By Mariyah Saifuddin
In a recent episode of Tech-Driven Business, Todd Kackley, vice president and CIO of Textron, joined host Mustansir Saifuddin to do a deep dive into the ever-evolving landscape of Gen AI. He shared real-life insights in how companies can successfully implement a generative AI solution and the pivotal role leaders play in crafting a scalable, well-governed and future-proof data analytics and AI infrastructure.
As organizations navigate their journey with Gen AI, management teams must look at the complete picture. Kackley shared some great starting points.
Third-Party Capabilities vs. In-House Models
In the podcast episode, Kackley highlighted the Textron journey with their first production Gen AI solution. Here’s what stood out:
Teams must look at the bigger picture when making a decision on whether to use third-party generative AI model. Even if your organization has experience working with AI and building models, Kackley pointed out that a third-party model may be of value “simply because of the capability and the massive amounts of compute and investments required to go off and build a model” and even more importantly, if it is secure. In other words, can you work within the confines of your data tenant, leverage the benefits of the OpenAI, and protect your sensitive data.
As Kackley shared, in-house models may make sense where there are small-scale opportunities to build lambda-based models or others using technology that does not require a massive data center type of computing horsepower required for larger models like Open AI or Azure AI.
Kackley pointed out that scenario covers most use cases being seen right now. In other words, think about what outcomes your teams want to achieve. If it is for quick analysis, summarization, or content generation it may not make sense to build a model when third-party proprietary models already exist to do that. For instance, think about third-party turnkey solutions for contract evaluation, proposal generation, service centers, etc. They can accelerate your ability to implement quickly and get value quickly. For Textron, the solution for its recent implementation of its inaugural generative AI solution meant leveraging a third-party proprietary generative AI model because it offered not only scalability but also ensured security and compliance, essential elements for Textron.Kac
Proprietary Solutions vs. Open Source
While organizations must consider cost, according to Kackley, strategy must drive whether they choose a proprietary or open-source solution. Open source can drive innovation and accessibility, but leaders cannot overlook factors like long-term viability, security, and vulnerability management. For instance, “the release of ChatGPT as an open source technology has fundamentally changed the perception of artificial intelligence and put the technology in the hands of the everyday consumer. You don’t need to have a Ph.D. in a neural network to go off and understand how to interact with artificial intelligence.” But this fast adoption of technology comes with risks. Teams should consider the following:
Security: How secure is the solution? How much time and resources are needed to maintain security to keep up with the vulnerabilities of such open-source technology?
Longevity : Will open-source technologies be consumed by another company or become proprietary, resulting in a fee-based application that you are now committed to?
Cost: Hand in hand with longevity, do you want to invest in a low-cost solution that only has a couple of years life cycle?
Durability: Will you find your applications being consolidated such that you have to reinvent what your technology stacks look like often?
As Kackley shared, strategy “really forced us to stand back and look at all of these investments to make sure that there’s going to be viability in the solution provider to continue to support, do investments, and continue to evolve that product set, whether it’s a small individual capability or a larger enterprise-wide program.”
Strategy Drives Success
Textron’s success with AI-driven solutions illustrates the important role that strategy plays when deciding what path to take. For Textron, it meant focusing AI efforts on service-center optimization to streamline access to critical information, empowering maintenance technicians to resolve issues efficiently and minimize downtime. From a data perspective, Textron had a lot of data to maintain as well as disparate sources of data,
Access and accuracy can be challenging so Kackley’s team focused on how to accelerate access to content to help them approach how to troubleshoot. The result? AI solutions are “going to drive a measurable impact in the time that these maintenance technicians normally spend in front of a computer.”
Kackley provided invaluable insights into establishing a scalable, well-governed, and future-proof data analytics and AI infrastructure. He emphasized the need for organizations to adapt to the evolving landscape of Gen AI, acknowledging its inevitability and the imperative to integrate it securely and effectively into business processes.
Gen AI is not just a trend but a transformative force shaping the future of technology-driven businesses. Organizations must embrace this evolution, recognizing that users will find ways to leverage Generative AI to enhance their workflows and drive business outcomes.
Interested in more Tech-Driven Business podcasts? You can listen here
About the podcast guest
Todd A. Kackley is vice president and chief information officer for Textron Inc. In this role, he leads the business unit chief information officers and the Textron Information Services (TIS) organization. He oversees Textron’s Information Management Council and manages Textron’s information technology supplier and outsourcing relationships.
Prior to his current role, Kackley was executive vice president and chief information officer for Bell, where he developed and executed IT and digital strategy, aligning business systems, infrastructure, cybersecurity and development capabilities to the needs of the business.