Yasser Khan, is the CEO of ONE Tech an AI-driven technology company that designs, develops and deploys next-generation IoT solutions for OEMs, network operators and enterprises.
What initially attracted you to Artificial Intelligence?
A few years ago, we deployed an Industrial Internet of Things (IIoT) solution that connected many assets across a wide geographic location. The amount of data that was generated was immense. We aggregated data from PLCs at sampling rates of 50 milliseconds and external sensor values a few times a second. Over the course of a single minute, we had thousands of datapoints being generated for each asset we were connecting to. We knew that the standard method of transmitting this data to a server and having a person evaluate the data was not realistic, nor beneficial for the business. So we set out to create a product that would process the data and generate consumable outputs, greatly reducing the amount of oversight that an organization needs in order to reap the benefits of a digital transformation deployment—heavily focused on asset performance management and predictive maintenance.
Can you discuss what ONE Tech’s MicroAI solution is?
MicroAI™ is a Machine Learning platform that provides a greater level of insight into asset (device or machine) performance, utilization and overall behavior. This benefit ranges from manufacturing plant managers who are looking for ways to improve overall equipment effectiveness to hardware OEMs who want to better understand how their devices are preforming out in the field. We accomplish this by deploying a small (as small as 70kb) packet onto the microcontroller (MCU) or microprocessor (MPU) of the asset. A key differentiator is that MicroAI’s process of training and forming a model is unique. We train the model directly on the asset itself. Not only does this allow for data to stay local, which reduces cost and time of deployment, but it also increases the accuracy and precision of the AI output. MicroAI has three primary layers:
- Data ingestion – MicroAI is agnostic to data input. We can consume any sensor value and the MicroAI Platform allows for feature engineering and weighting of the inputs within this first layer.
- Training – We train directly within the local environment. The training duration can be set by the user depending on what a normal cycle of the asset is. Typically, we like to capture 25-45 normal cycles, but this is heavily based on variation/volatility of each cycle captured.
- Output – Notifications and Alerts are generated by MicroAI based upon the severity of the anomaly that is detected. These thresholds can be adjusted by the user. Other outputs generated by MicroAI include Predicted Days to Next Maintenance (for optimizing service schedules), Health Score, and Asset Life Remaining. These outputs can be sent to existing IT systems that the clients have in place (Product Lifecycle Management tools, Support/Ticketing Management, Maintenance, etc.)
Can you discuss some of the machine learning technologies behind MicroAI?
MicroAI features a Multi-Dimensional Behavioral Analysis packed within a recursive algorithm. Each input that is fed into the AI engine impacts the thresholds (upper and lower bounds) that are set by the AI model. We do this by providing a one step ahead prediction. For example, if one input is RPMs and RPMs increase, the upper-bound threshold of bearing temperature may go up slightly due to the faster machine movement. This allows the model to continue to evolve and learn.
MicroAI is not reliant on accessing the cloud, what are the advantages of this?
We have a unique approach to forming models directly on the endpoint (where data is generated). This brings data privacy and security to deployments because data does not need to leave the local environment. This is especially important for deployments where data privacy is mandatory. Furthermore, the process of training data in a cloud is time consuming. This time consumption of how others are approaching this space is caused by the need of aggregating historical data, transmitting data to a cloud, forming a model, and eventually pushing that model down to the end assets. MicroAI can train and live 100% in the local environment.
One of the features of the MicroAI technology is its accelerated anomaly detection, could you elaborate on this functionality?
Due to our approach of behavioral analysis, we can deploy MicroAI and instantly start learning the behavior of the asset. We can start to see patterns within the behavior. Again, this is without the need of loading any historical data. Once we capture sufficient cycles of the asset, we can then begin to generate accurate output from the AI model. This is groundbreaking for the space. What used to take weeks or months to form an accurate model can happen within a matter of hours, and sometimes minutes.
What’s the difference between MicroAI™ Helio and MicroAI™ Atom?
MicroAI™ Helio Server:
Our Helio Server environment can be deployed in a local server (most common), or in a cloud instance. Helio provides the following functionality: (Workflow management, data analysis and management, and data visualization).
