How Marketers Can Start Integrating AI in Their Work

may18-28-200024729-002-Stephen-Swintek

STEPHEN SWINTEK/GETTY IMAGES

According to Constellation Research, businesses across all sectors will spend more than $100 billion per year on Artificial Intelligence (AI) technologies by 2025, up from a mere $2 billion in 2015. The marketing industry will be no exception.

AI holds great promise for making marketing more intelligent, efficient, consumer-friendly, and, ultimately, more effective. Perhaps more pointedly, though, AI will soon move from being a “nice-to-have” capability to a “have-to-have.” AI is simply a requirement for making sense of the vast arrays of data — both structured and unstructured — being generated from an explosion of digital touchpoints to extract actionable insights at speeds no human could ever replicate in order to deliver the personalized service consumers now demand.

Interestingly, in many cases, the sophistication of AI technologies has already advanced further and faster than most marketers’ ability to actually make use of them. On the one hand, there are the technical challenges of gathering and normalizing data inputs — the act of making different types of data comparable — connecting them to a unified view of the customer, and then aligning the AI-driven decisions to real-world actions. On the other hand, there are also real philosophical, ethical, or at least policydecisions to be made on the value exchange between marketers and consumers when data is shared and used to optimize marketing experiences.

The good news is that, as an industry, we are starting to see meaningful progress on both fronts. For businesses looking to keep pace with innovation and leverage AI, there are steps they can take today. But first, what are some examples of how AI can help make marketing more effective?

Using AI in Marketing

Smart marketers are developing, partnering to build, or integrating AI into their tech stacks to get better at what they do. AI is already being used in ad targeting and customer segmentation, but there are more possibilities in store such as:

  • AI-powered chatbots use all the customer data at their disposal to answer questions and give advice to customers considering making a purchase. Take Sephora’s Kik bot, which quizzes customers about their makeup preferences and then follows up with specific product information.
  • AI-enhanced image search allows users to upload pictures of products they are interested in, to find relevant shopping ideas. For example, companies such as CamFind let you snap a picture of something in the physical world, and get information related to it. Say you see a poster for a movie you’d like to see.  Snap a photo, and CamFind will show you movie recommendations, times, and locations.
  • Personalized training routines and nutrition information can be created based on data from other consumers with similar lifestyles. For example, UnderArmour leveraged IBM Watson’s AI to create a “personal health consultant” that provides users with timely, evidence-based coaching around sleep, fitness, activity, and nutrition.
  • Optimized advertising uses AI to make decisions based on the full range of data available — including unstructured data such as sentiment and mood. For example, IBM, this time as a corporate marketer, teamed up with MediaMath (where I am the CMO/CSO) to activate true AI-driven programmatic marketing using its Watson Cognitive Bidder, to extract predictive signals from exposure to large amounts of data.

How to Put AI Into Practice

There are three key areas of consideration to help you get started leveraging AI for marketing today: affirm your consumer data policies, make your data actionable, and select the right AI partner.  Let’s look at each in more detail:

Affirm your consumer data policies: Before getting started, it is important to confirm your policies regarding the handling of consumer data, transparency, and control.  Your company needs to make sure that you’re complying with the EU’s new General Data Protection Regulation (GDPR). Consumers should be able to interact with connected devices — from web browsers to mobile phones to voice assistants — knowing that their data is being used in transparent ways, in a manner consistent with their preferences and expectations, to which they have explicitly consented.

