Tech Experts Give Their Best Advice for AI Integration

Mediaplanet sat down with experts in the field of technology to discuss seamlessly integrating AI.

The past few years have brought an unprecedented wave of chatter about the impact artificial intelligence will have on our world and livelihoods. But this sphere of technology has been in existence for over 50 years. To what do you attribute this new focus on Artificial Intelligence—to what do you attribute this new focus on AI?

Artificial Intelligence has found its sweet spots in several applications today due to a combination of enabling factors.

To begin with, thanks to proliferation of devices, there is an abundance of data generated that enables AI algorithms such as machine learning to be trained faster. The falling cost of high performance computation and popularity of cloud adoption models means that today multi-layered artificial neural networks, also called deep learning machines can be deployed quickly and cost effectively. The AI research community also have made significant advances in areas such as cognitive algorithms that can simulate human capabilities such as real world knowledge, intelligent reasoning, etc.

The result – a quantum leap in both in capability as well as in potential applications. It can be said that while the last 50 years has been the research phase of AI, the next decade will see insertion of AI in more and more mission critical applications in the industry that will touch our daily lives.

Cognitive interfaces that are powered by chatbots that understand human language and that synthesizes a personalized response thereby driving meaningful conversations will be a new norm in customer engagement.

Already, cognitive chatbots such as the one deployed by CogniCor give advice to users on banking transactions, sell insurance policies in  fully automated mode  and in also help customer service agents quickly respond to queries. Over 4Million automated conversations have been processed through CogniCor deployments and our Virtual Agent Creation platform is used by companies to setup virtual agents for themselves. IT research firm Gartner has predicted that by 2020, 85% of customers will experience customer service without having speaking to a human being. This means that the apps and automated response systems that are currently dealing with these requests will soon be replaced by more advanced AI within the upcoming years.

As more and more cognitive interfaces are deployed, the enterprise world will see a new level of service expectations from the users. As users gets accustomed to engaging AI driven systems, they will expect businesses to know their digital self from their previous interactions and data and expect businesses to interact with automated systems which will display cognitive intelligence and will deliver results.

What applications of Artificial Intelligence can businesses realistically implement into their organizational models now?

Delivering an exceptional online customer experience is a challenge for every e-commerce enterprise. With the use of AI, a smart digital sales assistant, users can get an experience that compares with the that offered by a human assistant when it comes to conducting a pre-sales conversation. This is completely possible today and has proven to generate more sales and cross sell opportunities. A case in point is that OCBC bank of Singapore had generated $40Million new mortgages from conversations that originated in a chatbot channel.  Users could navigate complex topics such as product selection or customer on-boarding where the AI enables personalized interaction experiences possible.  Coupling this capability with devices such as Alexa will open up a completely new way of doing commerce.

Another advantage of having AI system is to augment the experience that is delivered by human agents. Currently companies use this  to help their customer service executives quickly and consistently answer complex customer queries that involve data gathering and reasoning over several policies. In an age where several products are launched and policies changes frequently, this reduced the time to train the company’s agents.

In automating internal workflow to increase efficiency? 

Without a doubt, AI enabled systems have drastically reduced process times, thanks to the automation and autonomous decision making capabilities. Taking and example from one of CogniCor’s recent deployment for IT service support, the ordering of a laptop for employees which involved selection,  order placing and status tracking is now fully automated and instantaneous. Insurance application in the US has been another area where CogniCor was able to implement a system that reduces health insurance application processing time by bringing a “virtual underwriter chatbot” to help the agents answer queries and process a customer’s application faster. Efficiency gains from such bold moves enable companies to be online at core and keep a healthy bottom line in operational expenses and at the same time delight customers.

In gathering actionable insights from big data to drive strategic decision making?

Chatbot driven Conversations generate huge amounts of data that companies can use to drive targeted decisions. IOT enabled devices also generate vast data points. AI algorithms use these to identify patterns and enable enterprises to predict likely customer behaviour or market trends. Advertisements and marketing campaigns could be run using this insights and also made more targeted towards specific market segments. This kind of predictive analytics enables businesses to anticipate user’s needs in advance and offer exceptional levels of service efficiency.