Workflows for managing assets – A hierarchy of where they are deployed and how they are used. (e.g., setup of all customer facilities globally, specific facilities and sections within each facility, individual stations, down to each asset in each station). Furthermore, the assets may be setup to perform different jobs with different cycle rates; this can be configured within these workflows. In addition is the ability for ticket/workorder management, which is also a part of the Helio Server environment.
Data analysis and management – Within this section of Helio, a user can run further analytics on the AI output, along with any raw data snapshots (i.e., Max, Min, and average data values on an hourly basis or data signatures that triggered an alert or an alarm). These can be queries that are configured in the Helio Analytics designer or more advanced analytics brought in from tools such as R, a programming language. The data management layer is where a user can utilize the API management gateway for 3rd-party connections that are consuming and/or sending data in coordination with the Helio environment.
Data Visualization – Helio provides templates for various industry specific reporting, which allows for users to consumer Enterprise Asset Management and Asset Performance Management views of their connected assets from both the Helio desktop and mobile applications.
MicroAI Atom is a Machine Learning platform designed for embedding into MCU environments. This includes training of the multi-dimensional behavior analysis recursive algorithm directly in local MCU architecture—not in a cloud and then pushed down to the MCU. This allows for accelerating the build and deployment of ML models through the auto-generation of the upper and lower thresholds based on multivariant model that is formed directly on the endpoint. We have created MicroAI to be a more efficient way of consuming and processing signal data to train models than other traditional methods. This not only brings a higher level of accuracy to the model that is formed but utilizes less resources on the host hardware (i.e., lower memory and CPU usage), which allows us to run in environments such as an MCU.
We have one other core offering called MicroAI™ Network.
MicroAI™ Network – Allows for a network of Atoms to be consolidated and mashed with external data sources for creating multiple models directly at the edge. This allows for horizontal and vertical analysis to be run on the various assets that are running Atom. MicroAI Network allows for an even deeper level of understanding to how a device/asset is performing in relation to similar assets that are deployed. Again, due to our unique approach to forming models directly at the edge, the machine learning models consume very little memory and CPU of the host hardware.
ONE Tech also offers IoT security consulting. What’s the process for threat modeling and IoT penetration testing?
Due to our ability to understand how assets behave, we can consume data related to the internals of a connected device (e.g., CPU, Memory Usage, data pack size/frequency). IoT devices have, for the most part, a regular pattern of operation—how often it transmits data, where it sends the data, and the size of that data packet. We apply MicroAI to consume these internal data parameters to form a baseline of what is normal for that connected device. If an abnormal action occurs on the device, we can trigger a response. This can range from rebooting a device or opening a ticket within a workorder management tool, to completely cutting network traffic to a device. Our security team has developed testing hacks and we have successfully detected various Zero-Day attack attempts by using MicroAI in this capacity.
Is there anything else that you would like to share about ONE Tech, Inc?
Below is a diagram of how MicroAI Atom functions. Starting with acquiring raw data, training and processing in the local environment, inferencing the data and providing output.
Below is a diagram of how MicroAI Network functions. Many MicroAI Atoms feed into MicroAI Network. Along with the Atom data, additional data sources can be merged into the model for a more granular understanding of how the asset is performing. Furthermore, within MicroAI Network multiple models are formed allowing stakeholders to run horizontal analysis on how assets are performing in different regions, between customers, before and after updates, etc.
Thank you for the interview and your detailed responses, readers who wish to learn more should visit ONE Tech.
Dimitris Vassos, CEO, Co-founder, and Chief Architect of Omilia – Interview Series
Dimitris Vassos is the CEO, Co-founder, and Chief Architect of Omilia, a global conversational intelligence company that provides advanced automatic speech recognition solutions to companies and organizations in North America, Canada, and Europe. Dimitris has significant experience in the field of applied speech and artificial technology, specifically, natural language understanding (NLU), speech recognition, and voice biometrics.
What initially attracted you to AI?
Human-Machine interfaces have mesmerized me since I was a child. In 1984, I had one of the first home computers. I remember I had programmed it to control our home lighting using sound recognition. Back then, there was no speech recognition technology available, so I had it recognize patterns of sound (i.e. clapping).
