Forum Discussion
How to get started with Braze AI
My company is a B2B banking platform, and our current product marketing approach follows a traditional deterministic model, relying on 'if-then' logic. However, we want to move beyond static flowcharts and drip campaigns to a more dynamic, data-driven next-best action model.
I’d love to hear recommendations on where to start. What are the key steps to making this transition successful? And how can I best equip my data team with the insights and resources they need to build an effective system?
2 Replies
Would add here:
- That Braze is set to launch its B2B account / opportunity data model rather soon based on my understanding, which will be a piece of the puzzle.
- In terms of "insights" you look to acquire assuming you're collaborating with a sales team - again a piece of the puzzle in B2B - is you need a data point highlighting when you'll be involving a sales rep into a conversation as a next-best action. Essentially multi-dimensional lead and account scoring; our Braze integration here at SalesWings is purpose-built for this. An "if-then" approach is essentially all you need. This could be leveraged to trigger an email alert to sales from within Braze, or you can manage a hand-off inside your CRM. If you do continue your lead nurture track you can tailor message content and content cards based on buyer-decision-stage and favor more bottom-of-the-funnel content for buyers that are more ready to purchase or are ready to expand.
We use Braze here at Braze to market to our customers, so here are a some things we've learned that could help!
- Personalized Paths for next best use case, feature, etc ----> This is a part of the Braze Experiment Step, and will flow users to the correct next step based based on how a first random sample of users responds to the sets of messages. From there, it determines what messages are resonating best with each user, and redirects future users like that profile to the correct step. The challenge with B2B is of course that the Canvas is user-based vs. account based, but it could be an interesting proof of concept to run!
- Depending on how many products, actions, etc. you are trying to influence, a Catalog + AI Item Recommendation could be powerful. We tested this out with some use case recommendations and saw some success. If you don't already use Catalogs, investigating those would be the first step.
- When we're running new concepts by our engineering counterparts who we'll need support from, we generally go through a campaign planning exercise as if we were in onboarding. So what is our idea, what data would we need to make this idea possible, and what data would we need to measure the success of the campaign. We can then work with them on determining levels of effort and timing.