For companies looking to implement any of these applications of AI into their business model, what should they be doing internally to prepare for this digital transformation?

Companies need to prepare internally embrace the AI wave  and adopt an agile mindset to capitalize on the opportunity so that they are not left behind when sands are shifting in the industry.

Best results deployments with enterprises came when there is a dedicated team within the organization that has a strong commitment to deploy such systems to production  partners with a technology vendor that has strong grasp on the underlying technology and has expertise in deploying to industry. Equally important is the mode of collaboration and deployment cycle. As AI systems are not 100% deterministic, organizations must be prepared to expect initial setbacks and course corrections when the system learns from feedback and stabilizes.

Start fast and start small, iterate infinitely. Instead of thinking of AI as a replacement, initial deployments must focus on leveraging AI to augment current capabilities. Customers are vital feedback elements of AI system, hence the release of the technology must involve customer in the loop. Also as the disruption is lead by many startups,  a willingness to work with startups and ability to engage with them cutting across bureaucratic barriers within the organization is essential.

The past few years have brought an unprecedented wave of chatter about the impact artificial intelligence will have on our world and livelihoods. But this sphere of technology has been in existence for over 50 years. To what do you attribute this new focus on Artificial Intelligence—to what do you attribute this new focus on AI?

Recent advances in AI technology—specifically the deep learning branch of ML—are enabling us to leverage all this data in new ways. And we can do it quickly, thanks to the startling pace at which AI tech is becoming commoditized. The Big Four all offer machine learning as a managed service (Google TensorFlow, AWS Machine Learning, Microsoft Azure, IBM Watson). This is augmented by the rash of startups focused on niche and vertical AI use cases, making it increasingly easy for companies of any size to add AI-powered capabilities to their products. No surprise the hype is at peak volume.

On the business side, there’s a sobering reality. The open sourcing and rapid commoditization of new technologies is an inevitable trend, and is moving more quickly than ever. David Cancel, CEO of Drift, recently went so far as to opine “Product-based differentiation is going away. Act accordingly.” What he means, and I agree, is that companies must move beyond differentiating on features and tools. Try Googling the next cool product enhancement on your roadmap. Chances are you’ll find 15 GitHub libraries, 5 StackOverflow how-to threads, and 3 startups on ProductHunt that already do it.

What applications of Artificial Intelligence can businesses realistically implement into their organizational models now? 

• In improving customer experiences and product value?
• In automating internal workflow to increase efficiency?  
• In gathering actionable insights from big data to drive strategic decision making?

Workflow automation is just one obvious target for enhancing user experience. Reacting to—or even anticipating—user intent and granting a wish (booking travel, purchasing goods, processing orders) in a fraction of the time, across multiple business systems and apps, is one example we’re already seeing in the wild. An even deeper moat can be created by seamlessly integrating a bot-to-human handoff behind the scenes (machine-qualified leads passed to sales reps); or by bringing aggregate benchmark data to optimize the workflow (predicting best time to purchase airline tickets based on seasonal travel patterns and price fluctuations using Big Data).

New business model optimization opportunities are also emerging thanks to AI. The attractiveness of monetizing advanced insights (vs. simple reporting) is a good example. The first wave of cloud technologies based their value prop on efficiency and productivity. With machine learning enabling the extraction and contextual delivery of predictive/perceptive insights, companies seizing on these capabilities can promise “work smarter” vs. “work faster,” and demand a premium accordingly.

For companies looking to implement any of these applications of AI into their business model, what should they be doing internally to prepare for this digital transformation? 

Ensure you’re building an internal culture that values and prioritizes patience, experimentation, and innovation. Getting AI right is playing the long game, and requires immense strategic discipline and focus in the face of ever-present pressures to fight fires, maximize short-term revenue, and cater to existing customer demands quickly. The payoff in the medium to long term will be immense—but you can only get there if the vision is clear, constantly evangelized, and embraced across all your teams.

The past few years have brought an unprecedented wave of chatter about the impact artificial intelligence will have on our world and livelihoods. But this sphere of technology has been in existence for over 50 years. To what do you attribute this new focus on Artificial Intelligence—to what do you attribute this new focus on AI?

