Generative AI is a business game-changer, but only if it’s built on solid foundations

The technology is transforming a wide range of processes, including customer service, content creation and software coding.

By WP Creative Group

November 10, 2023

• Facebook
• Twitter
• LinkedIn
• Mail
• Link

Generative AI is no longer an experimental technology whose impact is years away. It’s here, and ready for prime time. So it’s no wonder that C-suite executives and managers are thinking hard about it as they come under pressure to integrate the technology into their plans and strategies. Cloud services innovator AWS is on a mission to demystify generative AI for these leaders, while helping them get started with it.

This commitment was on display in October at AWS ExecLeaders Data and Generative AI Day, where AWS leaders and customers shared valuable insights about the technology. Generative AI will “drive rapid innovation at all levels … this is not just theory and buzzwords; it’s real,” said co-host Tom Soderstrom, global enterprise strategist at AWS.

Generative AI can create new artifacts like images, music and text that are novel, high quality and customized based on user inputs. That content could show up in everything from financial reports to chatbots and virtual assistants. Companies can also use it to automate time-sensitive and error prone business operations, including software coding, and improve customer experiences and speed up content creation. There’s at least 63 generative AI use cases spanning 16 business functions, adding the equivalent of up to $4.4 trillion to the global economy through increased productivity in key sectors.*

But with so many potential applications, how can organizations know where to start? To answer this question, leaders must establish clear goals that are informed by business priorities. “My biggest advice for selecting the right use case [for generative AI] is to fall in love with the problem, not the solution,” said Swami Sivasubramanian, vice president of data and AI at AWS. For example, organizations looking to improve developer productivity can use Amazon CodeWhisperer to automate code generation and those seeking to easily create and customize data visualizations using natural language can leverage generative business intelligence (BI) capabilities in Amazon QuickSight. “At Amazon, we work backwards from our customers’ needs before we propose solutions,” said Sivasubramanian.

Organizations must understand that AI is not something that can be switched on like a light. They must first have a plan to source, store, vet and secure the data that’s needed to power generative AI applications. Some large language models include billions of parameters, so it’s essential to implement “a modern data strategy” to manage it all, said event co-host Kelli Such, Americas data strategy leader at AWS.

Foundations for success

That starts with building a strong data foundation, including capturing a wide range of training data to create foundation models that can be used for a variety of tasks. “Data is the key to building foundation models,” said Sivasubramanian.

To make its solutions distinctive, an enterprise’s AI projects should incorporate as much of its own data as possible. “Your organization’s data is your differentiator. Everyone has access to the same generative AI models. So your data is the key to moving from generic AI to AI that knows your business,” said Sivasubramanian.

Customizing the foundation models, through fine tuning or through a process called in-context learning, will provide real business value, said Sivasubramanian.

Amazon Bedrock, which became generally available in September, makes leading foundation models, including Meta’s Llama 2 and models from Anthropic, Cohere, Stability AI and Amazon, available through a single application programming interface. This makes it easier and faster for AWS customers to get started with generative AI and drive business outcomes. The ability to work with multiple models is critical as “no one model will rule them all,” said Sivasubramanian. Amazon Bedrock also offers tools for model customization and privacy protection.

Data must be readily accessible to these models through systems and services that can store a range of organizational information, and also pull in external sources. “One of the biggest barriers to working with your data is the ability for your teams to locate the right data sources and access them when they need to,” said Sivasubramanian.

Governance is key

Organizations also need a data governance plan that stipulates where data lives, who has access to it and how it can be used throughout given workflows or business processes. The plan must set a high bar for data quality so that team members have trust and confidence in the data. “The quality of the data pretty much determines the sort of AI outcomes you’re going to drive and what sort of confidence you’re going to get,” said Gopinath Sankaran, VP for strategic cloud ecosystems at Informatica.

An effective data governance strategy must strike a delicate balance between control over data and access to data. The right people need to be able to discover, access and share information to encourage innovation, and the data must be freed from silos. Amazon DataZone supports the entire data discovery, access and usage lifecycle while providing complete visibility and control over how the data is being shared. 

