Exclusive CMO interview: Where Oracle is heading with AI in marketing
- 28 April, 2017 11:14
It’s official: Oracle is joining the throng of marketing and enterprise technology players jumping aboard the artificial intelligence bandwagon.
The vendor made its foray into the AI realm this week with a selection of AI-driven applications under its CX Cloud, aimed at helping organisations improve marketing, service and sales experiences with customers.
Adaptive Intelligence offers, as the AI app is called for marketers, is designed to optimise cross-channel interactions with recommendations on the most relevant content for an end customer at scale and in real time. Its AI-powered service app, meanwhile, is designed to provide predictive product failure, account health and recommendation capabilities to service agents, again with the aim of improving experiences.
All apps launched can ingest a user’s first-party data sets, as well as third-party audience data housed in the Oracle Data Cloud and sourced from BlueKai.
CMO had the opportunity to sit down with group VP product management, Steve Krause, during Oracle’s Modern Customer Experience Summit to find out what these applications tell us about the vendor’s longer-term approach to marketing and customer experience technology innovation. We also got the opportunity to discuss how AI will continue to impact the modern marketing function, the potential pitfalls, and the wider implications of customer centricity on the role of the marketer.
Q: Oracle’s Adaptive Intelligence offers across the CX cloud suite are the first solid statement from Oracle about its position on AI for marketing and customer experience. What’s the long-term market strategy for AI?
Steve Krause: It’s oddly enough, the second statement for us, in the sense it’s the one the world seemed to want to hear from Oracle. In typical Oracle fashion, we have been doing this for a while but just not doing a lot of PR about it.
For example, before the stuff we announced today in the Marketing Cloud, we applied machine learning and AI-style algorithms to look-a-like modelling in advertising, and we have it in email send-time optimisation, or the use of behavioural targeting based on time. We also have algorithms for extracting sentiment from social postings. But I don’t think it occurred to Oracle to make hay about it because what we were doing was making email sending better, or making advertising more efficient, or giving social marketing more capability, especially around listening.
Other companies have been doing some of that too and rebranding, like Einstein [Salesforce]. What we’ve said is we are in AI, have been in AI, and there are a couple of new tricks in AI that you wanted to hear about, so here they are. The Adaptive Intelligence offers on the marketing side, for example, legitimately are next-generation recommendation engines.
With chatbots, we’re playing in that area because clearly something is happening there. If you give people the tools so they can participate and plug it in to the rest of their marketing stack, that seemed like the right thing to do.
Q: As every vendor starts grouping a whole bunch of things under this term ‘AI’, are you concerned marketers will think AI is just a shiny new tool they plug in to solve all their problems?
SK: I hope not. There are other vendors trying to act that way, and at some point the truth will come out: There is no free lunch here. There are things technology does that looks like magic from the outside, and AI is going to provide some more of those. But from the point of view of a marketer having an ‘AI’ button that they push and all of a sudden, everything works out great, no, that’s not it.
In marketing, more so than other fields, we have creative aspects, strategy, and so on. We love the idea that AI is the next step in automation. Think about your recommendation engine: A human is never going to deal with the scale problem of having 10,000 SKUs and a million customers and making the right connections. But that doesn’t mean the rest of marketing is automatable in the same way.
The people who are creating the creative and beautiful images in catalogues or in advertisements worth looking at – we want to keep doing those things.
Q: Questions are starting to emerge that with the push towards data-driven and technology-led marketing, we lose brand and creative capacities across our teams. Do you see this happening?
SK: The pendulum swings back and forth. I read something the other day that talked about the AI discussions being potentially flawed because we’re talking about it being measureable on a single scale of intelligence, from something stupid to something smart. But there are multiple types of intelligence. There’s the kind of intelligence that can calculate very large numbers.... there’s also intelligence around being creative, and for strategic thinking. All have their level where they can be reasonably automated and where machines can perform tasks reasonably.
When it comes to AI, you almost have to look at them separately rather than as just one big wave where AI will handle everything. Some of these things may move faster, some may not move much at all.
Q: Looking at the capabilities Oracle now has grouped under ‘AI’, how will we see these offerings mature and grow?
