Nand Kishor is the Product Manager of House of Bots. After finishing his studies in computer science, he ideated & re-launched Real Estate Business Intelligence Tool, where he created one of the leading Business Intelligence Tool for property price analysis in 2012. He also writes, research and sharing knowledge about Artificial Intelligence (AI), Machine Learning (ML), Data Science, Big Data, Python Language etc... ...Full Bio
Nand Kishor is the Product Manager of House of Bots. After finishing his studies in computer science, he ideated & re-launched Real Estate Business Intelligence Tool, where he created one of the leading Business Intelligence Tool for property price analysis in 2012. He also writes, research and sharing knowledge about Artificial Intelligence (AI), Machine Learning (ML), Data Science, Big Data, Python Language etc...
Data science is the big draw in business schools
710 days ago
7 Effective Methods for Fitting a Liner
720 days ago
3 Thoughts on Why Deep Learning Works So Well
720 days ago
3 million at risk from the rise of robots
720 days ago
Top 10 Hot Artificial Intelligence (AI) Technologies
Microsoft's CMO: AI Will Make Marketing Less Manual
Marketing may be a trillion-dollar industry, but its supply chain needs a makeover, says Microsoft's CMO.
"What makes marketing such a unique case is that the labor component is much higher than any other industry," said CMO Grad Conn. "I think there's a lot of opportunity to optimize the way we're working in marketing and using AI to help us make better decisions."
Microsoft's recent partnerships underscore Conn's view.
In late March, for example, it teamed up with Publicis Groupe to migrate SapientRazorfish's data products to Azure, Microsoft's cloud platform. And Adobe announced it would standardize its marketing, creative and document cloud on Microsoft Azure while the companies collaborate on artificial intelligence and analytics.
Conn spoke with AdExchanger about AI's implications for marketers and advertisers.
AdExchanger: How can marketing automation improve?
GRAD CONN: Because we sell so many products, the interesting challenge for us is that our customers are often in different buying states at a single point in time.
The typical marketing system has this point of view that everyone is a new customer and doesn't consider their multiple states. The real challenge is the way marketing is set up today apart from customer care, which is kind of crazy.
Customer care is a repository of your very best, most loyal customers and your most valuable source of future business.
Other than better organizational alignment, what can be done about it?
The technology systems are not well constructed to handle marketing and customer care together. Customer service as a software category is quite mature. Some of the systems have been around since the 1980s. It's so mature it probably needs to be reinvented because some of the systems are so old.
Whereas in marketing automation, which is more acquisition focused, it's relatively new. Any software category that's less than 15 years old is a pretty fresh one.
I don't necessarily think of marketing automation as a new technology category.
Ironically, traditional marketing automation systems are almost out of date because they're very email-centric. And email is a very important marketing channel, but social is so critical and not all systems are equipped to manage social handles.
And then [in] advertising automation, which is a super new category, there's hundreds of tools in that area. Our new Adobe relationship is a pretty cool step to get more unification between what Adobe's doing and what we have in Microsoft Dynamics, which is sales automation, ERP and customer care.
Where everyone is going is trying to get away from complicated marketing stacks to a more simplified stack with a common customer ID. There is no real standard for a common individual ID, but you could argue Facebook has potential. We've got to get a lot smarter about the applications we use to manage the systems around a common ID.
Where does your acquisition of LinkedIn play a role?
The LinkedIn professional graph, I've heard, is more accurate than most company's org charts because people have the tendency to update their LinkedIn profile before they do anything else when they change jobs. We also have an Office 365 graph, which is part of a suite and allows you to see your interactions with co-workers and use as a tool to optimize your performance.
We're just at the beginning stages with LinkedIn, but we'll have some pretty exciting announcements this year around the possibilities for a combination of the professional and office graph.
What's Microsoft's position on AI in a sea of corporate competitors like IBM, Google, Salesforce?
It really is so new, and there really is so much opportunity, that it's kind of exciting to see people innovating. I think we're at the beginning of what will be a 20-year journey. I don't see this as competition. The key part of a lot of this stuff is natural language processing and voice recognition. Back in the early '90s, the voice recognition error rate hovered in the 90% range.
The whole field stagnated for a decade, and about two years ago, a bunch of Microsoft researchers decided to revive an older technology called neural networks and put it into our global Azure platform, which is a giant global computer now with virtually unlimited computing capability. They started to train it and, in October, it was the first time they got the machine error rate below the human error rate.
How do those insights translate into your business applications?
Around the holidays, we launched something called Skype Translator, which, leveraging this technology, allows a near-real-time translation of languages. And these are the things that will really start to unlock AI.
In AI, the challenge in a lot of traditional systems like sales automation systems or CRM is there's still a lot of manual input. Where we're moving toward is using more chatbots or voice recognition to help sellers sell instead of getting sellers to fill out more forms.
How will AI improve ads?
Typically in the global GDP, labor is about 20% of total GDP, but in marketing it's 51% and that 51% is locked in very inefficient processes and manual optimizations. It's not that machines will do all the work for us in the future, but I would like to know what are the core components of what makes a really effective communication [using AI], and that will make me a lot smarter.
Can you give us a hypothetical?
There's a technology called RevJet, which might not be exactly AI, but it used to be that an agency would present five creative concepts to you, you'd argue about it and one would somehow win and maybe we'd do a focus group or some copy tests around it.
Now, when an agency gives us five concepts, we put them in RevJet and they're trafficked in real time against real media. Once a concept starts to win, we take that winning concept and go back to the agency and say, "Create a multitude of elements [like new headlines and colors] for this winning concept." RevJet real-time optimizes all those elements until you get a winning piece of creative.
What results does it drive?
We see 10- to 50-fold increases in click-through rates as a result of that optimization process, and agencies love it because instead of shopping their ideas up and down the hallways, it turns it into more of a fact-based decision. I think that's where AI can play a role: [to] help us understand what's working to move much more quickly.