Make your data actionable: There are three pillars here to consider:

  1. Common identifier: With the explosion of devices, digital identity today is deeply fragmented, leading to head-scratching experiences where, for example, a gym franchise might roll out a great new membership offer…to someone who is already a member. The first step is to establish a common identifier, usually an alpha-numeric string, across various touch points and data sets to help create a unified view of the customer. Emerging solutions such as DigiTrust are helping marketers tie together their various touchpoints in a safe, respectful, and scalable way.
  2. Data gathering: AI can make sense of all your data and extract insights from it, but only if you can collect and normalize it before activating it. Choose a data management platform (DMP) that incorporates the identity solution selected above and can handle data from a wide range of sources, so you can collect, organize, and centrally manage all your data, segment it into granular audiences, and activate it for marketing in real-time.
  3. Data end points: The best customer experiences cut across touchpoints, whether they are “paid” (online, Connected TV, or digital audio ads), “owned” (your stores, websites, call centers) or “earned” (PR, blog posts, social media). Although there are as yet few platforms that can deliver marketing across all of these touchpoints, technology such as IBM’s cloud-based Universal Business Exchange can build upon the common identifier and data gathered to drive consistent execution across a range of platforms and tools.

 Select the right AI partner: With the policies, data assets, and pipes in place, the stage is set for AI to make optimal decisions and drive real business performance. For best results and fastest time to real return on your investment, be sure to choose a partner that has true AI, not simply rules-based decisioning, which is impossible to scale for the volume of data and combinations of interactions that marketers are managing today. In addition, be sure to choose partners with real experience addressing your particular use cases and working with your existing technology partners. Finally, as with any vendor relationship, be sure to check on your partner’s ethical and philosophical approach to AI, as this is still an emerging technology that demands thoughtful guardrails to produce effective results in a responsible manner.

AI is no longer just hype — it’s a “have to have” that you can start integrating into your marketing today with the steps outlined above.


This article was originally posted on –

https://hbr.org/2018/05/how-marketers-can-start-integrating-ai-in-their-work


About The Author

Dan Rosenberg is CMO/CSO at MediaMath.

Preparing your business for an AI future

Artificial Intelligence is a powerful tool that has the potential to fuel new ways of working, improve decision-making, and increase efficiency in businesses across the board. In fact, innovations in AI are happening so fast that the future is nearer than it ever has been before, so the adoption of this new technology is key to future-proofing your business.

However, because everybody’s talking about AI, it can be hard to make out who is speaking the truth…and who’s just hyping it up. This can make it difficult to know where to start when deciding the ways in which AI can help your organisation’s operations, and cloud any judgement on where to start.

Know which problems AI is best placed to tackle

A common misconception surrounding AI is that it describes some kind of superintelligence that can be applied to any situation and, like a human can, have an intuitive idea of what needs to be done, with limited information. This is simply not the case. AI is only good at solving really specific problems, and you have to map out a clear target towards which it can be trained.

This is why the problems you choose to tackle with AI must be well-defined and goal-orientated. The task itself can be complex and difficult to solve, but it must have a clearly defined objective, with clearly defined parameters. It’s important to be able to measure the success of the AI algorithm according to specific KPIs.

What’s more, AI needs vast amounts of data to learn from. The amount of data you have must be one of your first considerations when defining the problem you want to tackle. On one hand, if there isn’t enough data produced by the operation towards which you are deploying your AI, then there won’t be enough data for it to be taught the minute differences and combinations that it needs to be able to accurately identify patterns and make predictions. On the other hand, if there aren’t enough data-points in the process in the first place, then it might not be the right place for you to implement AI, as it won’t be able to offer you any meaningful insights.

Taking the time to consider the problem and data involved first will save you money and time on experimentation and development further down the line. This doesn’t mean that you should just limit the scope to the easiest problems, but it means defining your goals to the point that you don’t waste time training an AI algorithm that misses the point of what you’re trying to achieve.

Potential AI projects should be orientated around minimising the costs and making strong predictions for core business problems, opportunities or challenges. Just like any other technology in business, AI should be viewed as a tool that can help make your organisation more effective, profitable or streamlined.

For example, when it comes to sales, an AI model can accurately predict a business’s sales using patterns found in historical sales data. In manufacturing, AI can predict any machine malfunctions before they even happen, potentially saving a company thousands of man hours and millions of pounds. In farming, AI can use satellite imagery and weather data to deliver accurate predictions that tell them what to plant, how often to water it, and how to fertilise it – saving a lot of guesswork and money.

