4 ways AI radically changed how marketers do their jobs

We talked to AI-savvy marketers at the Bank of Montreal (BMO) about a few of the most profound shifts. Apologies in advance for the PTSD.
 

 
Then: A deluge of data
Will Duong, senior manager of BMO’s leads management engine and data strategy, has been with the bank for nine years – long enough to remember when customer data collection was like trying to keep grains of sand from slipping between his fingers. The fact that BMO was collecting data from scores of channels made things especially tough. Duong found himself manually collecting information from a vast array of data streams, dropping it into multiple spreadsheets and struggling to merge it all before using the insights to design campaigns.
 
Now: User-friendly insights
A process that used to take weeks or months is now instantaneous. Entering a few key inputs related to demographics or user behavior can generate detailed suggestions around who to target, at what time and on which platform. Resulting campaigns can then be deployed and coordinated in real time across channels as varied as email, web, SMS, social media or mobile apps.
 

 
“We have about 400 always-on campaigns running,” said Duong. “I don’t have a production team creating files, picking them up, manually processing them, loading them to another server and monitoring them for each campaign execution.” Instead, IBM Watson Marketing handles all the touchpoints. “We can just let it run and then optimize.”

Then: One for all
For lots of marketers, personalized targeting was something of a white whale. They did have customer data – mountains of it, actually, ranging from demographics to browsing habits and spending patterns. But parsing all of it to identify fleshed-out, specific customers required a Herculean manual effort. For many, the only realistic solution was to craft broad messaging that appealed to anyone and everyone, then blast those messages across every conceivable platform.

Now: To each their own
AI engines have stepped up to take in and synthesize data across channels and devices – customer signals, behavior, profiles, environments, contexts and other variables. Armed with a deeper understanding of their consumers, marketers can now determine their ideal targets and tailor personalized messaging that plays to customers’ interests, locations and much more.
 

 
“We’re triggering transaction-based offers in real time and in a specific context,” said Duong. “We’ve moved from traditional direct marketing channels like direct mail and telemarketing, which don’t deliver a response rate that’s worth the cost, towards real-time interaction – leveraging the moments when customers interact with us across channels and talking about what matters to them.”

Then: Isolated touchpoints
Over the years, as more platforms for engagement emerged, marketers responded by integrating new channels – one per year in the case of BMO. While this rewarded them with more customer data, each platform largely functioned as its own island. The result: Marketers knew a lot about how customers interacted with platforms in isolation, but very little about how those platforms interacted or could be sequenced.

Now: Mapping the entire path
For the first time, marketers can draw a coherent map between those islands, identifying customers’ affinities, favorite channels and online behaviors. For a company like BMO, that translates to the ability to optimize their messaging and offer strategies across each channel.

“We’ve done the work of integrating all the different channels,” said Duong. “AI enables us to optimize what we’re doing within them. Being able to stitch the journey from physical to digital channels and demonstrate benefits of each allows us to show the rest of the bank how to monetize channel engagement and data analytics.”

Then: Waiting for insights
Think about the sheer number of steps involved when segmenting customer databases: exporting gigabytes of data from multiple silos into spreadsheets; creating local copies; generating analytics; creating segments; and finally re-uploading the data. It took days before marketers were even able to compare their company’s data with industry benchmarks.

“We’d been using monthly tech models running on a certain cadence, and by the time they got to the customer, they might be 30 days stale,” said Duong. “The customer could easily, based on their spending pattern or behavior, have switched into a totally different segment.”

And on the UX side, customers were bailing at the first sign of trouble with marketers none the wiser.
 

 
Now: Real-time analysis, immediate action
Today, you can type in a simple description into an AI-powered system – “males, 28 to 34, income greater than $75K, east coast United States, who prefer high fashion” – and boom, your segment is available on-demand. This immediately usable, useful data is changing the game. “Taking in real-time intents across all the different components allows us to better target and optimize our interactions,” said Duong. “And we’ve seen a three-x lift in response rates when we tie in digital intent.”

AI can also observe where a consumer is getting snagged on poor UX, troubleshooting it in real time or alerting a human marketer to quickly take action. Those more pleasant experiences lead customers to stick around.

So there’s little doubt that AI is making marketers’ lives easier and bottom lines blacker, and these early returns are just that: We’ve only scratched the surface of the potential benefits an AI-powered industry offers. It won’t be long before everything “now” is “then” again.