Persado's Assaf Baciu Is Engineering AI to Understand How You Feel
The co-founder of marketing AI firm Persado on how the company’s growing engineering team stays on top of the latest development tools.
You know the feeling when you delete email after email, or scroll through countless social media ads, from marketing companies whose "personal" messages just don't land and only end up annoying you?
Assaf Baciu knows that feeling, too—and the software company he co-founded, Persado, is working to alleviate it by using AI-optimized language to help marketers reach the right audience with the right message at the right time. Before co-founding the firm in 2012, Baciu worked in product management and enterprise strategy for speech-recognition technology tools (think Siri and Alexa) at BeVocal and Nuance Communications. As Persado's SVP of product and engineering, his goal is to use AI to create truly engaging and resonant communication.
We teach the machine to tell different types of stories, and then the machine learns what stories to tell to which person.
— Assaf Baciu, Co-founder and SVP of Product and Engineering, Persado
“The purpose is to connect emotionally," he explains. Whether it's using excitement or achievement, the language of the email should make a user want to engage with the content—and feel good about the experience once they do.
So far, Persado's approach has proven effective in reaching marketers' audiences, as shown by a nearly 70 percent rise in click-through rates for customers (including American Express and Staples). But improving the AI—which relies heavily on computational linguistics, statistics and machine learning—is an ongoing process, and Baciu works closely with Persado's engineering team through each iteration. We caught up with Baciu to learn more about his team's development process and his vision of the future of AI.
MSF Hub: How do you explain Persado's technology to people who have never heard about it before?
Assaf Baciu: When you're in front of somebody and you talk to them, you learn over time: what makes them laugh, what makes them interested, what they don't understand. You adjust your tone, your vocabulary, your style, and other elements. We taught our platform to do the same using AI.
MSF Hub: What are some of the AI features your engineering team has been working on recently?
AB: We spent the last year developing and applying an application of "narrative intelligence" to our platform, which gives the machine the ability to tell a story, rather than construct some concatenated phrases. If you sell a travel package, for example, there are different narratives that you can tell. Wanderlust is one narrative. A brand could say, "Oh the places you'll go with this wonderful package from X." Or they can talk about prestige, or comfort or rejuvenation. We teach the machine to tell different types of stories, and then the machine learns what stories to tell to which person.
MSF Hub: How do you determine what product area to work on next?
AB: We are in AI, bringing something that is completely new to marketers and learning from them continuously. Our clients are primarily Fortune 500 to Fortune 2,000 customers, so they have a big say in what we develop. But we developed something that some companies don't know exists, and they sometimes don't even know what to ask for.
MSF Hub: What is the development process your team uses?
AB: Our development process is agile—but it’s not two-week sprints. We do between four weeks and six weeks for different applications. We adjusted the process a bit so we're able to bring innovation to market while doing relatively small increments.
MSF Hub: What kinds of challenges do you find your team faces the most often?
AB: Usually the big hurdles come when there are elements that are beyond their control. For example, sometimes we have released something and the API changes, and suddenly our customers suffer because a feature doesn't work properly. I help by bringing the ecosystem partner to the table as fast as possible to help our team resolve it.
MSF Hub: Persado’s evolution seems very much tied to how AI is growing and changing. What kind of advances are you most excited to see in the coming years?
AB: The ability to have a conversation in chat, for example. It's not yet there—even with Google's demo about ordering a haircut, it's still a canned demo. It's not something you can put in market today. But I think that it will be developed in the next five to 10 years. So the ability to really move from what we do today, which is personalized language across your marketing journey, to a full conversation will drive more complete and satisfying interactions.