It’s been well documented that search is no longer just a search box on a webpage with a list of links. It’s literally breaking out of the box; it’s changing in some dramatic ways. Search has slowly been integrated into more & more technologies/devices, apps & sites we use every day, from phones to gaming consoles.
Consumers today have an infinite amount of information at their disposal and new devices to get things done. As a result, search has had to evolve beyond the box. Search is the connective tissue that helps us connect with the people, places & things we care most about.
Those changes just scratch the surface of where search is headed. I want to paint you a much broader & more powerful picture of the future of search. One that has incredible impact for consumers and opportunity for marketers.
As we think about where search is going, the construct of the time we live in is influential to how search is going to be shaped.
Painting a picture of the future of search relies upon a number of major evolutions taking place in the world of digital. Improvements in AI, Machine Learning, and natural language processing are the transformative forces at play in digital that are rewriting consumer experiences.
These advances have a direct impact on search, transforming it from a reactive service requiring precise data inputs to produce ranked outputs to an empowered, predictive service - leveraging algorithms that tie in incredible amounts of data from a huge array of sources to think ahead of a consumer’s needs that deliver personal results on any device.
As we begin to explore this evolution, the first question we need to answer is – what exactly is artificial intelligence?
AI is the science and engineering of making intelligent machines. It’s an expression that describes an entire corpus of study and work which itself is the combination of three building blocks: machine learning, human learning and data science. In many ways there is a strong analogy between AI and raising a child.
Just like children get their foundational learnings from their parents, teachers and by the school books they read, machine learning is based on known properties, and the machine learns from the data. Think if/then scenarios. If your child behaves well, then will be treated by Santa. If he sees a puddle, then they should not to stomp in it to keep their feet dry. This is how machine learning works also: If you liked that book, then you’ll probably like these ones also. This is just one small example of a very complex field.
Kids learn fast that if they cry and shout they get your attention… Now, you will certainly want them to assimilate that such a behavior is not a normal mode of expression. Human learning is how we make course corrections to the machine learning that’s happening. Cortana, Microsoft’s digital assistant, has a team behind the scenes working on human learning so she can get smarter. The human learning give Cortana more personality, and her responses to queries are more human because of this.
Data science is the third brick of artificial intelligence. Data science is the discovery of unknown properties, or connections, in data. In this case, the machine is presented with a massive amount of data and asked to find connections in it. This is how we might discover that cars with a remote locking system in hotter climates have a greater chance of failing. We didn’t know there was a connection between these pieces of data until we went looking for that connection.
So where does that data that feeds machine learning, human learning and data science come from?
It comes from us! Artificial intelligence comes from us. It is us.
Artificial intelligence is only as intelligent as the data it takes in. It is only as fair – as human – as socially acceptable - as the data it takes in.
I’d like to share an example of AI, which to a certain extent illustrates how humans can influence how intelligent a bot can be.
Who remembers Tay, our first experiment as a Twitter bot? Tay learned from her inputs, which were hijacked by some people who wanted to influence her negatively.
In this case, Tay incited high emotion from people who engaged with her or read about what happened with her, even if Tay, herself, did not express emotion.
So, what can we reasonably expect from artificial intelligence?
The progress made in computing & processing power has enabled a fast acceleration of 3 technologies which underpin AI: object recognition, natural language processing and speech. If the AI can see, speak and listen, it is not far from being able to exchange with human beings transparently.
Capitalizing on the progress of machine learning around object recognition, natural language processing & speech, we have seen our expectations towards AI graduate from the most basic to much more advanced outcomes.
According to Silicon-Valley analyst, Ray Wang, there are 7 intertwined outcomes for AI, based on what we are now able to program via machine learning.
While nothing in AI would be considered easy, Perception is an example of early machine learning, now totally engrained in our daily life. Drawing on existing data, the machine delivers information about what is happening now. The weather, traffic – things that are measureable and reportable.
For humans, learning to express their perception is pretty simple – we’re at ease with describing our surroundings. It is dark. I am hot. Or, I am joyful - we learn how to do this almost immediately as children.
To illustrate Perception as an outcome, we can look at what we did at Microsoft to learn more about facial recognition. If you’re interested in seeing how we did, if our perceptions of your age are accurate, go to http://how-old.net and send your pic.
Next, notification. If I didn’t have my calendar delivering notifications, I would be a mess – late to meetings or unware of commitments because I can’t hold my schedule in my mind.
We learn notification early as well, perhaps starting with letting Mom know we’re hungry. It never stops – in school, we notify the teacher that we have the answer.
Suggestion is another area where we have grown to be familiar with this outcome, and is now engrained in our daily life. You love this machine learning with your Spotify account. If I listen to a song and I like it, the AI suggests more songs for me to enjoy. And you can always retain that Human Learning capability to ensure that the AI never drifts too far from Justin Timberlake… The early suggestions were basic, but imagine what can influence them today: demographic, geographic, day, time, weather, behaviours… The data sets are huge but we are now capable to process them in no time and identify some new, more obscure connections.