During my studies, I learned about speech signal processing, and immediately saw this as the most promising human-machine interface technology. I had a mission to make machines understand humans using natural language understanding.
What was the inspiration behind founding Omilia?
My first job was with IBM in its voice solutions unit. I was developing and delivering voice interaction solutions for automated telephone inquiries. Very soon, I realized that technology and products available could be made better. After a few years of frustration, I decided to take things into my own hands. That’s when I met my partner, Pelias, and we founded Omilia with a mission to re-invent the voice automation industry.
What are some of the solutions currently offered by the Omilia Conversational AI?
Omilia provides human-like human-to-machine communication experiences and technologies in to help companies improve the customer care experience. Our main solution offerings are:
- Conversational AI Self-Service for Customer Care, over both voice and text channels
- Customer ID & Authentication using Voice Biometrics
- Transcription Services focused on Customer Service
The Omilia Cloud Platform (OCP) offers ready-to-consume micro-services that enable companies to integrate our state-of-the-art capabilities into their customer care processes, while offering critical solutions for those companies that wish to benefit from our team’s extensive experience in deploying large-scale conversational AI solutions for over 15 years now.
Our differentiator in the market is that our capabilities have evolved over time and with that experience are ahead of the curve, offering robust solutions that deliver real business value.
Could you discuss some of the speech recognition challenges that may be faced from having a global clientele speaking in different languages and with different accents/pronunciations?
A language is a system of communication defined by people and influenced by their country or community. As such, accents and pronunciations become a significant part of any language. So, developing a language for AI technology involves collecting a large sample of spoken language from a group of native speakers and using it for training our AI models. It requires considerable effort to perform this process. Our diverse clientele propelled us to strategically invest in building out capabilities in various languages. Today, our solutions support over 24 languages, and the upfront investment and assets that we have accumulated are very significant.
Why do you believe that voice biometrics is important?
In customer service, perhaps the most challenging part of a caller’s experience is authentication – the process by which a customer representative verifies the identity of the caller. We all have called in various contact centers and have spent time answering trivial questions that are designed to prove our identity. One facet of the problem is that today, such personal information is easily discoverable by anyone making security of authentication processes alike compromised. At the same time, the authentication process itself is often very tedious, lengthy and adds frustration to the customer experience.
Voice Biometrics solves this problem. While one’s voiceprint is not 100% unique and secure, it is significantly more accurate and secure than knowledge-based information. When done right, voice biometrics can provide a more satisfactory customer experience. For me, “done right” is defined as seamless for the customer.
At Omilia, we specialize in adapting the technology to the customer. Good tech is invisible tech. This is the design principle that guides our products and services. With our deepVB solution, we are achieving ultra-high accuracy rates, from the first few seconds of the customer speaking, and without any need for the customer to say anything specific. It is an innovative breakthrough that is solving the problems of traditional biometric deployments with tedious enrolment processes and lengthy speech samples. Omilia’s Voice Biometrics works seamlessly across all voice channels.
Voice Biometrics is an important ingredient in the customer service experience, making the authentication process seamless, and thus improving overall customer experience and business value.
Voice biometrics has often been poorly deployed in the marketplace, how does Omilia tackle this challenge?
The most common mistake when deploying voice biometric services is designing the customers’ experience around the limitations of the technology. This is wrong. If customers were to use the technology exactly in the way it was designed, without deviations, it would perform well, but the reality is that customers have their own objectives they are set to accomplish on a call and technology can either meet those objectives, or not. The solutions must be designed around the customer experience, not the other way around.
So far in the market, most biometric deployments in customer service have not delivered the benefits they promised. In contact centers, most deployments have relied upon legacy technology, which hasn’t kept pace. As with voice self-service, at Omilia we persistently refuse to settle for anything less than stellar customer experience. So, we set out to define what such an experience would look like and what the technology enabling that experience should be able to do. The result is our deepVB engine, which is at the core of our services and solutions, and which not only leverages the latest in Deep Learning, but also works hand-in-hand with our Speech Recognition engine and Dialogue Manager, to smartly get informed from the context of the interaction. We have solved the Conversational Authentication problem, not just voice biometrics.
