of it like a ChatGPT that will literally create and organize a journey map for you.

How to use it:

• Journey Assist is great for creating journey outlines and combatting Blank Page Syndrome. Tell the prompt you want an abandoned cart flow and within seconds you'll get the structure you need to get started.
  
• Journey Assist is great at adding clarifying titles to each journey node, so you know the intention behind each step of the workflow.
  
• Journey Assist also has several preset prompts to help you get started, which means you might not even have to add in a prompt. Now, this mostly reduces Blank Page "scaries," but it is still an extremely valuable generative AI tool within the Iterable platform.
  
  

Things to remember:

• As with all AI features, the input matters most. Writing a descriptive, specific prompt will give you a better result.
  
• Journey assist isn't the best at selecting a trigger event, so you'll still need to understand your custom events and trigger options to set up the flow in the best way..

‍

Channel Optimization 

What it is: Channel Optimization is a node that determines which send channel (Email, SMS, In-app, Push, etc.) would be optimal for a particular message step in a given journey, based on the user's message engagement history.

How to use it: In our opinion, this is a feature most brands miss simply because they're not used to having the option, and we think it not only helps boost engagement, but also reduces the complexity of multichannel Journeys. Use it by dragging the Channel Optimization node from the Journey sidebar into your workflow map.

‍

Next Best Action 

‍What it is: This feature is one of the most exciting use cases we've seen come out of Iterable lately. Something we've experienced as marketers throughout our careers, and as consultants working with teams, is a starting point of action that indicates what the next step should be to improve a message or journey. 

Marketers are often so involved in the day-to-day of their workflows that they have a hard time seeing the "forest through the trees" in terms of their overarching strategic initiatives. 

How to use it: Next Best Action points you (the marketer) in a direction, and you can then decide if you agree or disagree. It's hugely impactful for small lifecycle teams because it empowers them with a simplified decision tree, which allows them to take quicker and more decisive action.

‍

Predictive Goals 

What it is: Modular has been particularly impressed with Iterable's efforts on Predictive Goals. While this feature requires a significant amount of data (minimum 6 months), it can help determine who in your audience is most likely to help you achieve your goals, which makes segmentation that much more methodical and meaningful.

Another really important aspect of this feature is Iterable's willingness to allow the AI to explain how it reached a conclusion about its audience recommendations, which helps streamline and clarify AI audits.

A note: You'll notice this theme throughout the article: AI is only as good as the data it receives. A disorganized marketing environment with confusing, undocumented, or redundant events can be counterproductive to marketing efforts as a whole, and especially for AI. Take the time to reduce technical debt on the marketing team before implementing this feature.

Frequency Optimization 

What it is: This is an important feature from Iterable that gives AI the power to cap (on a personalized basis) the number of email, SMS, and push notifications each user receives, based on past engagement data (i.e. opens and clicks).

How to use it: Let Frequency Optimization choose a personalized cap for the number of email, SMS, and push notifications you send to each user in an Iterable project, so each user gets the right number of messages for each message channel and type.

‍

Send Time Optimization 

What it is: Send Time Optimization (STO)  is another feature that is more "algorithmic" than "AI", but this feature attempts to determine the best time of day to send a message per recipient by analyzing historical engagement data. 

The catch? You need historical data.

How to use it: Modular's recommendation is to first focus on generating that historical data of user engagement, then run several A/B experiments for STO to find statistically significant differences between using the feature vs. scheduling standard sends. 

This feature is also best suited for automated journeys as blast messages are typically more urgent and time sensitive for contacts.

‍

Brand Affinity™ 

What it is: Brand Affinity at Iterable is a score based on a user's open and click engagements in the platform. 

How to use it: Use Brand Affinity to identify actively engaged users (i.e. "loyal" contacts), as well as segment users who could either use a brief message quarantine, or perhaps a different communications approach.

‍

Copy Assist 

What it is: Copy assist is a generative AI feature that gives marketers new ideas for content like subject lines and preheader text. 

How to use it: AI-generated content is typically very general and "sales-y" (i.e. not great), so we recommend using this feature as a starting point, rather than a finished output.

Braze

Braze is another automation powerhouse that has been making strides in AI lately. One thing to note about Braze's AI suite is that, outside of the tools themselves, the platform has excellent documentation on each of its AI features, including potential use cases to make it easier to decide how you want to adopt it.

Intelligence Suite

Intelligent Timing

What it is: Intelligent Timing determines when a user is mostly likely to open or click on a message, then holds the send of a given campaign until that time – making it easier for you to know that you're optimizing for engagement.

How to use it:  Use this feature for sending recurring campaigns that aren’t time sensitive, automating campaigns with users from multiple time zones, or messaging your most engaged users.

Intelligent Channel

What it is: Braze's Intelligent Channel feature automatically sends users messaging through the channel that has the highest likelihood of engagement – for that particular user.

How to use it: Most brands don't use this feature out of old habits. We think it not only helps boost engagement, but also reduces the complexity of multichannel Canvases. We recommend testing this feature in your next Canvas automation.

Intelligent Selection

What it is: Intelligent Selection regularly analyzes the metrics of a given Canvas or campaign experiment that's running and will automatically adjust the audience size for each variant based on user engagement.

How to use it: This is a great way to have more confidence in your testing, and to get results faster than a typical A/B experiment. Best for repeated campaign sends and triggered messages, rather than one-off promotions and announcements.

