Tell us about your interaction with smart technologies like AI ML, and Automation for contact centers.

AI and ML are near and dear to our hearts at Talkdesk. We have pushed AI as a layer in Talkdesk architecture where it empowers all Talkdesk products. We apply advanced Machine Learning to address practical problems in contact centers on a daily basis and try new ideas without any fear.

How did you start in this space?

I joined Takdesk about two years ago. My goal was two-fold 1) Make Talkdesk a totally intelligent and automated contact center and 2) Turn Talkdesk into a data-focused and data-driven company. I, along with several of my colleagues on the Talkdesk leadership team, came from outside of the contact center industry. This out-of-the-box perspective from the Talkdesk team is infusing a new and creative approach into an established industry.

What is Talkdesk and how does it help workforce transformation?

Talkdesk is an enterprise cloud contact center that empowers companies to make the customer experience their competitive advantage. With enterprise-class performance and consumer-like ease of use, Talkdesk easily adapts to the evolving needs of Sales and Support teams and their end-customers, resulting in increased customer satisfaction, productivity and cost savings.

The core tenets of the Talkdesk approach are:

• Powerfully Simple: Get started fast with an implementation that is done in weeks, not months. The consumer-app style interface is so intuitive, agents and reps can get started with little-to-no training.
• Seamlessly Connected: Provide access to the information and tools agents need to better serve customers.
• Endlessly Adaptable: Instantly leverage the new capabilities Talkdesk delivers in three updates per year. Make changes quickly with clicks, not code.
• Enterprise-Class: Talkdesk offers the scalability, reliability, and security required by large global organizations.
• Actively Intelligent: With Talkdesk iQ, companies can leverage the power of Artificial Intelligence to drive better customer engagement, contact center operations, and decisions.

This last tenet is, of course, leads directly into our discussion today. Talkdesk is an intelligent contact center. The word intelligent is very crucial here. Human Intelligence is innate; it’s baked into our architecture and we cannot separate it. Intelligence grows, matures, and — hopefully— improves over time. This evolutionary approach to intelligence is applied to the products we build at Talkdesk. The same philosophy and strategy. We have baked AI into our company DNA; in every module of our product, from the backend all the way to the user-facing modules. AI should not be a feature to add on – it must be the core of a contact center operation as it improves routing, agent coaching, agent assist, speech science, workforce optimization, workforce management, workforce engagement, omnichannel optimization and virtually all aspects in a contact center.

How does Talkdesk bring AI to the center of retail customer experience?

One of the main challenges for retailers is providing an Omnichannel experience, brand recognition, brand loyalty and to improve the overall customer experience. A modern contact center should be able to assist with this mission. Customer experience should be channel agnostic and whether customer calls, emails or sends a text they should experience the same quality of service. Studies have shown that even one bad experience might change a customer’s view or stop a purchase. Retail itself is very competitive.

In such a brutal environment that has seen the demise of well-known and stable brands such as Circuit City, ToysRUs, Robinsons May, Borders, and Sport Chalet, every single interaction matters in both the short and long term. Contact centers play an extremely important role in generating, maintaining and strengthening brand loyalty. Infused throughout the platform, Talkdesk iQ mines billions of interactions to reveal insights and trends and drive predictive recommendations, better customer engagement, and improving contact center operations, resulting in better customer experience.

What is your opinion on the use of personal data for building AI ML models, and then using the same technology to chase/follow customers without their knowledge?

There are some factors to be considered here:

• AI and Machine Learning are built based on data. Without data, AI cannot do much. New frontiers in Machine Learning such as Deep Learning algorithms are data intensive. They need lots of data points to learn, which is to say that AI without data is useless.
• Privacy should be at the heart of any data science model, flow, product, and operation. Data, by definition, belongs to the individual who owns this. Privacy-preserving AI has always been the highest priority within the Machine Learning community as researchers take this issue seriously.
• Communication with the user is key. If the user understands that by providing data to a Machine Learning engine they may enjoy some benefits through intelligence-enabled customer experience, then they might be willing to contribute. Medical informatics is a good example of this issue. Algorithms such as advanced machine learning to detect tumors from X-rays in the early stages needs mountains of data to learn and improve. Participation in these data and information sharing projects could literally save lives.

Which technologies are involved in the Digital Transformation of Contact Centers and Customer Care Support companies?

Digital Transformation has been a dream for contact centers for many years. Moving from an on-premises, hardware-based type of technology to a cloud software-based type of technology needs lots of work. We cannot deny that some of the players in the Contact Center as a Service (CCaaS)  field have added some cosmetic surgery on the top of their platform to help retain clients in the short-term. Sooner than later, the cracks become apparent and they will lose it to the digital wave. In fact, this is already underway. The following are the most important blocks you need to have for Digital Transformation (DT).

