Navigating the Gen AI Landscape

With the proliferation of Generative AI (Gen AI)-applied technologies such as chatbots, ChatGPT, and various cybersecurity tools, it’s no surprise that Gen AI has made its way into the insurtech ecosystem. It is now a driving force behind natural language interfaces, decision making, and other innovations that are revolutionizing insurance solutions and cloud technology. Special guest Michael Schwartzman, Sr. Azure Specialist, Core & App Innovation at Microsoft, joins Michael Mirel, Director, Cloud Platform Engineering at Sapiens, to discuss the intricacies of cloud technologies within the context of Gen AI, and how some of Microsoft’s products are supporting the expansion of Gen AI within the greater insurtech landscape.

Michael Mirel|Michael Schwartzman

Michael Mirel: Hello! Welcome to the Sapiens Insurance 360 podcast. I’m your host, Michael Mirel, Director, Cloud Platform Engineering at Sapiens, and I’m so glad that you’re out there listening; this is where we discuss the latest news, trends, and issues from across the insurance solutions and technology spectrum.

In today’s insurtech ecosystem, you can’t talk about insurance solutions and cloud technology without including Generative AI, commonly known as Gen AI, the artificial intelligence technology that can produce various types of content, including text, imagery, audio, and synthetic data in response to prompts. In today’s episode, “Navigating the Gen AI Landscape,” we’ll take a deep dive into the intricacies of cloud technologies within the context of Gen AI, exploring real-world implementations, potential pitfalls, security risks, and strategies for addressing challenges. We’ll also examine the role of AI in development processes and our partner Microsoft’s landscape in analytics and cloud.

And from Microsoft, today’s very special guest is Michael Schwartzman, Sr. Azure Specialist, Core & App Innovation at Microsoft, who will be walking us through today’s topic. Michael is a seasoned professional in the cloud computing industry. Based in Israel, he is focused on assisting global ISVs with building and selling SaaS solutions on Azure. Before joining Microsoft in 2022, Michael was a Lead Cloud Solution Architect at VMware, where he specialized in VMware Cloud on AWS, further solidifying his expertise in cloud applications.

Michael, thanks so much for joining us today!

Michael Schwartzman: Thank you for having me. Michael! Happy to be here.

Michael Mirel: So Michael, let’s start with a basic question. What practical applications do we observe in Gen AI architecture and what are the primary concerns or pitfalls encountered during implementation?

Michael Schwartzman: Right. So in the realm of practical applications, with Gen AI, Gen AI is revolutionizing everything from automated customer service to cybersecurity. And the idea is that for customers, these can provide natural language interfaces for use cases such as summarization and searching for internal data sources, a generation of content and code. But also lately we see more advanced use cases where Gen AI is being used as part of an orchestration engine to perform actions and configurations in the system.

And so I can give just a couple of examples of what we’re doing in Microsoft. So we have in Microsoft the Azure portal, where we have a Copilot chat where you can search recommendations and ask questions about your subscription and data. Also, we can generate around scripts in a batch and PowerShell and automatically generate a graphic from natural language. We have the Copilot implemented in the rest of our products and stack anything from, you know, helping you to summarize your email threads and then helping you write better emails, or helping you to create better PowerPoint slides, etc., etc. And we also have the GitHub Copilot stack where developers can write code and ask questions about their code base. And we see, I see customers which I cover, and we have a PR that was announced recently with customers of mycheckpoint, they’re a cybersecurity industry leader in the area of firewall management and firewalls, and they implemented the chatbot system where you can interact with the different technical entities such as identity firewall rules, natural language like asking specific questions about security posture, but also using it to orchestrate and generate new firewall rules and policies like Gen AI and then natural language interface.

So basically what all of this does is it creates more efficiency and helping to generate a lower barrier to entry from a technical standpoint for analysts. And they’re helping folks who are technical to accomplish more with less repetitive tasks. Additional use cases that we see in general in the markets is automated customer services. So companies like Zendesk and Salesforce use Gen AI to our chatbots and they help with support ticket response, and provide customer services and fraud detection. Companies like MasterCard are using Gen AI for real-time fraud analysis, detection and preventing of fraudulent transactions. Content creation platforms like Canva and Adobe leverage Gen AI to suggest design layouts and create personalized marketing content. Microsoft also has offerings in this area. You can create a, like I said, PowerPoint slides and graphics with Microsoft Designer. And so that is something that we’re actively working in.