Our children learn a nice drawing will trigger a smile from their mother, or that it’s time to wash their hands before a meal. Over time, we don’t even have to remind them; they just know it’s what’s next…
…and it becomes automation. A suggestion or a recommended action can grow into automation based on learning your preferences. If you passionately follow the progress of your favourite team, it will start to automatically inform you of their performance. If you always make a reservation for 7pm on Saturdays, your AI will start to automatically fill in the date and time on your reservations.
Predictions can be the hardest machine learning to train, because so many variables can affect this outcome. Think of a child who sees Dad packing a suitcase; based on past behavior, this toddler knows that this means Dad is leaving for a few days, which is sad. But sometimes it also means that the child gets to travel with Dad. What factors will alert the toddler about what outcome to expect?
At Microsoft, we have developed a program called Bing Predicts which combines and models all the data signals we can find, and comes up with incredibly accurate predictions. We initially explored popularity-based contests like American Idol, for which the web and social signals are very strong and highly correlate with popularity voting patterns.
Bing Predicts could accurately project who would be eliminated each week during American Idol and who the eventual winner would be. More complex, we then turned to sporting events and even world political challenges. During the World Cup in Brazil, our team predicted accurately with 100% accuracy the winners of the final elimination round. In order to successfully predict a sporting event outcome, the number and type of signals we incorporated quadrupled from what we used to predict a basic popularity event like American Idol. This is because we recognize that popularity alone does not predict whether a team will win. A fan base has however special insight into the abilities of their teams, and those fans are having constant discussions about their team. This is called the “wisdom of the crowd.” We weighted their knowledge against player and team stats, tournament trends, game history, location and even weather conditions. This is how we were successful in our predictions.
Cortana Intelligence with Bing Predicts is currently in preview for our enterprise customers. We can integrate our rich Bing Predicts tool with your company data and goals to solve problems no marketer has ever been able to touch before.
For example, by monitoring social signals, anonymized and aggregated Bing search data, opted-in Internet Explorer session data layered over a concert promoter’s own data, we were able to predict the audience size of specific concerts in specific cities. This allowed the concert promoter to book the right size venue and provide the right amount of support, which varies a lot depending on performer and city. There is nothing like this on the market right now.
The next logical step after prediction is prevention. Again Bing Predicts shines in this category:
By analyzing large samples of search queries, Microsoft scientists have been able to identify internet users who are suffering from pancreatic cancer even before they were diagnosed.
The researchers focused on searches conducted on Bing that indicated someone had been diagnosed with pancreatic cancer. From there, they worked backward, looking for earlier queries that could have shown that the Bing user was experiencing symptoms before the diagnosis. Those early searches, they believe, can be warning flags.
(The search data was anonymized, so they couldn’t actually contact the searchers.)
Situational awareness for AI comes close to mimicking human behavior in decision making. We see situational awareness as a combination of many aspects of AI, from object recognition to conversational speech.
An example of this is the next video.
These 7 outcomes are complex and require a lot of training and time to accomplish. They are also interconnected and not mutually exclusive. They actually build upon each other to offer the benefits of AI to us, users.
When we extend this logic to digital assistants, we can summarize this in what we call the 3Ps, which is shaping where search is today & where it’s headed. [Personal, Predictive and Pervasive]
And these assistants are spreading to all form factors and becoming more pervasive. While these assistants today are predominantly in mobile devices, you can begin to seem them in TVs, speakers, homes, wearables, cars and more.
Take Cortana, it’s already embedded into devices running W10 – including PCs, Mobile Phones and Mixed Reality devices like HoloLens - even in devices like Samsung washing machines and services like Microsoft Edge, Skype and more. And coming in 2017, we’ve partnered with Harmon Speakers to bring a premium speaker with Cortana to market.
Today, the search results you receive may very well be different than another person for the exact same query. Search results today are more context aware then ever – understanding location, past behavior, demographics – to deliver more personal results.
The rise of digital assistants – Cortana, Siri, Alexa, Google Assistant – is a direct manifestation of the future of search being personal. To provide this relevant information that relates to you and only you, they access, with your permission of course, your email, search history, location, interests, the people you care about and more to be able to deliver this predictive and personal experience.
Personalization is not only about the results delivered but also the ways in which you can interact with search. We think of a system that allows us to speak and interact in the way that is most natural to us in that moment. Search must be able to understand and respond in that same form. That is a massive challenge for any search engine, but it also represents massive opportunity, with 65% of smartphone owners using voice assistants.