• You need a solid infrastructure that provides data at any access point to the platforms to interact with customers and agents. This is probably the most essential step toward DT.
• Your architecture must be scalable. Unfortunately, “contact center industry” and “outage” are very familiar terms and appear together frequently. One of the main reasons is the lack of scalability capabilities by the legacy, on-premises platforms.
• Data must be captured and analyzed from all channels including speech, text, video, user behavior, actions, and interactions.
• Your contact center needs a comprehensive infusion of Artificial Intelligence.

An on top of everything, you need to define your strategy and make sure your organization goes through such cultural change to accept, understand and leverage AI capabilities.

As Head of AI, what kind of skill you train for? What skills do you look for in your team?

We are looking for bright, focused and hard-working individuals with the following expertise:

• Core Data Science team. This is a group of hardcore machine learning specialists, data scientists, and data mining engineers. Their main job is to build AI modules that address important practice problems.
• Data team. The data team’s main job is to deliver a data platform to be consumed by all other teams. The data team is specialized in data architecture, database infrastructure design, database optimization and building fast data platforms.
• Product manager(s). This is one of the key components of any Data Science team, connecting the Data Science team to the company product and to the business team. The role of this team is to get the scope of the product aligned with the company strategy and product roadmap and to make sure the Data Science product is at the heart of the roadmap on a monthly basis.
• Business Intelligence (BI) specialists and BI engineers. The BI team role is already known and has been widely discussed. We chose to be a proactive BI team. For business owners, there are always some gray areas that they need to be informed by a data-oriented BI team. Our BI team not only delivers what is needed for business decisions but they also proactively find new insights to be delivered to different segments of the company.
• You might ask, “Where are the software engineers?” Great question.

At the core of every data science team are experienced, software engineers. This team is the conductor of the Data Science symphony. They make the pipeline, define the right software architecture and eventually deliver the product. Without the right software engineering team, you might end up building nice prototypes that never get to the product line. To transition our prototypes to reality, we teamed up with our production team – engineers who fight on the front line. Each product team at Talkdesk is helping us to build AI as part of their own product.

Tell us more about your approach to building real-time ML models. How do you see the involvement of AI software and Automation in such models?

The era of a typical Machine Learning module needing 24 to 48 hours to crunch data and provide insight is over. This is still the case for some of the scientific discoveries, but in a contact center world where users expect real-time experience, ML has to be real-time or near real-time when it’s needed. Data Science architectures mush design the flow of the data processing and ML algorithms carefully in order to mine data off-line to learn models and then apply such knowledge in real-time. In addition, real-time learning for some use cases is inevitable.

What is your opinion on “Weaponization of AI/Machine Learning”? How do you promote your ideas?

I spent several years on the application of AI and social network analysis in counter-terrorism and fraud detection. The “bad guys” got sophisticated over time. They were intelligent, they were learning fast and adapting quickly. We had to design new systems and algorithms to predict and detect fraudulent activities. Intelligence comes with data, context, intention, and goal to aim for. AI systems are no different – these elements are crucial in the future of AI. Having said that, I do believe a world with AI will be a better world. Look at the AI revolution everywhere – soon it will be part of the curriculum from elementary schools through colleges, regardless of the field of study.

Are AI Scientists and Engineers governed by any data privacy regulations/code of ethics? Could you tell us more about it?

Of course, data comes with privacy and regulations and data scientists are the stewards of these protections. There are many regulations such as GDPR that provide a restrictive framework for consumer protection. There are two general directions on this subject.

• Those who make sure about privacy in data and design their algorithm without having access to several data points.
• Those to build signatures of individuals and try to build their models based on these signatures. Advertisers and related entities make a signature of browsing behavior and follow people on different websites using cookies and other types of technologies. While both the AI developing community and advertisers need to follow these regulations, their view to data, provoke and regulations are totally different. This difference in motivation makes the stewardship of privacy protection more important than ever.

What AI ML Research Labs, start-ups and Big Data communities are you keenly following?

Microsoft research labs, Amazon ML works and AI for social goods at Harvard by Milind Tambe (graduate school advisor).

What technologies within Machine Learning, Big Data, and Neural Analysis are you interested in?

I am interested in Deep Learning and Context-aware Machine Learning. Those algorithms that take context to the heart of their inference. Also, I am in favor of those models that provide a macro view of behaviors and then look for micro-level details.

As a tech leader, what industries do you think would be fastest to adopting Analytics and AI/ML with smooth efficiency? What are the new emerging markets for these technology markets?

Any industry which lacks data and is still dominated by old technology will not adapt quickly. Hospitality, travel, fitness, insurance are good examples. Those industries which fundamentally are going through some transformations will adapt faster. Examples include the modern auto industry, healthcare, and education. Also, I think we will see the amazing outcomes of AI in healthcare and education.

What’s your smartest work-related shortcut or productivity hack?

I use data and data analysis in every aspect of my life from minor choices like where to get my coffee, all the way through to making major business decisions. Trust your data and analyze it correctly. Doing so saves significant time and energy in every corner of your life.

Tag the one person in the industry whose answers to these questions you would love to read:

Vipul Patel, Chief Data Scientist at SAP