While it’s true that AI has been around for quite some time, practical uses of AI have come a long way in the past few years.  It was kicked off with the likes of Siri and Alexa, but has continued to see momentum as the technology is rapidly maturing and expanding the universe of use cases it can be applied for.  Alexa has become a mainstay in many homes, but we are also seeing AI in many other areas of everyday life including customer service.  Today’s AI is easier for companies to implement, can self-learn and connect with users in a more human, personal and contextual way, and starts showing value much quicker than in the past.  This unprecedented focus on it is a direct result of this maturity – the technology is actually working and providing real ROI for businesses and consumers alike.

What applications of Artificial Intelligence can businesses realistically implement into their organizational models now? 

• In improving customer experiences and product value?
• In automating internal workflow to increase efficiency?  
• In gathering actionable insights from big data to drive strategic decision making?

If you aren’t seriously considering adding an AI component to your customer experience, you’re already behind.  Today’s customer wants immediate answers to queries and resolution to issues and while it is nearly impossible for organizations to staff a contact center to answer all queries in near real-time, AI is stepping in to help meet these needs.  In some cases, the technology is being used to empower customers to self-serve and in others, it is being used to help agents quickly become smarter about a customer so they can provide a better, more personalized experience.

For companies looking to implement any of these applications of AI into their business model, what should they be doing internally to prepare for this digital transformation? 

First and foremost is understanding what you want to use AI for.  It’s not one size fits all.  Some scenarios will be AI appropriate and some won’t. Understand which parts of your business make the most sense for automation and which may still require a human touch.  When implementing a customer-facing solution, companies should consider a hybrid approach – one that puts chatbots on the frontlines but has an easy and elegant way to escalate to a human where necessary.

The past few years have brought an unprecedented wave of chatter about the impact artificial intelligence will have on our world and livelihoods. But this sphere of technology has been in existence for over 50 years. To what do you attribute this new focus on Artificial Intelligence—to what do you attribute this new focus on AI?

In today’s world, organizations are competing on customer experience more than ever, so it’s no surprise that we have narrowed in on AI as a means to drive better engagement, loyalty and ultimately lower costs. At the same time, computational processing power has grown exponentially, and we now have an abundance of data from customer interactions across an ever-increasing number of connected devices. Add in cloud’s low-cost storage and you have an environment where AI can be the catalyst for real solutions that can enable machines to have intelligent conversations with people – predicting what a customer needs and providing personalized, automated service across both traditional and digital channels.

What applications of Artificial Intelligence can businesses realistically implement into their organizational models now? 

• In improving customer experiences and product value?
• In automating internal workflow to increase efficiency?  
• In gathering actionable insights from big data to drive strategic decision making?

AI can let consumers engage with brands in the same natural way they engage friends and family – simply texting, typing or talking into their device to gain immediate access to information. Businesses today can deploy a single AI-powered virtual assistant that can understand, and in some cases predict, a customer’s request (no matter which channel that customer chooses to engage) and either provide an immediate solution or put that individual on a path toward their desired outcome. The result is a better, more streamlined customer experience that ultimately drives loyalty while keeping down costs. 

Customer service contact centers are being inundated with requests across an increasing number of platforms. AI can automate answering the simple questions that come in, allowing human agents to focus their time and energy on the customer conversations that are more complex. AI also makes the authentication process more efficient by letting customers avoid the frustration of pins and passwords and instead authenticate through biometrics. That means by the time a human does get brought in the loop, that customer has already been verified and the agent can quickly understand the situation, resolve the issue, and inform the AI-powered virtual assistant of how to handle it in the future.

AI is the technology that can derive meaning from data and when that concept is applied to customer service, the impact can be tremendous. Organizations today can deploy solutions that let them listen to customer engagements across channels and better understand how individuals are engaging through advanced analytics. Classifying and mapping each interaction not only helps businesses uncover areas of potential weakness in their product or service, but also guards against fraud. As interactions are analyzed, the findings are also fed back into the AI learning loop in real-time, so every conversation that follows is smarter and brands can predict what a customer might do next.

For companies looking to implement any of these applications of AI into their business model, what should they be doing internally to prepare for this digital transformation? 

