Lucy could help with research, personality profiling and media optimization. But a spokesperson told TechRepublic that the cognitive assistant is meant to help marketers, not replace them.
Equals 3 managing partner and co-founder Dan Mallin talked with TechRepublic's Tonya Hall about how his company's AI assistant could help marketing executives. Here's their conversation:
Hall: Tell us about Lucy. What does Lucy do?
Mallin: Lucy is kind of the marketing analysis, marketing assistant to help marketers fulfill their dreams, right? There's capabilities in research and finding information capabilities in the audience and in personality profiling. And then, finally, media and media optimization.
Hall: There's all kinds of challenges we have now. Technology is a resource for us. What are the most relevant challenges that marketers face today and how is artificial intelligence helping them?
Mallin: Marketers have, well, we've created our own problem. We have better and better capabilities of collecting information and we collect information in many many systems. A marketer, in theory, should be leveraging data coming out of 20, 30, 40 different places. What that requires though, is access, login capabilities, knowledge training, and the ability to apply all that information.
With the capabilities of machine learning and artificial intelligence, Lucy collects and centralizes the ability to access all of those systems and leverage that content, which in effect, dark data to most people and allowing you make enlightened decisions from your dark data.
Hall: You've said that there are two sides of artificial intelligence in marketing, tell us what you mean by that.
Mallin: Most people think of artificial intelligence as the abilities for that bot or capability to make a decision and execute something. That's certainly part of artificial intelligence. We also focus on augmented intelligence. The idea is quickly gaining access to information but allowing knowledge workers and smart people to spend their time not collecting, and collating, and confirming that research or information, but thinking about it, strategizing about it, and making a better decision from it. It's augmented intelligence. You + Lucy is better than you alone or Lucy alone, and that's where Equals 3 comes from.
SEE: IT leader's guide to the future of artificial intelligence (Tech Pro Research)
Hall: Not everyone, even though it's 2018 and we have technology surrounding us, not everyone is comfortable with technology. How can you make analytics, data-driven decision making, and business intelligence understandable and actionable for even the least technical of employees?
Mallin: One of, again, capabilities in this artificial intelligence space, is machine language. Excuse me, the ability to speak, or talk, or type, ask questions, or interact with data. Not using queries or things, but just the way that you think or talk. Ask me a question like, "Hey, do we have a SWOT analysis on this brand? How did one brand compare?" Those types of conversations that you can have with Lucy, allow anyone to gain access without knowing how to do that or where it's coming from.
The other thing is, we're focused on not finding documents, or creating charts, or doing various things. The idea is to find answers, so if you ask a question, I don't want 100,000 documents to come back, I'd just like a few answers or the answer. Sometimes, the answer is a number, a paragraph, a sentence, or a chart, but it's the answer to my question.
Hall: You work with IBM Watson. In what ways does IBM Watson contribute to Lucy's features and functionality?
Mallin: The way this artificial intelligence space really works is their APIs, application programming interfaces. We use the Watson technology from inside our application, but we call an API and get a piece of information back and it's very geeky and under the hood how that all comes together. There are a dozen different APIs that are in use inside of Lucy and the Watson API stack is a big component of how we have enabled Lucy to do a bunch of functionality, in addition to our own created IP.
Hall: How does Lucy generate predictive media models to forecast optimal allocations and campaign results? I mean, does Lucy actually eliminate the media buyer's job?
Mallin: I think it's anything but eliminating the job. On the same note, you know, if you look at just pure programmatic, I mean yes, there's some job loss there. But, the idea is to take the strategies from the media side of the house, run scenarios, and predict what the outcomes will be, track those outcomes, but also interacting with the media buyer or planner is part of that.
An interesting component, we're doing something right now whereby zip code in the US, where we're comparing media spends, image zip code, gap to leader. Our client to their competitors and then optimizing the spend and predicting the outcomes where they could be, hopefully, the leader in all zip codes. But, underspending in zip codes, they're already the leader in, overspending in zip codes where they need to claim a better outcome, and optimizing exactly what's happening in each zip code.
- Special report: How to implement AI and machine learning (free PDF) (TechRepublic)
- No, AI won't eat your job, say tech chiefs, and here's why (ZDNet)
- Cheat sheet: How to become a data scientist (TechRepublic)
- Five tech jobs that AI and automation will make radically more efficient (ZDNet)
- Demand for AI talent exploding: Here are the 10 most in-demand jobs (TechRepublic)