Choose the right AI solution for you

A new generation of tools is emerging that provide end-to-end AI, allowing organisations to operationalise the technology at speed and for a reasonable cost. Whereas before, AI had to be developed in labs by highly-skilled technicians, this new generation of tools means that much of the nebulous technical work is done for you. Using an accessible interface, you can set the parameters, input your data, and define the desired goal that you want your model to be applied to.

The rapid development happening in AI software and hardware means that you want to choose a solution that’s scalable and future-proof. At the same time, there’s no point in implementing an AI that’s not bespoke to your enterprise. It’s therefore important to invest in an AI platform that’s flexible and capable of adapting to your specific needs. That’s why it’s important to review the technical capabilities of the platform and have a good idea of the different types of AI itself. For example, traditional Machine Learning can solve far fewer problems than Deep Learning, so opt for the latter if you want a viable solution that lasts for posterity.

In-house competencies must also be a top consideration – while it’s important to pick a solution that you can tailor to your own needs, you’ll also want to make sure that your tech team are able to manage and maintain your AI tools without the need for too much extra training or external consultancies.

Budget of course underpins all of the above considerations. Not only must you consider the future costs of running your AI solution, but you also want to make sure that the solution is robust and won’t start requiring expensive hardware – in this case, cloud-solutions are your best bet to keep the total cost of ownership to the minimum.

Spread the word

If you think that AI has the potential to transform your company, educate your organisation and senior management.

Because of AI’s complexity, demonstrating its possibilities to non-technical audiences and those working in other departments within your organisation may seem like a daunting task. But taking the time to make sure your colleagues are informed will not only lead to the faster adoption of AI by decision-makers, but it will also improve the efficiency with which you implement AI into your company once the decision has been made.

Company-wide basic knowledge of AI will be hugely valuable and will encourage certain changes in data practice management. This doesn’t mean everyone has to be or become an expert. The point is to begin to shape the mindset and strategy of the organisation around the core principles that allow AI to work, and what it’s particularly good at doing.

To harness the real power of AI, you can’t just go to an AI consulting firm and ask them to optimise your profits. The process of building an AI model must start within the organisation itself. To be more specific: it must start at the top.

 


This article was originally posted on –

https://www.itproportal.com/features/preparing-your-business-for-an-ai-future/


 

Luka Crnkovic-Friis, CEO and co-founder, Peltarion
Image Credit: Shutterstock/Mopic

Constellation Research : 46% of Companies are Investing in AI for Sales and Marketing Purposes

Artificial intelligence (AI) adoption is relatively modest across large enterprises, but marketing and sales organizations are embracing the technology most quickly.

Forty-six percent of companies are investing in AI for sales and marketing purposes, and 50% are deploying AI projects for commerce and customer service, according to a report released Thursday by Constellation Research. By contrast, 48% of firms said they do not plan to invest in AI for finance, legal and administration.

It’s much easier for marketers and sales teams to adopt AI because of the breadth of plug-and-play options in the market, said Courtney Sato, director of research development at Constellation and co-author of the report.

“They’re already using a CRM platform or some enterprise software that has an automated tool set built in,” she said. “All they have to do is turn it on and pay an extra fee.”

Despite marketer uptake, the report, which surveyed 50 C-level executives at large companies across 12 verticals, found enterprise investment in AI is modest across the board. Most companies (51%) are investing less than $1 million in AI, although 60% plan to grow their budgets by at least half this year.

“Even early adopters are not spending a lot on AI yet,” Sato said. “The budgets are pretty modest compared to everything we’ve been hearing about AI.”

Most companies are in the early stages of adopting AI. Forty percent of respondents said they’re investing in building a data lake and big data analysis capabilities, which can take up to two years. Just 30% are investing in predictive analytics and machine learning, while 22% are working on natural language processing and image recognition. And only 8% are building deep learning neural networks.

“Companies are investing in the foundational technologies,” Sato said. “They need data lakes and statistical analytics, so they’re going to invest in those first. They don’t have the appetite for the more cutting-edge technologies.”