From pitfalls and concerns, we have a couple of them. So first of all, is the area of data security and privacy. So customers are really keen to understand that they are within the compliance area with the regional data storage and processing laws. This is crucial when deploying GenAI globally. Second, there is the privacy, so establishing strong data to handle orders and ensuring models respect privacy is a mandatory legal compliance and for customer trust. Intellectual property, clarifying the IP ownership and IP-generated output is essential to avoid legal disputes and encourage innovation. Cybersecurity, adopting a comprehensive cybersecurity posture, including protective measures, monitoring rapid response and recovery strategies. Scalability, designing an infrastructure that can accommodate growing or fluctuating workloads without service interruption is crucial and high availability, ensuring no single point of failure, a sort of fault tolerant system design is possible across multiple data centers and cloud regions.

Michael Mirel: Sounds awesome. I can share also from Sapiens’ side that we are starting to adopt the Gen AI slowly and the first thing as we start looking into it is first of all, how to make the Gen AI very secure, with respect to identity networking, authorization, token generation. And this is very critical for organizations that are starting to adopt Gen AI and integrate these two in their core systems. The first thing is to look into the security domain That’s very useful and very awesome information. Thanks for that, Michael. And now in addressing the challenges posed by Gen AI implementations, what strategic approaches should the organization consider?

Michael Schwartzman: So security risks in Gen AI are significant, especially when it comes to privacy, as AI systems process large amounts of sensitive information, they become a prime target for cyberattacks. So to mitigate those concerns, we need to provide a robust encryption, regular audits, and transparency around the AI framework that explains the decision and processes, making sure that there is accountability and compliance with regulations such as GDPR and making sure that we take into consideration the entire architecture and design of the application, not only the Gen AI engine like you mentioned previously. So making sure that the networking, security, identity, and token generation is being done in a secure way from the entire architecture of the organization.

Michael Mirel: Thanks, Michael. That’s so important for our listeners to hear and consider. So, another critical question: How does the development process integrate AI technologies, such as GitHub Copilot, and what advantages does this offer?

Michael Schwartzman: So tools like GitHub copilot, they are transformative for developers. Like I said before, we can use it to suggest code, help debug, and even automate repetitive tasks. And this at the end leads to increased efficiency and allows developers to focus on the more complex and innovative tasks. And Microsoft was seeing Copilot not only speed up the coding process, but also serve as a learning tool for less-experienced developers with capabilities such as asking questions on code repositories with our latest announcement of the GitHub Copilot enterprise, and use cases such as summarization of pull requests, which is also part of the enterprise ecosystem, helping to debug terminal errors or compilation errors, helping  to write terminal commands. This is an area of great investment for Microsoft and we see really rapid and constant innovation in the area of GitHub where we see it coming from code repository, platform, and CIC platform to Gen AI platform for the organization.

Michael Mirel: Awesome, eventually it’s everything about efficiency. Okay, so one last question that should provide a nice endnote to our conversation today. Could you provide insights into Microsoft’s landscape regarding analytics and fabric, and how does it complement the Gen AI ecosystem?

Michael Schwartzman: Sure. So Microsoft’s analytics in the fabric of the landscape is extensive, encompasses tools such as Azure sign-up analytics for big data processing, Azure machine learning and different set of tools. And this infrastructure supports the Gen AI ecosystem by providing a necessary platform for storage, computing analytics, and ensuring that the applications are there scalable, reliable, and secure. But the latest innovation that we have in the area of the fabric is that we’re able to take and integrate seamlessly all of those services to assess a solution and experience where Microsoft is eliminating the integration. Basically all of the products are being seamlessly integrated and being consumed as a SaaS solution, which is really nice. And also the elements of cost savings because there is shell compute that can be shared. So the different products as a singular tool, you can mix and match different services and not bring up compute capacity, whether it’s overprovisioned or unused.

And another great point I would say fabric is that it’s based on one leg, which is based on an open format where you can basically read and integrate with external systems, which are not necessarily Microsoft based. And also, it has a concept of shortcuts where you can read from a different non-Microsoft or Azure database; doesn’t have to necessarily even sit on an Azure, and you manage all of your different products and services permissions and now but from a single place on an open format, that can seamlessly integrate to other systems and also read from other systems. And on top of that, we also enable Gen AI inside of the fabric solutions. So you can use natural language to basically query and generate reports and different phases of your analytics process to the AI-driven Copilot of Hub.

Michael Mirel:  Sounds very awesome, very nice innovation from Microsoft. Thank you so very much for this very insightful session, Michael! We could go on for another hour on this very important topic and I know that I speak for all of us when I say that I can’t wait to see how Gen AI and its future applications will continue to evolve.

And to our listeners, as always, thank you so much for spending your time with us today. We love hearing from you, so please reach out to us on our channels. And don’t forget to subscribe to the podcast! As always, we’ve got more coming, to don’t forget to tune in next time for another episode of Sapiens Insurance 360.