The way we search today is dictated very much by the devices we have and use. We’ve gone from typing on a PC, to talking and tapping on our mobile phones. But with the average consumer owning at least 4 connected devices, it’s clear that search is no longer just on your desk or in your pocket. As more connected and smart devices permeate our world, search will not just be on our desk or in our pocket but will be available on our wrists, in our appliances, in our cars and in emerging technologies like Mixed Reality.
We are expected to reach 25 Billion internet connected devices by 2020 and consumers will expect ways to interact and search on these devices seamlessly. With more smart devices around us, it’s clear that the future of search is pervasive – always available when called upon and always around us whatever device is near us or whatever website we are on.
Instead of searching on your own, these smart devices will be always on, learning about your needs and habits and ensuring you have the most appropriate experience possible.
Predictive search capabilities have been available in search boxes since 2004 – those suggestions helping you complete searches were just the beginning of predictive search. Predictive search has become a welcome part of our internet interactions, helping us search faster, find results quicker, and discover answers to questions we didn’t even know we had.
Today, search is pretty good at giving users what they're looking for, but less at telling them what they want to find and worse at finding things they don’t even know yet that they wanted. However, advancements in artificial intelligence and machine learning are improving predictive search capabilities.
As predictive search becomes more powerful, assistants like Cortana, Google, Siri are embodiments of what predictive search can be. These assistants can deliver important information about the traffic on your morning commute, your updated flight itinerary, and the results of last night’s football game on your phone, without you even asking or searching.
With your permission, assistants like Cortana today leverage the data it collects and aggregates from its users through Search, Mail, Maps, Calendar and more. It understands who you are, what you are doing and where you are doing it to predict what you want based on user behavioral patterns.
To bring all 7 AI outcomes – and the 3 Ps - to life at once, I want to share this final video with you. Conversations As A Platform with its conversational bots & deep machine learning were for us the vision of an imminent future of how brands and businesses will interact with their customers.
The rise of messaging apps where consumers now spent most of their time and technological advancements in NLP, AI and machine learning are creating new experiences in which consumers will search with and engage with that will be personal, predictive and pervasive.
Conversations as a platform unlocks a more human, personal way to discover, search for, access and interact with information. This new platform will enable us to interact with devices more intuitively, using natural language, conversations, evolving us from mechanical keyboard and mouse to touch and beyond. This allows us to have an on-going conversation and relationship in context.
This new platform includes personal digital assistants that knows you, knows about your world and is always with you across all your devices helping you with your everyday tasks. It has the perfect understanding of the context you are in.
This new platform includes bots, with the capability to take the power of human conversations, and apply it to everything. Think of bots as new applications that you converse with. Instead of looking through multiple apps, or pages and pages of websites, you can call on any application as a bot within this conversational canvas.
Search will never again be a constrained to writing in a search box. Instead, search will be a partner that can listen and communicate in dialogue with a consumer on any platform and any device. Thanks to new, more natural interfaces, voice search is becoming increasingly possible and accurate. That's the world that you're going to get to see in the years to come. We want to give consumers a canvas where they can communicate and search for information not just in a transactional fashion but in a way where they can have on-going relationships and dialogs.
CaaP can help brands engage with consumers more deeply across the full cycle of customer engagements. Conversations as a Platform provides an opportunity for brands to build an on-going relationship and dialogue with consumers and we want to provide brands a better understanding of the full cycle of customer engagements across any messaging service.
We’re in the early stages of exploring these opportunities but our focus today is to create a robust bot ecosystem with developers.
So how should marketers prepare and adapt for the future of search?
Personalization: With more competition for consumer’s attention and their expectation to deliver experiences for them, it’s important for marketers to begin to deploy personalization tactics. Basic use of targeting like remarketing, demo targeting will be tables takes as consumers expect unique results and experiences that are delivered to them.
Pervasiveness: With search expanding outside the box and available in any connected device, marketers will need to consider search not just as a bottom funnel tactic but across the entire CDJ. Searches will no longer just happen on a phone or PC but in TVs, wearables and more. Already today, searches happen across an entire consumers CDJ – from researching, consideration, comparison, purchase, and services, search. And as search becomes more pervasive, consumers will search upper funnel.
Predictive: The power of search data and capabilities have improved the ability of search to predict what consumers want. These capabilities extend beyond just providing consumers better search results but can enable marketers to make better decisions and deliver better experiences. Tools like the Cortana Intelligence suite are already utilizing search data to help companies and marketers improve demand forecasting, gain greater insights on ad campaigns and determine public impression on their consumer products. What marketers should do is harness the power of search data and insights beyond just search but for their business overall.
Privacy: Be transparent about how you talk to consumers and what data you use but also who you partner with. Ensure that your customer have controls in place and be sure that there are benefits to consumers when you do collect data.
