Accessing AI for your business
How the IBM Garage can help you put AI technology like Project Debater to work for you
AI for you right now
The ability to reason has long been the thing commonly acknowledged as what separates humans from other animals; is that also the thing that has separated humans from technology? The nuance and rich context required to be able to effectively reason — and even further, debate — feels innately human. To think about technology and Artificial Intelligence (AI) replicating that level of sophistication sparks our imagination like science fiction. But, as IBM’s Project Debater shows, it’s not some far off potential capability — it’s here today. Project Debater uses the Watson speech/text API, and the IBM Cloud offers several AI services, including Natural Language Understanding, Natural Language Classifier and Discovery service. The IBM Garage can help you take advantage of those technologies in ways that quickly match your business goals. With us, you can get AI technology on the job for you and your particular challenges in just weeks.
Where does AI fit into my business strategy?
AI’s broad goal is to help you make better decisions, and statistical models at the core enable that decision-making. Some of those models are built to emulate skills people use naturally, such as language or vision.
Suppose you are an expert on finance and listening to a discussion about leveraging debt helps you understand that the broader topic is cash flows. You may not even consciously recognize you’ve tapped your experience and mental connections to make that leap; however, indeed this is what is happening in your brain, pulling relevant, related information to enable you to continue the discussion from leverage, to debt management, to cash flow.
If you were going to make a business decision based on a similar set of contexts, an AI system could support any decisions you would make by extending the domain automatically and providing you with additional information. It is more than just a simple keyword search. Instead, an AI system would first extend the relationship of topics using the same mechanisms that people use, the words we use. Second, it derives the statistical relevance of those topics to the core. It would extend your topic of “leverage” to debt management and finance cash-flow case studies. Finally, it would give you the ranking of those studies to your topic. With that, you could make a better decision on the subject than you could otherwise, pulling from a vastly greater expanse of information and models than you could store in your brain.
Project Debater as one use-case of many
Project Debater shows how the context improves decision support for an AI system. In public debates, the context is the world’s knowledge on a given topic, framed to support a specific hypothesis. The project debater video shows a beautiful example of AI working within that context. Even the moderator and debate partner were amazed at how successful that was.
Business leaders depend on understanding all the relevant information in their context to making a strategic decision. They want to frame a decision and understand all of the impacts, not just the short-term, one-off examples. They want it specific to their domain with appropriate detail.
An AI system like Project Debater provides that ability to pull in all of the relevant knowledge in a domain and position it according to its statistical relevance. It is a core tool in an evidence-based business where information is used to mitigate risk and uncover opportunity. But it applies to any use case that involves considering all sides of a decision. It could apply just as well to decisions around extending business capabilities to helping customers choose from a variety of choices. Debating claims for insurance or health benefits, debating merchandising for retail… the applications are endless.
Is AI a trustworthy tool?
We all need to trust anyone or anything that’s giving us advice. If I gave you advice on a complex decision, I’d be willing to defend my approach and reasons for providing my conclusion. My personal integrity is also important. My values should be consistent with yours for me to be an effective advisor.
This is true for AI systems as well. From the core model all the way up to the interfaces to the data, the conclusions must be valid and consistent with our business values. The advice must be consistent with your business values. You need to be able to see and understand if the statistical weights for training sources are accurate. You should agree that the rankings for advice supporting the supporting decision are defensible.
IBM understands this need and has addressed it in several ways. One way is that systems like Watson OpenScale can detect and avert bias within the model itself as well as ensure that the model can scale to systems that need it. This helps ensure that the AI models remain transparent in the help they provide. Another way relates to what data you use for training. We provide ways for you to directly influence what data trains your models. That way, you can control the set of information your models uses to support your decisions.
Flexible ways to use the information
Another remarkable thing about Project Debater is how you use it — with your voice. The natural and historic way to have a debate is speaking, and it does a great job at that. It follows the state of the art of interaction of these kinds of systems, where the systems meet you the way you want to use them. They don’t require special buttons or complex commands to work. Instead, you just naturally interact with them because they were designed with people in mind.
The IBM Garage is ready
The IBM Garage’s goal is to innovate with our customers. These kinds of AI solutions are a great example of what we use to do it. We use Design Thinking in the IBM Garage Method to collaborate with you to understand what we’re trying to do. Then we pull in the best technology to meet that objective as a Minimum Viable Product (MVP). That way, the MVP confirms that your idea will work. Quite often, an MVP goal is about making better decisions. In those cases, the Garage will use AI technology to accelerate and support that goal.
The Garage covers both interesting aspects of technology demonstrations like Project Debater. If you’re amazed by the AI to support decisions, we have years of experience with AI to provide that kind of support. If you’re thrilled by the easy, modern design of the interfaces, we have design experts that can deliver that as well. Even better, we can deliver the solutions quickly, using agile approaches. We work with you to make sure that the solution meets your objectives for transformation.
Stephanie Trunzo, Global Head of the IBM Garage
Tony Efremenko, IBM Garage
Bring your plan to the IBM Garage.
Are you ready to learn more about working with the IBM Garage? We’re here to help. Contact us today to schedule time to speak with a Garage expert about your next big idea. Learn about our IBM Garage Method, the design, development and startup communities we work in, and the deep expertise and capabilities we bring to the table.
Schedule a no-charge visit with the IBM Garage.