How they’re doing it

Depending on what the AI is being used for, companies have three options for deploying it: build their own infrastructure, leverage computing power from providers like Google or Amazon or license off-the-shelf tools from platforms such as Salesforce or Oracle.

Each option has its pros and cons, Sato said. Building AI infrastructure is a huge and long-term investment, whereas working with a vendor or cloud provider can allow companies to start using AI more quickly. But organizations choosing to leverage or license should be wary of getting locked in with a cloud provider and aware that out-of-the-box AI tactics aren’t very flexible.

“Contracting with a vendor is helpful if you don’t want to build the back-end infrastructure,” Sato said. “The drawback is it’s not as customizable, and then you face lock-in.”

Companies that want to own their AI technology need to source from a limited and competitive talent pool. Eighty percent of respondents told Constellation they need to bring on more employees to implement AI solutions, and 40% said they’ll need to hire a “significant” amount of new talent.

Attracting talent will be tough, especially when competing with large tech and consulting firms with deep pockets, Sato said.

“If you have someone good, big tech companies are going to come after them,” she said. “It’s going to be a difficult decision for the employee not to go there.”

Because of these challenges, most companies are taking a hybrid approach to adopting AI. Twenty percent of respondents said they are both building and working with a cloud provider or vendor, while 26% said they’re doing all three.

“This is the fastest way to dip your toe in the water and then decide which strategy works for which project,” said R “Ray” Wang, principal analyst at Constellation.

Friction and oversight

Companies are cautious about deploying AI because it will fundamentally change the way their organizations function. It will also demand new roles and skills and affect roles beyond manual jobs.

Fifty-four percent of executives told Constellation they’ll need to understand how to restructure their business to accommodate AI, while 50% say they’ll need to acquire data expertise. Forty-eight percent said they need to learn how to manage an AI-augmented team.

Just 6% of executives said they don’t expect their roles to change as their organization adopts AI.

“Most executives are going to have to understand how an algorithm works and how to identify false outcomes,” Sato said. “After a while, I’m not sure how much value an executive that doesn’t have those competencies will be.”

Many executives are anxious that AI will replace their jobs and automated technology won’t be able to do tasks as well as humans, Sato said.

“There is fundamental distrust of the technology and what it’s going to do to people’s livelihoods,” she said.

There are also looming questions about how AI will be regulated. The General Data Protection Regulation in Europe and the Facebook-Cambridge Analytica debacle mean US data privacy regulation could be around the corner – and companies should put it high among their priorities.

“AI can create personal data,” Sato said. “You might give consent to share information A and B, but when those two things come together, they make information C that you didn’t consent to share.”

AI, however, gets smarter by consuming as much data as possible. Tech companies may now be gathering as much data as possible before potential regulation, Sato said.

“It’s against their interest to curb data collection,” she said. “They can’t help themselves.”

 


This article was originally posted at AdExchanger – https://adexchanger.com/research/constellation-research-ai-piques-marketers-interest-but-overall-adoption-is-slow/


Author

Alison Weissbrot

 

How The Development Of AI Has Advanced The Technology Available For Chatbots

 

Shutterstock

Artificial intelligence (AI)-powered chatbot platforms are taking the world by storm and revolutionizing diverse industries, but before AI was commonplace in chatbots, they were powered by different technologies that were not necessarily intelligent. Chatbots existed as early as the 1960s, and they have rapidly evolved since the development of AI. Today, chatbots are capable of performing incredible feats that come very close to human thinking and thought-processes.

This mimicry of the human thought-process comes from a chatbot’s ability to set goals and pursue them autonomously without any specific instructions — an ability referred to as a chatbot’s agency. This ability (AI) is the reason why chatbots are making a splash in areas such as customer service and sales today.

Chatbots Revolutionizing Customer Service And Sales

I think chatbots are the future, especially in customer service and sales. Here’s why:

1. Intelligent Chatbots Can Provide Preemptive Customer Service

Feeding on thousands of chats and their data, smart chatbots can be trained to recognize important patterns that indicate distress, frustration and annoyance and respond to issues before they escalate.