CaaP: With conversational UI on the rise, marketers can begin to transform customer engagement with CaaP. Be sure to invest in a Bot strategy since consumers are already spending their time in messaging apps in an highly engaged manner. Not only will you improve the customer experience and reduce the need for human assistance, customers will be able reach you on any device in more personal and natural way in the right context.
Key message to land: Personalization: The Future of Search will be more personal. Search will deliver more personal results and allow for more natural ways to interact with computing. Pervasive: The Future of Search will be pervasive. Search technology will be inside the apps, sites and devices we use everyday making them smarter and more useful. Predictive: The Future of Search will be predictive. Search will predict what consumers want next but also search data will empower organizations with predictive analytics. Conversations as a Platform: The combination of search being personal, pervasive and predictive is informing new platforms that we’re creating, like Conversations as a Platform, where consumers will engage with digital assistants, bots and people to achieve more.
The Future of Search - Adela Popilkova
Search is changing
The next evolution of search
expected Internet of Things
units by 20205
74% of people trust search engines
almost as much as they trust
the website or brand4
Breakthrough in speech recognition
Lowest ever recorded word error rate with Microsoft
reaching human parity9, 10
will come from
image or voice by
of online consumers get frustrated
when irrelevant content appears674%
embedding AI into their
apps by 20181
1/2 of all developer
teams will be
The Cortana search box*
monthly active users, with
questions asked to date8
By 2020, the
customer will manage
of the relationship with
an enterprise without
interacting with a human3
New headsets with potential
to significantly alter the way
users interact with
are coming to
* in the Windows 10 task bar. 1. "IDC FutureScape: Worldwide IT Industry 2016 Predictions — Leading Digital Transformation to Scale", Nov 2015, IDC FutureScape, https://www.idc.com/research/viewtoc.jsp?containerId=259850. 2. “Internet Trends 2016 - Code Conference”,
KPCB, 1 June 2016, http://www.kpcb.com/internet-trends. 3. “Gartner Customer 360 Summit 2011”, Gartner, Mar-Apr 2011, https://www.gartner.com/imagesrv/summits/docs/na/customer-360/C360_2011_brochure_FINAL.pdf, 4. "SMX EAST RECAP: Catalyst partners with Bing
to study search in the media mix, search + other channels", Bing Ads, 18 Nov 2016, https://advertise.bingads.microsoft.com/en-us/blog/post/november-2016/smx-east-recap-catalyst-partners-with-bing-to-study-search-in-the-media-mix,-search-other-channel. 5. "Gartner
Says 6.4 Billion Connected "Things" Will Be in Use in 2016, Up 30 Percent From 2015", Gartner, 10 Nov 2015, http://www.gartner.com/newsroom/id/3165317. 6. Denenholz, Jeff, "Online Consumers Fed Up with Irrelevant Content on Favorite Websites, According to Janrain
Study", Janrain, 31 July 2013, http://www.janrain.com/about/newsroom/press-releases/online-consumers-fed-up-with-irrelevant-content-on-favorite-websites-according-to-janrain-study/. 7. Cortana Dev Center, April 2017, https://developer.microsoft.com/en-us/cortana. 8.
“Why 100 million monthly Cortana users on Windows 10 is a big deal“, TechRadar, 20 July 2016, http://www.techradar.com/news/software/operating-systems/why-100-million-monthly-cortana-users-could-be-a-bigger-deal-than-350-million-windows-10-installs-1325146. 9.
"Historic Achievement: Microsoft researchers reach human parity in conversational speech recognition", Microsoft, 18 Oct 2016, https://blogs.microsoft.com/next/2016/10/18/historic-achievement-microsoft-researchers-reach-human-parity-conversational-speech-
recognition/#sm.00000jv1wtv9nacq7r266s8uucy4p. 10. "Technology Quarterly: Finding a voice", The Economist, 1 May 2017, http://www.economist.com/technology-quarterly/2017-05-01/language#section-5.
Nearly 75% of online consumers get frustrated
when content appears that has nothing to do
with their interests.
Personal Digital Assistants
Rise of Digital Assistants like Cortana are able
to provide results and information that relates
to you and only you.
Natural Language Processing
The methods of text inputs are being
complimented with more natural and personal
experiences, like conversations and voice.
More Connected Devices
Expected to reach 25 Billion internet connected
devices by 2020 and consumers will expect ways
to interact and search on these devices seamlessly.
Search as the Intelligence
Connected devices will be smarter because of
search technology and knowledge.
Persistent and Ambient
Smart devices will be always on and always around
you, learning about your needs and habits and
delivering information and knowledge to you.
Proactive Personal Assistants
Assistants like Cortana are delivering relevant,
personalized information to users, all but
eliminating the need for search as we know it.
Advancements in AI and machine learning are
improving predictive search capabilities and
will soon help users find things they don’t even
know yet that they wanted.
Applying search data
Companies are utilizing search data to help
them predict patterns and behaviors in their