SK: There are two answers to this. Firstly, you’ll notice what we’ve been talking about this week is AI applications. There’s AI Offers for marketers, but there’s also an offering for service, and how to let a service rep be smarter by quickly bringing the information they need quickly to bear to be responsive to what a customer needs. We all know that knowledge is often lying around, but getting to it quickly can be very difficult. That’s where augmentation can help.
Oracle’s strategy is that we’re not building one AI ‘thing’. What we’re going to try and do is learn from the past. People who try to build generalised AIs don’t tend to do very well. Even Watson has many different uses – the part of Watson that knows how to play Jeopardy is not going to be very good at drug discovery, for example. We’re at that point of saying where AI is going to work is where the domain can be constrained down, and where we can also build in the other things a marketer needs to get done.
The strategy is to have these core algorithms underneath. By the way, we had a product before any of this stuff called real-time decision, based around adaptive, real-time machine learning. So think of a core, from which you have applications – for marketing, service, statistical process control.
AI is not going to be marketed as the wonderbox. What it’s going to do is make smarter the things you’re already doing, such as making the right offer to the right person, or having shorter customer services calls with better call resolution. Look for those capabilities also coming in other areas where Oracle plays.
The second answer to this question is that just as people in Oracle were solving problems down at the feature level... you’re going to continue to see that in various products where AI is an infusion of those smarts. There are two kinds at least for marketing for AI usage. One is to make the end customer experience better; there’s also the use of the technology to make the marketer’s processes more efficient, smarter, and so on. You’ll see some more of that as well.
Q: As experience and engagement lead business strategy, how much responsibility is it of the marketer to extend their focus into service management?
SK: You’re seeing companies reorienting out of functional silos and situations where marketers are more than just marketers, and a service person is about more than just service. Centres of excellence are being developed, and digital transformation projects come in where if it’s not the organisations themselves, it’s some other authority saying ‘thou shalt work better together on behalf of the customer’.
Historically, you thought you were working on behalf of your org, but it doesn’t really make sense to spend all this money trying to understand who our customers are, then building tech to take that understanding and turn it into a different experience for you versus you. But by the way, I hope you don’t have a service ticket because as soon as you do, it all vaporises.
At the overall enterprise level, it’s the company’s responsibility to make those parts work together. And different companies will find different ways to do it. A key issue is around functional incentives, which are set usually around things right down in the silo.
It’s also a mentality. You can destroy your brand promise and all your advertising and engagement efforts with one bad customer service interaction. When a company decides to really organise around customer thinking, the concept of customer service as a cost has to go out the window.
What Oracle believes, and what is favourable to our business because we have so much breadth, is that just like there’s an obvious need for marketing clouds today because there are too many fragmented systems, there will be an obvious need for taking it up another level and making sales, service, commerce, marketing and customer-facing engagement work together. It makes sense: The customer doesn’t distinguish between the part of a brand that’s service versus a follow-up interaction from marketing.
It’s going to take a long time, because things change slowly when it comes to organisational incentives and traditions. But companies serious about customer experience, especially companies bred to be this way, will force everyone else to change.
Q: As experience becomes a holistic thing between the customer and the brand and their products and services, it raises the question about the relevance of the whole concept of marketing. Do you think ‘marketing’ as a term is a problem now?
SK: There are many different connotations to marketing - the average person sometimes thinks of marketing as the negative thing that’s forced upon them. In our minds, the right way to do marketing is always in response to something. If you are foisting stuff on people, the game is already lost. The whole ambition with doing orchestration and connecting your marketing together is to act based on triggers relating to a customer’s context.
The reality is marketing in the past - and a lot of it today - is still based in a world where the marketer wants to control the timetable, what the customer sees, and how many customers they can get to see it.
We don’t call that modern marketing. We want to create that world where the term marketing means relevance and utility. Jay Baer’s book talks a lot about ‘Youtility’ and there’s a lot in that. If you can redefine marketing as something that most people would say yes to, then we want to build the technology to support that.
- Nadia Cameron travelled to Oracle Modern Customer Experience summit as a guest of Oracle.