Let’s agree not to squabble over the term “artificial intelligence” (AI).

Let’s just go ahead and stipulate that what most people mean by AI is not artificial general intelligence, or AGI, but some combination of ML modeling techniques and rule-based inferencing software. We’ll call this “lite AI.” 

So, sure, let’s concede that lite AI is a long way from AGI, but let’s also take a moment to appreciate what lite AI is actually capable of. Believe it or not, lite AI is producing outcomes that seem … if not miraculous, then (from the points of view of the people who stand to benefit from them) kind of close to breakthroughs.

Look at what lite AI is doing for sales and marketing – or, more precisely, at how lite AI is transforming a number of common digital use cases in sales and marketing. Whether it’s helping professionals to more effectively target consumers across different channels, or giving them tools they can use to refine (and empirically test) different kinds of messages in their outreach to consumers – so what if this isn’t AGI? AI by any other name is still just as neat.

There’s even an optimism – naïve or otherwise – that lite AI will transform the creative aspect of marketing, perhaps reducing the frequency (if not the intensity) of late-night stimulant-fueled hair-denuding whiteboard-smashing brainstorming sessions.

“Creative … has been this mystery black box. It has been people in a room being like, ‘I have an idea. Maybe this will work,’” observed Lisa Spira, head of content intelligence with Persado, during a recent episode of DM Radio, a weekly data management-themed radio program. 

With ML-infused sales and marketing, Spira argued, the creative process is less of a shot in the dark: not a science, per se, but (as in something like linear perspective in painting), the application of science to art. “There’s just so much opportunity in changing creative to learn what … large audiences respond to. It’s a space that hasn’t been explored yet,” she said.

Words and things

Poets and marketers are alike in at least one respect: both know that phrasing matters. 

Phrasing is Persado’s métier. At a high level, Spira explained, Persado’s software helps marketers aptly and optimally to refine their messaging. However, she stressed, Persado is less interested in enabling marketers to develop general-purpose messaging than to craft targeted messages: i.e., in saying the right things in the right way at the right time to the right people. 

“It’s gonna tell you which combinations of words work … for these different people at different times and we’re able to explain language and break it down into components and interestingly, not necessarily [just] the components,” Spira told DM Radio host Eric Kavanagh. “We have our own ways of breaking down language … that shows you what forms, what type of emotion you put behind language, what types of ways you write certain text, [how to] format things and how you can highlight different pieces and test all of that in an environment that will show you … what moves the needle.”

Not surprisingly, Persado’s software has a lite AI god-from-the-machine at its core. But Spira, who said that she has a background in linguistics, assured Kavanagh that this is less the stuff of magic than the rigorous application of neurolinguistic theory to different kinds of problems. In sum, she said, if you can measure what it is you want to improve, Persado’s software can help improve what you’re measuring.

“When we start out working with someone, you tell us what you’re able and interested in measuring. If that’s email-opens or that’s clicks or that’s something else – like watching a video or downloading, or all the way through to a conversion to a purchase,” she said. “If you’re able to measure that, that gets fed back from you into our AI, so that [we] can understand how well individual words and … segments of language did at contributing to that, and then it can continue to learn and refine.”

Who are you? (Who, who? Who, who?)

In a sense, each of us already has a globally unique identifier (GUID) – i.e., our DNA.

Alas, marketing professionals are not yet permitted to identify people on this basis.

This is to emphasize the noncontroversial point that marketing has struggled for a long time with the problem of identifying customers. It has had difficulty linking customers across different channels, or resolving customer identities in the aftermath of address changes, name changes, and other interposing vicissitudes. Honestly? It still has trouble de-duplicating customer database records.

According to David Henry, director of product marketing with Amperity, a company that develops a B2C-oriented customer data platform, lite AI – that is, the combination of ML modeling techniques and rule-based inferencing software – has the potential to change this: big time.

“We have multiple patents for our AI-based identity resolution and it basically does what a human would do,” said Henry, who guested on the same DM Radio episode, explaining that AmpID, Amperity’s core identity management product, functions as a kind of ML-powered replacement for human reasoning. “If … one day [a customer] put[s] in 123 Zenobia St. and another day [they] put in 1234 Zenobia Street, we can kind of discern with our AI that that’s most likely the same David that we’re talking about … and then extrapolate that across hundreds of millions of customer records.”

From there, Henry told Kavanagh, organizations can more effectively target and fine-tune their approaches. He positioned Amperity’s AmpIQ customer intelligence product as a complement to AmpID. Basically, AmpIQ functions as a kind of lab for testing messaging, engagement, and activation. 

“Let’s say you know [that] David Henry responds best to email marketing, [and] he’s opted in to email and SMS. But we have some functionality to run A/B tests … [because] we’re not going to assume that what David’s done in the past is what he’s going to do for perpetuity,” he explained. “But we do have a variety of machine learning models that are built into the platform that can predict things like, what’s the right channel for this audience? Where are they most likely to make their next purchase, whether that’s in a store or online. What products or categories are they most interested in?”

It might seem table-stakes, but identity management is at once a foundational technology and a kind of force multiplier: if a company is able consistently to link customers across different channels (in spite of interposing vicissitudes), it has the potential to fine-tune both the context and the content of its outreach to these customers. 

The search ain’t over. Far from it.

There’s a perception that search-engine optimization (SEO) is a stagnant space and that SEO itself has been mastered. Maybe. It might also depend on how one looks at it, however. After all, for better or worse, Google is still the 800-pound gorilla of Internet search. Moreover, as DM Radio’s Kavanagh pointed out, Google experiments with its SEO algorithms on a daily or weekly basis. What’s frustrating is that this experimentation is neither uniform nor predictable: Google might tweak its search algorithm in just one region, for example. If it likes what it sees, it might test these changes in other regions before incorporating them into its search algorithm. One way or another, marketers need to stay on top of what Google is doing.

“As Google changes their algorithm, you need to change and update your site,” stressed Christina Mautz, a senior vice president with Moz Inc., a company that develops SEO-oriented products and services. Mautz drew a connection between Moz and Persado, which aims to optimize content for different kinds of consumers: Moz helps companies optimize their SEO (with an emphasis on keywords) for different kinds of search engines. “You also need to think about optimizing your content for search engines, because it’s that content that gives you … the authority of your website that is measured by domain authority,” she argued. Mautz told Kavanagh that a “domain authority” score is Moz’s way of measuring how highly a website is likely to rank in a search engine’s results page.

Yes, keywords are an obvious focus of SEO optimization, she conceded, but not in the way you’d think. “You really want to be watching [for]… are you getting the conversions you’d expect. If you’re not, then you might get people to your site, but then they’ll bounce and you’ll see … they’re not converting. That means your content is not performing well. Google will pick up on that, [so] you’ll rank lower,” She said. 

In other words, the SEO problem is one of maximizing both rankings and relevancy. On the one hand, you’ve got to use the right keywords so your company secures a top-level placement in Google’s and Bing’s page results, driving traffic to your web site. On the other hand, you’ve got to ensure that the products and services your company offers are actually relevant for the people Google or Bing direct to your web site. 

Once again, the onus is on sales and marketing professionals to refine their keywords to better target the right users at the right time. Sound familiar? 

“You want to look at your site to check … the referral traffic you’re getting, how many backlinks and the domain authority of those backlinks,” Mautz told Kavanagh. 

Not surprisingly, she said, Moz uses (lite) AI to help its customers with this and other SEO challenges.