For some organizations, that may require a change in mindset. “This ends up being primarily a cultural challenge,” said Mark Schwartz, enterprise strategist at AWS. “The cultural change that’s necessary is to convince employees to change their thinking. Instead of thinking of the data as belonging to them, they must think of themselves as stewards of the data for the rest of the organization.”

One company that is focused on building a data-driven culture is Carrier. “That’s the foundation we’ve been putting in place in the organization,” said Bobby George, chief digital officer at the leading global provider of  building and cold chain solutions. “The process of innovation and decision making is not about collecting the data, it’s really about driving value from the data.” At Carrier, information is democratized and shared widely to drive action.

Real-world results

The market for generative AI services could reach $200 billion by 2032, according to Deloitte, which earlier this year partnered with AWS to establish a generative AI practice to help clients navigate the technology. As proof-points of how generative AI is already making impacts, AWS ExecLeaders Data and Generative AI Day featured numerous companies that are engaging with AWS to successfully deploy the technology.

• Salesforce: Salesforce enables customers to Bring Your Own Lake (BYOL) and Bring Your Own Large Language Model (BYO LLM) to seamlessly unite all of their data across Salesforce and AWS for better, faster insights and predictions. The arrangement allows customers to bring LLMs hosted in AWS to generative AI apps and services in Salesforce without copying it or moving it, helping to maintain security and compliance. “For a lot of use cases, our customers are asking us to really allow them to use the data in place,” said Gabrielle Tao, senior vice president for product management at Salesforce.
• FOX: The news and entertainment giantis using behavioral data to power AI and generative AI services and is standardizing all of this information so it can be used across business units, which include FOX News, FOX Weather and FOX Sports. The goal is to “create an experience that makes sense to users and also capitalizes on the fact that we know you across all these areas,” said Lindsay Silver, senior VP for data & technology at FOX Corp.
• Booking.com: The online travel planning site is using customer and transactional data to create better travel experiences and more customer-centric products. “Data … is at the root of everything we do,” said Milo Milovanovic, senior director, Big Data technologies at Booking.com. “We really try to understand customer behavior.” The company recently embarked on a project to push all of the data it captures to the public cloud and manage it through cloud-native tools and services. “AWS has been a great partner for that,” said Milovanovic. Among other benefits, using the cloud can hasten time-to-market for new services. Earlier this year, Booking.com launched an AI trip planner to enhance the travel planning experience.
• Startups: It’s not just established brands that are embracing generative AI. Innovative startups are working with AWS to use the technology to accelerate growth. Theia Insights is using it to generate reports that give investors real-time insights into the financial markets. OpenAds gives publishers tools to integrate AI-driven chatbots and personalized ads into their sites. And Krikey AI lets users quickly create animated videos, games and more. All three companies came through an AWS incubator program. “AWS is eager to support them in developing incredible, generative AI solutions,” said Soderstrom.

Building trust with responsible AI

Even as they extolled AI’s tremendous upsides for business, many of the thought leaders who spoke at the event cautioned that the technology must be used in ways that are responsible, ethical and respectful of customer privacy. “Responsible AI goes beyond just the technology,” said Diya Wynn, responsible AI lead at AWS. “It extends to people, process and culture so that [organizations] can unpack areas of unintended consequences.”

Those consequences can include outcomes that exacerbate inequities in areas like housing and healthcare. “Responsible AI recognizes the threats that are coming,” said AWS Enterprise Strategist Tom Godden. “Poorly trained data can perpetuate biases, and lead to bad decision making.” For an AI program to be truly responsible, it needs to have human oversight. AI should be “augmenting human oversight, not replacing it,” said Godden.

Generative AI can take business productivity and innovation to new heights. For organizations to realize these benefits, their leaders must commit to fully understanding how to use it effectively and responsibly.

Generative AI can help take your business to the next level. Learn more.

---