Being able to understand intent in speech analytics can be tricky, how is Omilia able to do this so successfully?
The ancient Greeks had a saying “good things are acquired by expending labor”. This is true for everything we do at Omilia, and it is certainly true for our Natural Language technology. For more than 15 years now, we have been painstakingly designing, building and refining a unique approach to Natural Language Understanding (NLU), which is not a singular technology, but rather a combination of multiple technological approaches, all fused together to deliver Conversational NLU for Customer Service. Conversational AI companies are continuously looking for the best Machine Learning algorithm to solve for NLU. However, most of the time, what makes a difference is not the best algorithm in the lab. What has rendered results for us is the approach of going the extra mile to find solutions to real world problems, not technological ones, utilizing the best tools.
I often remind our team of how long it took us to get here. And develop the technological capabilities we have. We were successful because we were sought out to solve the problems that companies and their customers are experiencing, as opposed to simply looking at the technology.
Conversational AI is becoming a hyper competitive industry, what do you believe differentiates Omilia from competing products?
The field of Conversational AI is very broad and is applied in different ways in different markets and industries. In practice, only a small handful of providers can deliver sophisticated solutions that can compete with the most advanced solutions in the space in which Omilia is active. Omilia is one of those providers.
Omilia’s biggest strengths are voice and dialogue. We can facilitate real-time, unconstrained, conversational dialogues with customers over various communication channels, with ultra-accuracy and stellar customer experience. We have been doing so for many years, and our experience has positioned us to achieve the highest number of conversational voice self-service solution deployments in the world to this day.
The “Forrester New Wave” 2019 report assessing customer service in Conversational AI found that Omilia’s strongest product differentiation is its omnichannel, voice and speech, vertical specialization (pre-trained models for specific industries) and security and authentication which enables seamless identification and authentication of customers, which otherwise is a point of high friction for customer service.
What type of enterprises are using Omilia products?
Omilia works in over 16 countries serving global companies in sectors with a large volume of customer service interactions such as finance, telecom, insurance, healthcare, energy, retail, utilities, public sector and transportation, to name a few. Omilia’s ability to achieve 96% semantic accuracy, across 24 ASR languages and do so at scale while ensuring consumer fraud prevention, makes Omilia one of the few companies that can meet the high customer service demand with human-like conversational experience, saving companies millions of dollars.
Omilia’s products fit all enterprises, big or small. We are serving mid-market customers with our cloud-based offering of OCP miniApps®, which we consider a real disruption in the Conversational AI space. We utilize sophisticated conversational voice and text self-service for a company of any size, at the click of a button and with zero-coding. Now, any company can provide the highest level of sophistication in customer service that previously only the large enterprise customers could afford.
Thank you for the interview, readers who wish to learn more should visit Omilia.
Dr. Danny Lange, Senior VP of AI at Unity Technologies – Interview Series
Dr. Danny B. Lange is VP of AI and Machine Learning at Unity Technologies. Formerly, Danny was Head of Machine Learning at Uber where he led an effort to build the world’s most versatile Machine Learning platform to support Uber’s rapid growth.
What initially attracted you to Artificial Intelligence?
I built and programmed computers from a very young age and I was almost immediately fascinated by the idea of making these systems autonomous. What captivates me about autonomy is the challenges you as a developer have to overcome in creating a system made from sequences of rigid code that can safely respond to unpredictable and never-seen-before circumstances. The field of Artificial Intelligence (AI) has over the years provided us with increasing powerful tools from object-oriented programming, rule-based inference, to machine learning and more recently deep learning. It is the increased capabilities of these technologies that fuel the rapid progress in the field of AI.
You have been a leader in the space for many years such as being General Manager for Amazon Machine Learning in AWS, and Head of Machine Learning at Uber. What are some of the lessons that you have learned from these past experiences?
Machine learning is a truly transformative technology, but to realize the potential to its fullest, it is necessary to bring it to every corner of the enterprise. Repeatedly, machine learning has demonstrated its ability to create unimaginable optimizations and lift business operations to levels that cannot be achieved by ordinary human processes alone. However, true disruption only occurs when a critical mass of business processes are operated in this way. What organizations such as Amazon and Uber clearly have demonstrated is that if we make the machine learning systems broadly approachable and available to every team we experience a broad adoption that invariably leads to a virtuous cycle of continued improvements to the overall business as a network effect takes place
You’ve been the VP of AI at Unity Technologies since 2016. What was it about this company’s vision that excited you?
