Our experience isn't (currently) direct, but a couple of the biometric authentication partners that we're using are leveraging deep learning techniques and models to deliver an AI optimized answer on whether a given biometric factor is a match or not.
We are using deep learning to understand the intent of the customer at every touchpoint based previous interactions and other data attributes.
Audit and compliance, fraud prevention and procurement.
Routine tasks, other areas of automation.
We are just beginning and defining the use cases to leverage machine learning to improve patient experience. No current plan for Deep learning in my space.
We are in earlier stages of ML to predict customer attrition, new money calculation and fraud prevention using regression and classification. At this point our focus is on data prep and building a robust pipeline to achieve better results. We are yet to tap into deep learning.
At Foodbot AI we provide services like a loyalty program, table reservations, online ordering, etc. via conversational chatbots over the facebook messenger platform. Contrary to content grouping in mobile apps and websites, in chatbots the user interacts with the Bot via human-like conversations. To predict user intents, sentiments, etc. we use AI and take the respective action. Thus, without any human sitting on the business side we are able to convert our leads into users and make them use the available services.
We use AI/ML to better understand what's occurring in their environment through a rich telemetry stream provided by the customer that enables us to compare their experiences to the 300 trillion data points we have stored spotting problems that others have encountered and closing new ones based on a team of data scientists and subject matter experts. The end result is improved customer experience and insights. Happy to talk to you further.
We are using machine learning to do NLP and entity extraction from customer interactions to build relationship graphs between entities (users, customers, companies, products, deals/opportunities). We are using 3rd party software for sentiment analysis to tag the interactions, personality profiling to tag the users and customers and another third party piece of software to suggest content for users to send to customers. All are based on ML algorithms. We are also developing an engine to recommend the next best interaction between a user and a prospect / customer based on the data and tags resulting from the above. This engine allows external systems to in inject suggested interactions into the candidate set to be evaluated for presentation to the users, along side the ones generated internally by the engine. Because of constraints around access to the data and its sensitivity, the first first iteration of the engine is based on rules and heuristics. It is establishing the feedback loop between the interactions suggested, if the user accepted or rejected them, the time to the user actioning the suggestion, and the outcome. The objective is to have a future version that uses this data for unsupervised ML that will allow the engine to train itself on each client data set. This last piece is currently highly experimental, and we are unsure of the ability of the unsupervised ML to deliver acceptable results.
As a provider of machine and deep learning solutions we are very fortunate to be based in the UAE. We are seeing significant traction and appetite for solutions that make use of AI and we have already built solutions around Smart Waste Management and Smart Parking. The opportunities we see are around monetizing services that operate outside of the commercial economy. Examples include police and civil defence.
We are working in the fintech and e-commerce fields.currently we are working in some AI projects which will make a major changes in our product . We expect making a huge profit after lunching those projects
Some context: Stallion AI is built from the grounds up around advanced AI technologies including deep learning. We provide AI services, solutions and products to help organizations in the MENA region across different industries to implement AI today. Everything we do is AI and all the value we create for our clients and ourselves is derived from building and implementing advanced AI technologies. In addition to serving clients, we conduct active R&D in AI in our facilities in Canada and UAE. All our processes and products are built around AI from the grounds up and so we have “100%” traction. Some examples of where we found AI and ML techniques to provide the most value include: advanced Arabic & English chatbots, document understanding in English and Arabic, investment portfolio management products run purely by AI (i.e. no human intervention, no quant style rules). We are also seeing very significant ROI in using AI in select defense applications.
At Reaktor, we believe one of the best ways to get traction using AI is to build a basic understanding of the technology across the organisation. I’m not talking about turning everyone into experts, just teaching people to see through the hype, understand the basics so they can see through the hype to understand how it works and what it can/can’t do. Then you can have your whole business looking out for opportunities for AI to improve the business – not just the IT and digital teams. Also, getting any AI project off the ground takes collaboration across many teams and individuals. That’s going to be a lot easier when more people understand the technology and are open to using it. We recently developed the ‘Elements of AI’ online course with the University of Helsinki to help do this. It’s free and anyone is welcome to give it a try: https://www.elementsofai.com
We are using this across our org for various things such as threat detection, AWS tagging and other functions.
We have used Machine Learning for converting static Correlation Rules of SIEM or Threat Analytics solution to Dynamic, meaning you don’t need to trust the statically defined man made rules any more and you can be prepared for protection against the modern day zero day Vulnerabilities. Correlation rules are dynamically created, maintained, suspended and purged based on machine learning. We have built our own Dynamic Correlation Engine using Apache Spark for this. We are now working on converting data stores like Elastic search Logstash and Kibana (ELK) , Hadoop and other enterprise data lake solutions to Threat Analytics Engine with a ML AI engine called Korrelat.
Beginning use in IT audit as well.
We use it for learning consumer experience while using our portals, Internal employees application engagement.
We use Deep Learning to solve a multitude of robot perception problems. Our focus is mainly in developing smart robots and we use Deep Learning for tasks from obect detection to path planning.
Mainly computer vision projects for big entities in the Middle East (corporations and government). Vehicle recognition/re-identification, face recognition, multiple object detection, and recognition, etc.
At HeyDoc! we've been looking at AI to help with symptom checking and pre-diagnosis to help direct patients towards better understanding their health issues and symptoms they've been struggling with. A person's health is a pretty sensitive subject so at this stage it's not completely utilized (we're not fully there yet) as a solution but more of a guide to assist them in better understanding their symptoms and to help the medical specialists by automating some of the initial screening questions. Wearable devices are all around us feeding us with vital health and wellness data. We're working on better understanding this data via ML and thus help people better understand themselves. When we do that we would be able to pro-actively monitor patient and notify them when something is out of the ordinary, advise on potential solutions and connect them with the right medical consultant.