2. Smart Chatbots Can Reduce Training And Operation Costs 

While human customer service agents need to be trained from time to time, based on new products and emerging customer behavior trends, intelligent chatbots need to be trained only once. They are self-learning and update knowledge autonomously.

3. Autonomous Chatbots Can Handle Customer Service 24/7 

With human customer service agents, you have to hire in shifts to enable 24/7 customer service, which can double your operations and training costs. Autonomous chatbots can enable the same quality of service at no extra cost.

4. Intelligent Chatbots Can Scale The Volume Of Customers Served

From a logistics perspective, human customer support agents can face bottlenecks with support queries. Chatbots, on the other hand, can attend to customers simultaneously without deteriorating the quality of service provided.

5. Smart Chatbots Can Personalize Support With High Accuracy 

Sales managers may fear adoption of chatbot technology because they assume that bots take away the human touch and lack personalization. The opposite is actually true. Intelligent chatbots have the potential to greatly personalize conversations, based on their vast stores of highly-organized data that humans can’t compete with.

Stages Of AI, And How They Can Be Applied In Chatbots Of Different Kinds

Over the years, bots have evolved in their intelligence and abilities, shedding old avatars for newer and more advanced ones. To understand what AI technology’s development has done for chatbots, you have to be exposed to the stages of AI’s evolution and what each stage had to offer in terms of abilities.

Stage One: Cue Readers

Chatbots in stage one have no real ability to understand what you’re saying and simply pick up phrases from your speech that they then match to an internal database before spouting scripted answers. However, such chatbots have the interesting ability to insert amusing fillers in conversations to prevent their lack of knowledge from betraying their limited capabilities. That’s why developers don’t fully ignore stage one and give it some importance, as the later stages pay little attention to emotion and personality in conversations.

Stage Two: Form Fillers

Chatbots in stage two are designed to perform very specific tasks such as making a restaurant reservation or ordering food for delivery. These chatbots have built-in forms that specify all the information that they need to receive from customers and continue making conversation with people until they have filled all the blank spaces. Such chatbots not only go by important phrases but also follow certain rules written by their programmers.

Stage Three: Intent Readers

Chatbots in stage three are akin to talking encyclopedias and are trained to run search queries for the information they need to converse. Such chatbots are constantly indexing information in massive databases from which they pull facts when needed. These bots can do much more than just state facts — they can also identify the part of speech being used by customers and accordingly generate answers to satisfy those queries.

Stage Four: Connection Makers

Chatbots in stages one, two and three are inept at holding long-running conversations with customers. This is because they typically read each sentence of a conversation separately without making interconnections between sentences. Bots in stage four, however, store histories of conversations and can draw from them to make highly coherent and relevant answers. Such bots are even capable of learning based on previous conversations and can constantly improve their ability to have conversations.

Stage Five: Problem Solvers

Chatbots in stage five understand the core problem that a customer is trying to solve. While you could program bots in stages one to four to search for solutions based on certain phrases, bots in stage five can solve such problems more efficiently. For instance, if a customer wants to catch a bus to a destination at a certain time, bots in stages one through four can check if such a bus is available and relay the information, but bots in stage five can suggest alternatives if such a bus is not available.

Stage Six: Companions

Chatbots in stage six are what developers dream of and sci-fi authors write about. Built somewhat like the robot Sophia, chatbots in stage six should have the ability to maintain dynamic and broad conversations that are not limited by subject or depth of any sort. Such bots can replace humans completely in most situations that require autonomous-thinking.

Autonomous chatbots are changing the world every day, and I think a future where such bots run most business tasks is not far away, considering the continuing development of AI and the technology that supports chatbots.


Originally posted on Forbes –

https://www.forbes.com/sites/forbestechcouncil/2018/05/23/how-the-development-of-ai-has-advanced-the-technology-available-for-chatbots/


 

Author

Laduram Vishnoi

Artificial Intelligence Is Making Better Chatbots For Businesses

 

Just a few short years ago, having “conversations” in human languages with machines was pretty much universally a frustratingly comedic process.