Unity is a fantastic place for the AI enthusiast. We have a remarkable culture of solving hard problems for our customers – with AI being one of the mightiest challenges that i can think of. Our leadership is committed to power and drive the future of AI. We have the technologies, resources, customers, and partners to do just that. I cannot imagine a better place to work on changing the world.
You’ve spoken before about the importance of synthetic data, could you share with us what this is precisely?
Synthetic data is created by an algorithm as opposed to data captured from the real world. A real-time 3D engine with a realistic physics emulator, is the ideal tool to create realistic yet synthetic training data for a wide variety of applications ranging from object recognition in computer vision systems to path planning for navigational robots.
What makes synthetic data so important when it comes to building machine learning systems?
I have on many occasions called Unity the perfect AI Biodome. And it is true. Working with AI in the real world and using real-world data can be outright scary. Do I have to mention self-driving vehicles on the streets of San Francisco or face recognition systems deployed in public spaces? There are worries about safety, bias is always lurking, and privacy concerns often collide with common use cases. And then there is the scarcity and high cost associated with collecting the necessary amounts of training data. With Unity, we have not only democratized data creation, you also have access to an interactive system for simulating advanced interactions in a virtual setting. Within Unity you can develop the control systems for an autonomous vehicle without the risk of hitting and injuring anyone.
Can you discuss how Unity Simulation can assist companies with the generation of synthetic data?
With Unity Simulation we have taken a real-time 3D engine designed for human consumption whether that is gaming, film, or engineering – and turned into an cloudoptimized instance that not only runs at unimaginable high frame rates, but also allows for scaling to thousands of instances running in parallel. In this way Unity Simulation allows developers to generate experiences for their AI systems orders of magnitude faster than wallclock time. Until recently, this scale of data generation was only available to a few privileged corporations, but with Unity Simulation we have truly levelled the playing field.
Recently, Unity has teamed up with The Institute for Disease Modeling (IDM) to build real-time 3D in-store simulations that model COVID-19 spread. Can you discuss how Unity Simulation can effectively simulate the spread of COVID-19?
Computer simulation has been used for decades by researchers, engineers, problem solvers, and policy makers in many fields, including the study of infectious disease. Unity Simulation enables a special kind of real-time spatial simulation that can be scaled on the cloud to holistically study large, complex, and uncertain systems. We built a simplified demonstration project to simulate coronavirus spread in a fictitious grocery store and explored the impact that store policy has on exposure rates. By running tenth of thousands of simulations, we were able to identify the behaviors and policies that appeared to have the greatest impact on the spread of this terrible infectious disease.
Recently, you’ve been speaking a lot about Artificial General Intelligence (AGI). Can you explain what emergent behavior is and why it’s important for the development of AGI?
In just 100,000 years, the human race went from surviving on picking berries in the wild to putting a person on the moon. We know from archaeogenetics that the human brain has not changed significantly during that period of time. You can say there were no significant hardware upgrades to the processor. So what was it then that was so transformative? The key should be found in our ability to accomplish something together. We use the term emergent behavior of a system that does not depend on its individual parts, but rather on their relationships to one another. Emergent behavior cannot be anticipated by investigating the individual parts of a system. It can only be predicted by understanding the relationships between the parts. Emergent systems are characterized by the observation that the whole is greater than the sum of the parts. While I have repeatedly shown entertaining examples of emergent behavior in relatively simple multi-agent systems, just imagine what will happen when you have a plethora of AI systems collaborating at the speed of light.
Do you believe that there is a possibility that we can achieve AGI within the next decade?
When it comes to AGI we have to remember that it is all about the journey and not the destination. Nobody knows exactly when AGI will happen as it will not be at a specific moment in time, but rather a gradual change over time. It is in the nature of AGI that it will be hard for us humans to pinpoint just how intelligent a system at any given moment. Looking at the progress made over the last decade, I am sure that this decade will bring us plenty of interesting progress towards AGI.
