Taking Flight with Generative AI
Steer clear of turbulence with Slalom and Microsoft
In a recent interview, Microsoft executive VP of business development, strategy, and ventures Christopher Young explains the value of generative artificial intelligence, or generative AI (GenAI), in that it gives organizations “a set of co-pilots that allow people who have certain job functions today to do their jobs more efficiently, more effectively, to produce better output.” In other words, Microsoft’s Copilot technology has a similar function to a copilot on an airplane; it supplies assistance that makes the job of the person in charge a lot easier.
But even a pilot and copilot need help to successfully navigate takeoff, landing, and the journey in between. The ground crew, baggage handlers, air traffic controllers, and others also play important roles.
The same goes for your GenAI projects — they can run smoother and encounter less turbulence if you have the right crew backing you up. So, let’s put our seat backs into an upright position, stow our carry-on bags and tray tables, and prepare for the journey with Slalom and our Ground Control offering.
Your team on the ground
Within your business, there are likely multiple functions and processes that require you to sift through vast quantities of data to locate, refine, and deliver the insights, information, and intelligence your stakeholders need. Generative AI is the ideal technology for such activities, but leaders seem to be struggling with the best way to apply it effectively. Recent research conducted by Slalom found that “71% of executives don’t fully comprehend the scope of tasks AI can effectively augment or automate in their organizations.”
This is precisely where Slalom can help. We’ve developed an offering that we’re calling Ground Control. Ground Control is an approach to building additional intelligence into your generative AI practice so that you’re able to both solve immediate, individual needs as well as establish a repeatable process to leverage GenAI successfully throughout your business. By standardizing and speeding up the deployment of AI applications within your environment and ensuring that your AI and human workers are collaborating as seamlessly as possible, Ground Control can help you realize the value GenAI delivers at a broader, more effective scale.
Ground Control can work within your Azure OpenAI instance to help you:
- Work more securely with generative AI. Slalom has devised best practices around GenAI security, including the use of common integration patterns that keep your data secure within your Azure tenant. Because your data never leaves your secured cloud environment, Ground Control is an ideal approach for applications involving sensitive, confidential, and private information, such as legal, HR, and even software development.
- Keep your people not only in the loop but in charge. Ground Control can help you ensure that your AI applications are notifying people within their workflows when something requires human attention. This allows you to satisfy a fundamental tenet of responsible AI by including checkpoints that promote human accountability for your AI systems.
- Safely scale your GenAI program across the organization. Slalom’s approach helps you create repeatable AI applications that can support a wide range of use cases. Once you’ve launched your first project, we’ll help you prioritize other processes or areas of your business that will benefit from what Ground Control offers.
Ground Control in action
So, how exactly does Ground Control work? Policy grounding is a concept that’s core to the approach. Grounding in generative AI is the action of guiding an AI model with facts to increase accuracy. Policy grounding is more specifically the action of guiding an AI model with your organization’s policy standards. To illustrate how this works, let’s look at a use case that’s common within most organizations: creating a contract for a new client.
In this example, we’re using an organization’s own legal policies and standards to guide the AI model. These can be given to the AI in any format, such as Word documents, Excel spreadsheets, PowerPoint presentations, PDFs, and images. The AI works in much the same way as a classification algorithm by accessing the information given to it and configuring its output based on that information. Only in this case, there’s no code involved.
In this example, as with other use cases, the AI handles only specific tasks within a larger process. These types of tasks, where new information must be verified against existing, documented rules, policies, and standards, can be completed much faster by AI than by a human. But AI remains squarely in its support role — as copilot and ground crew — with humans still very much in the pilot’s seat as the ultimate authority.
The Ground Control approach can be applied to numerous scenarios throughout your organization, from HR to finance, logistics, procurement, customer care, and so much more. We’ve only just scratched the surface of all that can be achieved when our pilots, copilots, and crew are working together toward common goals.
Want to learn more about running a successful, well-balanced GenAI program in your organization? Download our whitepaper to find out more.
Slalom is a global consulting firm that helps people and organizations dream bigger, move faster, and build better tomorrows for all. Learn more about Slalom’s human-centered AI approach and reach out today.