Adobe Stock

Adobe Stock

Today that has changed. While natural language processing (NLP) and recognition is far from perfect, thanks to machine learning algorithms it’s getting increasingly closer to a point where it will be harder to tell whether we are talking to a human or a computer.

Business has capitalized on this, with increasing numbers of chatbots deployed, usually in customer service functions but increasingly in internal processes and to assist in training.

At ICLR 2018 in Vancouver, Salesforce’s chief scientist, Richard Socher, presented seven breakthrough pieces of research covering practical advances in NLP including summarization, machine translation and question answering.

He told me “NLP is going to be incredibly important for business – it is going to fundamentally change how we provide services, how we understand sales processes and how we do marketing.

“Particularly on social media, you need NLP to understand the sentiment around your marketing messages and how people perceive your brand.”

Of course, this raises some issues, and one of the most glaring is, do people really want to talk to machines? From a business point of view it makes sense – it’s incalculably cheaper to carry on 1,000 simultaneous customer service conversations with a machine than with the giant human call center which would be needed to do the same job.

But from a customer point of view, are they gaining anything? Unless the service they receive is faster, more efficient and more useful, then they probably aren’t.

“I can’t speak for all chatbot deployments in the world – there are some that aren’t done very well,” says Socher.

“But in our case we’ve heard very positive feedback because when a bot correctly answers questions or fills your requirements it does it very, very fast.”

“In the end, users just want a quick answer, and originally people thought they wanted to talk to a person because the alternative was to go through a ten minute menu or to listen to ten options and then have to press a button – that’s not fun and its not fast and efficient.”

Key to achieving this efficient use of NLP technology are the concepts of aggregation and augmentation. Rather than thinking of a conversation exclusively taking place between one human and one machine, AI and chatbots can be used to monitor and draw insights from every conversation and learn from them how to perform better in the next one.

And augmentation means that the machine doesn’t have to conduct the entire conversation. Chatbots can “step in” for routine tasks such as answering straightforward questions from an organization’s knowledge base, or taking payment details.

In other situations, the speed of real-time analytics available today means that bots can raise an alert when they detect, for example, a customer becoming irate – thanks to sentiment analytics – prompting a human operator to take over the chat or call.

Summarization is another highly useful function of NLP, and one which is likely to be increasingly rolled out to chatbots. Internally, bots will be able to quickly digest, process and report business data when it is needed, and new recruits can quickly bring themselves up to speed. For customer-facing functions, customers can receive summarized answers to questions involving product and service lines, or technical support issues.

Chatbots are a form of the ‘intelligent assistant’ technology which powers Siri or Google Assistant on your phone, or Cortana on your desktop. Generally though they are focused on one specific task within an organization.

One study found that 40% of large businesses have implemented this technology in some form, or will have done so by the end of 2019.

Among those, 46% said that NLP is used for voice to text dictation, 14% for customer services and 10% for other data analytics work.

Chatbots are also increasingly ubiquitous in collaborative working environments such as Slack, where they can monitor conversations between teams and provide relevant facts or statistics at pertinent points in the conversation.

In the future, chatbots will probably be able to take things even further and propose strategy and tactics for overcoming business problems.

Socher tells me “They will probably be able to help us craft marketing messages, based on understanding of the language of all the things that have been successful in the past.”

Another example could be customer service bots which can allocate resources to dealing with customer cases based on the classification and sentiment analysis of the conversations they are having.

As with all AI, development of NLP is far from a finished process and level of conversation we are able to have today will undoubtedly seem archaically stilted and unnatural in just a couple of years’ time.

But today, organizations are clearly becoming more comfortable with the idea of integrating chatbots and intelligent assistants into their processes, and confident that it will lead to improvements in efficiency and customer satisfaction.
Bernard Marr

 

Continue reading “Artificial Intelligence Is Making Better Chatbots For Businesses”