Is there anything else that you would like to share about Unity Technologies?
At Unity, we continue to see ourselves powering the future of AI and playing a significant role in the advancement of AI technologies. As our relationship with DeepMind is a clear demonstration of our technology is the perfect environment for researchers and developers to safely push the boundaries of AI. We are gearing up to support our customers and partners in creating virtual environments that operate at previously unseen scale to solve the challenges of tomorrow whether that is climate change, logistics, or health challenges.
Thank you for the amazing interview, I enjoyed learning about Unity and your views on AGI. Anyone who wishes to learn more should visit Unity Technologies.
Dr. Lingjia Tang, CTO and Co-Founder, Clinc – Interview Series
Dr. Lingjia Tang, CTO and Co-Founder of Clinc, is a professor of Computer Science at The University of Michigan. Dr. Tang’s research in building large-scale production infrastructure for intelligent applications is widely recognized and respected in the academic community. In addition to working at both Microsoft and Google, Lingjia received her PhD in Computer Science from the University of Virginia. Lingjia has recently received prestigious awards including ISCA Hall of Fame, Facebook Research Awards and Google research Award.
What initially attracted you to AI? When did you first discover that you wanted to launch an AI business?
In the mid-2000s I was performing research around large-scale systems that support various applications and how we can design servers as a software system to run those applications more efficiently. At the time, we were shifting from working with traditional web applications to more machine learning-driven functions. That’s when I started to pay attention to the algorithms associated with AI and gained interest in fundamentally understanding how AI applications work. Soon after, the research team I was working with decided to pivot and basically build our own AI applications as benchmarks to study, which is what led us to publishing our first few research papers and developing our first product, Sirius—an open end-to-end voice and vision personal assistant.
As an open source software, Sirius allowed people to build conversational virtual assistants on their own. At the time, this was a very limited capability for the general public and was really only controlled by the big companies, such as Google and Apple. However, we saw that we were filling a critical gap when we released the software and saw that it had tens of thousands of downloads in the first week! That was the turning point where we knew there was a lot of market demand for this type of software.
Come 2015, we launched Clinc with the mindset that we would be able to provide everyone – every developer, company, person—who wants to be able to build a virtual assistant with the access to expertise, tooling and innovation to be able to do that.
Clinc offers conversational AI solutions without relying on keywords or scripts. Could you go into some details regarding how this is achieved? What are some of the Natural Language Processing (NLP) challenges that had to be overcome?
What really sets Clinc apart from other conversational AI platforms on the market is its underlying AI algorithms that enable its “human in the room” experience, which understands messy and unscripted language. This allows for corrections to backtrack and “heal” mistakes made in human conversation and enables complex conversational flows—conversations that a real human would be able to understand. In contrast to a speech-to-text word matching algorithm, Clinc analyzes dozens of factors from the user’s input including wording, sentiment, intent, tone of voice, time of day, location and relationships, and uses those factors to deliver an answer that represents a composite of knowledge extracted from its trained brain. For example, if I ask my virtual assistant, “how much money did I spend on a burger?” it needs to understand that I am asking about money and spending, that I am asking specifically about a hamburger and that a hamburger is a type of food and should be matched to my recent spending at a restaurant.
Achieving this level of understanding is not easy. In general, we would break down conversational AI into two components: Natural Language Understanding (NLU) and dialog management. So, the challenge that we had to overcome was figuring out how to build a system that can extract key pieces of information accurately and can anticipate what the user is asking.
We are able to do this through sophisticated, contextual, top-down NLU, that is trained to be intuitive in nature to keep up with the natural flow of conversation, understanding slang and context. This is in comparison to competitive solutions that have a top down, rules-based approach to Natural Language Processing (NLP) that does not allow for conversational healing, meaning if the customer makes an error, the competitive solutions make them go back to square one, wasting time and only frustrating the user. We also use crowdsourcing to extract our language data to create a richer, diverse data set that can be immediately used to train AI models.
Could you discuss how deep learning is used with the Clinc AI system?