‍

Predictive Suite

Predictive Events

What it is: "Predictive Events" as a feature name can be a little confusing, but this feature aims to highlight which of your users are most likely to trigger a particular event that you specify ahead of time (i.e. a purchase). Once you set a prediction, users are scored from zero to 100 based on their likelihood of completing that event. 

How to use it: Segment and message users according to your predictive events goals to drive them to their next action.

Again, we need to shout out Braze's documentation on these features, which is full of how-to instructions and use case recommendations.

‍

Predictive Churn

What it is: Similar to Predictive Events, Predictive Churn identifies segments of users who are most likely to churn. 

How to use it: If you have a robust ecosystem but struggle to identify points of churn, this feature is a great addition to your toolstack and can help you hyperfocus on retention.

‍

Generative AI

AI Liquid Assistant

What it is: Braze's AI Liquid Assistant helps you generate personalized message content, logic functions, and more from a simple prompt, even if you're not a Liquid expert.

Templating languages can be difficult to learn, and often have weird quirks. For example, Braze's version of Liquid is slightly different than Shopify's, or Customer.io. Iterable's version of Handlebars varies slightly from the original Git repository. 

Brazes Liquid assistant explains the code it outputs, so it effectively teaches you as it helps you. 

How to use it: One of the most practical uses of generative AI is a bot dedicated to helping marketers write better templating code. In fact, Modular started creating custom GPTs for templating languages like Handlebars (Iterable), Liquid (Braze, Customer.io), and Django (Klaviyo) as soon as custom AI agents were made available.

But Braze is a step ahead of the game by offering one in-house. We can't overstate the importance of this.

‍

AI Content QA

What it is: Braze offers an AI Content QA feature that checks your message content for spelling and grammar errors, offensive language, tone, and more. While offensive language and tone are interesting features, we at Modular think the most valuable checks will be for spelling and grammar, as well as accidental content (i.e. extra code, Liquid errors, etc.). 

We see a lot of value in a feature like this. In fact, we are building our own AI-powered QA tool, called FlowProof, that optimizes last-mile proof checks for email teams –– all from the simple comfort of one Slack channel.

‍

AI Copywriting Assistant

What it is: Braze offers a basic AI-powered copywriting assistant to add content to your messages. This can help reduce the effects of blank page syndrome, and is particularly helpful when working from a pre-designed template instead of a copy-first approach.

How to use it: Our experience at Modular with AI-powered copywriting is that it's about as good as a junior intern –– it puts words on a page that you most likely completely rewrite, but it's a better starting point than zero. 

AI-generated copy also tends to have a particular cadence and "dialect". It's hard to describe, but it's just not quite… human. Use with prudence.

‍

AI Image Generator

What it is: Braze's AI Image Generator taps into the power of OpenAI's DALL-E 3 engine to create images from AI prompts. 

How to use it: at Modular, we have yet to see a realistic and effective use of AI imagery in marketing efforts. As we've noted previously about copywriting, AI-generated creative content tends to have its own "dialect", so-to-speak, which can leave readers and viewers somewhere in the uncanny valley. Use with caution.

Klaviyo

Insights

Personalized Product Feeds 

What it is: One extremely cool feature from Klayviyo's AI is the ability to create a personalized product feed that surfaces product to a user based on what they are most likely to engage with, as determined by AI. 

Many platforms require complex logic, templating code, and maintenance to determine potential "related items" that display to customers and encourage upsells. Klaviyo's feature, however, removes the need for that.

How to use it: Introduce personalized product feeds as an A/B experiment and test their conversion rates against your existing emails, perhaps starting with a high-engagement flow like an abandoned cart series.

‍

Industry-specific benchmarks

What it is: This feature is incredibly useful because it helps marketing professionals understand how their efforts line up against industry standards.

How to use it: Often marketing professionals work in a silo of their own company metrics. Use this feature to compare your metrics to real-time industry benchmarks when analyzing the health and success of your marketing ecosystem.

‍

Generative AI

Flows AI 

What it is: Klaviyo's AI will generate a journey map based on a prompt input. Flows AI is great for creating automation outlines and combatting Blank Page Syndrome. 

How to use it: Tell the prompt you want an abandoned cart flow and within seconds you'll get the structure you need to get started.

Segments AI

What it is: Klaviyo allows users to create a prompt to build a dynamic segment, then its AI will go ahead and build that segment. 

How to use it: Our experience at Modular is that AI models are typically not familiar enough with your data schema to build reliable segments, so use this feature with caution and always double check the output.

‍

Message Content AI

What it is: Utilize AI to create first-draft subject lines, pre-header text, email & SMS content, as well as inbound SMS replies.

How to use it: Again, our experience at Modular with AI-powered copywriting is that it's best for first-drafts –– it puts words on a page that you most likely completely rewrite, but it's a better starting point than zero.

‍

Optimization

Email IP & Domain Warming + Reputation Repair

What it is: At Modular, we handle a lot of IP warmups and platform onboarding engagements. One of the most useful AI tools we see from Klaviyo is a "Guided Warming" feature that allows you to warm up a new domain and IP utilizing AI to analyze user engagement and determine send size, audience, and segmentation so you can take the guesswork out of your migration.

‍

Another hero feature in this suite is Klaviyo's Reputation Repair AI. When your sending reputation is on the outs, this feature will give you a customized repair plan to improve your deliverability.

Customer.io

Customer.io's AI features are still experimental, but you can enable them in your account through Experimental Features (see how here).

Generative AI

What it is: Customer.io's Generative AI feature can be used to generate message content and liquid expressions. As always, a good prompt begets a good response. At Modular, we're excited to see what else Customer.io will bring to the AI table.