IDC: Human creativity will be vital to how marketers successfully use AI
- 11 March, 2019 11:44
IDC research analyst, Gerry Murray
Human creativity will be key to avoiding the diminishing returns of technology and data as artificial intelligence (AI) and machine learning become an everyday part of our marketing lives.
That’s the view of IDC research director, Gerry Murray, who took to the stage as the keynote of CMO and CIO’s recent Executive Connections events in Melbourne and Sydney to share how AI use cases are starting to proliferate across the marketing and customer engagement sphere. Importantly, he also discussed what marketing and technology leaders should be doing to realise their potential.
As Murray pointed out, AI is going into every single marketing tool out there today, and he noted more than 80 use cases identified by IDC as part of recent research. These stretch from virtual sales reps and social sentiment analysis, through to lead scoring, AI-powered content marketing, chatbots, recommendation engines and attribution analysis.
Some of this is being baked into existing martech systems and platforms, while other AI engines will be developed on a case-by-case basis.
“I go back to the analogy of sport – it’s a lot of different things, from swimming to skydiving, American football, surfing and Aussie Rules,” Murray told attendees. “Each of these sportspeople have natural talent; they’ve also invested an enormous amount in developing highly specialised skillsets to compete at the very highest level of their sport. You wouldn’t mix them around either and expect the same level of performance.
“The same is true of AI engines – they’re generally specific to a particular purpose.”
IDC’s own official definition of AI and ML is: “A set of technologies uses natural language processing, machine learning, knowledge graphs, video/image/speech analytics and other technologies to answer questions, discover insights and provide predictions or recommendations. Applications using these technologies hypothesize and formulate possible answers based on available evidence, can be trained through the ingestion of vast amounts of data/content, and automatically adapt and learn from their mistakes and failures.”
At present, AI is best used for task-level activity, Murray continued, such as email subject line optimisation, content recommendation engine and sales pricing solution.
“Today there is something for everyone, and a lot of things that marketing automation and AI/ML use, down to stack level. But there are also capabilities assisting with planning, budgeting, competitive intelligence and achieving insight for the CMO,” he said. “The key message here is your marketing mission drives your AI use case.
“It starts with the team, determining the use case, which drives the data required to support the decisions. That is what determines the type of engine you want to apply to the data.”
On top of this, teams must have solid metrics in play to compare AI performance against traditional activities.
“Lastly, you need to ensure all the data requirements are available to make the engine run,” Murray said. “The data could be sitting outside marketing, or you don’t have access to it, or there’s a lack of confidence in the data sets – CRM might not have the best reputation for example.”
Much like how a restaurant operates, Murry said AI is a cross-functional capability, requiring a diverse set of perishable ingredients coming to a team in the kitchen who boast of specialist skills and tools that can turn ingredients into a consumable product. You then need an AI ‘waiter’ playing the critical role between what the customer wants as an experience, and how the restaurant turns the ingredients into something that customer will enjoy, he said.
“There are all these other roles that connect the data-driven marketer to the data set and engineering that will do the back-end world. That key role is the business analyst,” Murray continued. “They are the ones talking to the marketer with goals, datasets and decision criteria who make sure those are built into the data set and that the science team brings it all together.”
And it’s this ongoing need for human intervention and creativity Murray positioned as vital to making AI a competitive advantage for your marketing organisation long term.
“Machines are going to become more and more capable of doing task-level stuff, and certain roles will find a lot of things they do are going to be automated. So what to do with that extra capacity for the marketing team? That’s where this pays off in the end,” he claimed.
“You want to start with tasks, but you also don’t want to limit yourself to using technology to do the same-old marketing better, faster and cheaper, thereby providing the same level of customer service.
“You have to be able to think differently about these problems.”
To make the case, Murray noted how the commodisation of marketing automation systems in recent years has taken away what was once a technology-level competitive advantage for early adopters.
“The same is true of data analytics – within the one industry, there isn’t a lot of differentiation between the data one bank has on its customers, and another bank,” he said. “Eventually, the competitive advantage diminishes over time because everyone knows everybody. So what do you do when everyone you compete against has the same capabilities technically?
“The one thing I don’t think is subject to these diminishing returns is the idea human creativity. That’s what drives the ability to start working with data in new ways.”
What this means is creativity is going to take on a brand new meaning for marketers, IT and the rest of the organisation in the face of AI-powered change, Murray said.
“All of these teams have start to facilitate and assimilate this concept and new culture in their organisations, which will require all of us to be smarter with using data in new ways to do things for customers,” he said.
To get there, Murray stressed how important it is to break down silos between functions, data sets and teams.
“Just within the marketing team this is tough… The social marketer doesn’t care about how to feed 100 people at an event, and the events person doesn’t care how to tweet 500 times per month,” he said. “But what’s important is how they each take the identity of the customers they both serve. What kind of data and attributes do they have and know about those customers, and gather in their interactions?
“If the events person shares the sessions each individual goes to with the social marketer, it’s a much stronger indication of interest than whatever a person happens to be tweeting about. If the social person can monitor what’s happening at the event, see whose posts get more retweets or have more influence with audiences, it’s a good set of information events people can use to incent and reward folks, as well as intervene and help people not having a good time to improve their experience.”
Murray said this kind of data give-and-take needs to go right through the organisation. “What about customers who haven’t paid? Perhaps I could do some analysis and look-a-like modelling in my marketing campaigns so I don’t bring in more of those types of customers?” he asked.
“I encourage you to start having these conversations around your organisations. Once you start having these conversations about data, you will find an enormous amount of creative potential you can unlock in your organisations.”
IDC’s top 10 current use cases for marketing and key business motivators
- Virtual sales rep - for those unable to follow up on 100 per cent of leads
- Social sentiment analysis – ideal for coping with large volumes of social traffic, powerful external influencers
- Lead scoring – for those with lots of leads and types, sales reps and partners
- AI-powered content marketing/dynamic content – to assist with large content inventory
- Attribution analysis – ideal for omni-channel marketing in complex, long cycle customer acquisition, or where there is large ad omni-channel spend
- Competitive intelligence – to help with keeping up with dynamic and disruptive competitive environment
- Chatbots – for very high Web traffic volumes and/or online support requests
- Cross-selling/upselling – for those with large customer bases with many product and service opportunities
- AI-powered merchandising – assisting with complex offers and inventory management
- AI-powered recommendation engines – to cope with complex buyer’s journeys and content matching, delays in decision making