Clinc is using a hybrid approach to deep learning where we use the traditional old-school model to some degree and leverage deep learning where needed. Specifically, we use deep learning to understand words and languages to determine the dialogue flow. Generally, our entire dialogue is a combination of deep learning and symbolic AI. We don’t use deep learning for language generation yet because, when it comes to our customers which are primarily in the banking industry, there are a lot of regulations that the virtual assistant must follow that dictate what they can and cannot say to their customers. So, there is still a lot of uncertainty around whether or not the deep learning will be able to follow those set language restrictions.
As of right now, I don’t think the conversational AI community is completely ready to fully adopt deep learning whereas the academic community is 100% all in, but I do look forward to seeing what the new models can do.
What’s the process for a company that wishes to customize the AI’s responses to target a specific audience? Could you give some examples of how Clinc is currently being used by clients?
We allow clients to either license a platform they can build on however they like, or take our fully built and trained chatbot, Finie, and customize it and integrate it into their apps or messaging services. Finie can handle matters related to balances, transactions, spending history, locating an ATM, making a transfer and more.
My favorite example of how a client has customized Clinc’s AI to target a specific audience is İşbank. As Turkey’s largest private bank, they turned to us to develop their digital banking assistant, Maxi, back in 2018. To infuse Maxi with a unique personality, İşbank held 14 focus groups to gauge what sort of traits and skills bank customers wanted in a virtual assistant. They also hired a voice actress to recite sentences in Turkish related to banking tasks. İşbank’s conversational banking team came up with these sentences by considering the way real people would phrase their needs. Upon our recommendation, the team paid participants on crowdsourcing marketplaces such as Amazon Mechanical Turk to supply different ways they might express the same questions, such as a request to view their balances (“what is my balance,” “how much money do I have in my account,” “show me the cash in my account”) or pay a bill (“pay my bill,” “bill payments”).
This example really shows how invested İşbank is in offering a digital banking assistant to help their customers better navigate their accounts. With Clinc, İşbank launched Maxi to more than 7.5 million people, in Turkish. Since its launch, İşbank has seen widespread adoption by more than 5.5 million users, with an average of 9.8 interactions per user. In recent months, as COVID-19 cases increased in Turkey, İşbank swiftly trained Maxi to be responsive to COVID-19-related queries. Since March 2020, Maxi has answered more than 1.2 million customer queries related to COVID-19, a more than 62% increase in usage.
What would you tell women who are interested in learning more about AI but are reluctant to get involved due to it being a male dominated field?
Off the bat, I don’t think there is any reason why AI is considered a male-dominated field. I think there are a lot of women pioneers in AI that are doing really well and are making an impact. I think AI coupled with social policy is a unique area that has the potential to have a lot of impact on people’s everyday lives. This is where I do think more diverse insights across the board would really benefit us, especially since there are a lot of conversations around AI bias involving race and gender. I believe that having a scoped community of AI developers will continue to have a disproportionate impact on society and policy.
For the women out there who are interested in joining the AI field, I highly recommend it especially if you are interested in making an impact. AI has had so much growth and innovation over the years and it really is an exciting time to be a part of it.
Is there anything else that you would like to share about Clinc?
Clinc is making huge strides right now. Personally, I have just stepped into a new role as CTO of Clinc and I am really excited to focus on how we can further work with developers and data scientists to grow the reach of our technology. As I look toward the future, I see the demand for AI-powered applications shifting to enable people who don’t have years of data science experience and machine learning background to be able to use it too. For example, you don’t have to have a graphic design degree to be able to use Photoshop. I think AI is heading in that direction where developers with no AI or machine learning training will be able to achieve results and produce high quality applications. Overall, we want to reiterate that we are not only devoted to the end-user but also to the developers, no matter what level, who show interest in our solution.
Thank you for the great interview, I look forward to followin your progress. Anyone who wishes to learn more should visit Clinc.
- Dimitris Vassos, CEO, Co-founder, and Chief Architect of Omilia – Interview Series
- Human Brain’s Light Processing Ability Could Lead to Better Robotic Sensing
- Game Developers Look To Voice AI For New Creative Opportunities
- Udacity Launches RPA Developer Nanodegree Program in Conjunction with UiPath
- AI Used To Identify Gene Activation Sequences and Find Disease-Causing Genes