The key, especially for large enterprises, is to consider the pieces of infrastructure they’ve already built and work to understand how they might leverage data that already exists to ultimately enable an AI application. AI is only as powerful as the information that it is fed. A virtual assistant for an insurance company, for example, must understand what to do when it learns a customer had a major life event. “That” process is determined by data that likely already lives inside the organization’s infrastructure. With the right analytics in place, an organization is well positioned to deploy a powerful AI solution, fed with the right information that will enable successful implementation.

The past few years have brought an unprecedented wave of chatter about the impact artificial intelligence will have on our world and livelihoods. But this sphere of technology has been in existence for over 50 years. To what do you attribute this new focus on Artificial Intelligence—to what do you attribute this new focus on AI?

We’ve spent the last 20+ years putting in place the core building blocks necessary to allow for the sudden accelerated growth of Artificial Intelligence (AI).  Over the last few decades we’ve increased compute, storage and networking capabilities to the point where we now have the core infrastructure for the “intelligent edge- intelligent cloud” computing model.   This model provides the three essential ingredients needed to scale AI.

Cloud – The cloud provides the compute capacity necessary to allow all developers and data scientists the scale needed to build and run their own machine learning (ML) models and AI Solutions.  Prior to the cloud only developers at institutions with compute/storage/networking available at scale could work on AI/ML problem sets. 

Data – The growth of data from intelligent devices and sensors at the edge, and from human generated data from social networks and business applications.

Algorithms – While we’ve had algorithms (and tools/languages) available for decades the rapid growth of tools (Cognitive services, Bot Frameworks, etc.) has made AI useable by mainstream developers and Machine Learning tool sets for data scientists means the broad community of solutions can be infused with AI. 

What applications of Artificial Intelligence can businesses realistically implement into their organizational models now? 

• In improving customer experiences and product value?
• In automating internal workflow to increase efficiency?  
• In gathering actionable insights from big data to drive strategic decision making?

Today business can infuse their existing business solutions with artificial intelligence to address everything from the user experience to individual and team productivity.  Whether it’s new input capabilities (speech, inking, gestures), new interaction models (business/virtual agents) or new business approaches (AI driven forecasting) traditional business experiences can be completely re-thought.  At the same time there is a new generation of solutions being created for every industry and user…from healthcare to education, to construction to mining, every facet of every business is being re-thought.   In the consumer space AI is being used to break down barriers to language, location and access.  AI running in the background is enabling the continued enablement of a globally diverse, yet connected and accessible world.  

ML works well at finding patterns in large datasets that are not easily identifiable by humans. Applying ML to quality control scenarios or identifying irregularities are good uses of it. For example, network operators can use ML to analyze network traffic and make adjustments in real-time so that users get a great call experience.  This can just as easily be used to identify patterns in customer support, in products and in processes that can be used to improve customer, employee and partner experiences.

A favorite example of mine comes from agriculture. Most people think of agriculture as a low technology business. Do you know which country produces the most milk per cow? Israel. By putting pedometers on cows, researchers were able to analyze them to figure out when was the best time to breed cows or to maximize the milk produced. Once you start gathering data the real limitation on strategic decision making comes from not being open to asking the right questions.  This type of thinking can be used for agriculture as well in terms of crops, fertilization, seeding and more. 

For companies looking to implement any of these applications of AI into their business model, what should they be doing internally to prepare for this digital transformation? 

Any organization that is looking to apply AI to their business model should start with the customer. It is tempting to apply AI to everything, but it is best to start small, make an impact and then grow from there. I like to suggest companies focus on three things

1. Build a simple business agent using cognitive services and the bot framework to get comfortable with the new generation of tools in this space
2. If you have unique data assets and data science capabilities use the new generation of ML tools to enhance a business process or insight using a blend of your unique data and publicly available data sets
3. Focus on one solution that is broadly applicable to your business, likely in partnership with a company or service integrator (SI) that has expertise in this space to get a sense for what’s involved in building real-world AI solutions

Overall doing nothing, with a wait and see attitude will put you behind in this space, but trying to apply it to everything at the same time will put you past what your capabilities are and what the technology will support today.  A measured and deliberate approach will provide the highest chance of success.