I sat down with Teresa Tung to learn more about the changing nature of data and its value for an AI strategy.
The success of AI depends on several factors, but the key to innovation lies in the quality and accessibility of an organization’s proprietary data.
I sat down with Teresa Tung to discuss the opportunities presented by first-party data and why it is so critical to creating value with AI. Tung is a researcher whose work spans revolutionary cloud technologies, including the convergence of AI, data and computing capacity. She is a prolific inventor, holding more than 225 patents and applications. And as Accenture’s global head of data capabilities, Tung leads the vision and strategy that ensures the company is prepared for the ever-changing advancements in data.
We discussed a multitude of topics, including Teresa’s six ideas.
Finally, we concluded with Teresa’s speech Advice for business leaders using or interested in AI
Susan Etlinger (SE): In your recent article, “The new data essentials”, you laid out the idea that first-party data is an organization’s competitive advantage. Would you like to expand?
Terese Tung (TT): Until now, data was treated as a project. When new information is needed, finding the data, accessing it, analyzing it, and publishing it can take months. If these ideas give rise to new questions, this process should be repeated. And if the data team faces bandwidth limitations or budget constraints, it will take even more time.
“Instead of treating it as a project – an afterthought – proprietary data should be treated as a critical competitive advantage. »
Generative AI models are pre-trained on an existing corpus of internet-scale data, making it easy to get started on day one. But they don’t know your business, your people, your products or your processes, and without this proprietary data, the models will provide you with the same results as your competitors.
Companies invest in products every day based solely on their opportunities. We know the opportunities offered by data and AI (improved decision-making, reduced risks, new monetization paths). So shouldn’t we consider investing in data in the same way?
SE: Given that much of a company’s proprietary knowledge is in unstructured data, can you talk about its importance?
TT: Yes, most businesses operate with structured data, data in tabular form. But most data is unstructured. From voice messages to images to video, unstructured data is high fidelity. It captures nuances. Here’s an example: If a customer calls customer support and leaves a review for a product, that data can be extracted by its components and transferred into a table. But without nuanced elements like the customer’s tone of voice or even profanity, there is no full, accurate picture of this transaction.
Unstructured data has always been difficult to use, but generative AI excels at it. In fact, it is needs the rich context of the unstructured data to be trained. This is so important in the age of generative AI.
SE: We hear a lot about synthetic data these days. How do you think about it?
TT: Synthetic data is needed to fill data gaps. It allows businesses to explore multiple scenarios without the significant costs or risks associated with collecting real-world data.
For example, advertising agencies can broadcast various campaign images to predict audience reactions. For automakers training self-driving cars, pushing cars into dangerous situations is not an option. The synthetic data teaches the AI – and therefore the car – what to do in extreme situations, such as heavy rain or an unexpected pedestrian crossing.
Then there is the idea of knowledge distillation. If you use this technique to create data with a larger language model (say, a 13 billion parameter model), that data can be used to fine-tune a smaller model, making it more efficient, more cost-effective, or deployable. on a smaller device.
The AI is so hungry. To be relevant, it needs datasets representative of good scenarios, boundary conditions and everything else. This is the potential of synthetic data.
SE: Unstructured data is typically human-generated data, so it is often case-specific. Can you explain more about why context is so important?
TT: Context is key. We can capture it in a semantic layer or a domain knowledge graph. This is the meaning of the data.
Think about every domain expert in a workplace. If a company generates a 360-degree customer data report covering domains or even systems, one domain expert will analyze it for potential customers, another for customer service and support, and another for customer billing . Each of these experts wants to see all the data but for their own account. Knowing trends within customer support can influence a marketing campaign approach, for example.
Words also often have different meanings. If I say “it’s hot for summer,” the context will determine whether I was talking about temperature or a trend.
Generative AI makes it possible to transmit the right information at the right time to the right domain expert.
SE: Given the pace and power of smart technologies, data and AI governance and security are a priority. What trends are you seeing or predicting?
TT: With new opportunities come new risks. Generative AI is so easy to use that it makes everyone a data worker. This is the opportunity and the risk.
Because it’s simple, generative AI built into applications can lead to unintended data leaks. For this reason, it is essential to think through the full implications of generative AI applications to reduce the risk of them inadvertently revealing confidential information.
We need to rethink data governance and security. Everyone in an organization needs to be aware of the risks and what they are doing. We also need to think about new tools like watermarking and confidential computing, where generative AI algorithms can be run in a secure enclave.
SE: You’ve said that generative AI can jump-start data preparation. Can you expand on this?
TT: Of course. Generative AI needs your data, but it can also help your data.
By applying it to your existing data and processes, generative AI can create a more dynamic data supply chain, from capture and curation to consumption. It can classify and label metadata, and generate design documents and deployment scripts.
It can also support reverse engineering of an existing system before migration and modernization. It is common to think that the data cannot be used because it is in an old system that is not yet cloud-enabled. But generative AI can restart the process; it can help you understand data, map relationships between data and concepts, and even write the program, including testing and documentation.
Generative AI changes what we do with data. It can simplify and speed up the process by replacing one-off dashboards with interactivity, such as a chat interface. We should spend less time processing data in structured formats and work more with unstructured data.
SE: Finally, what advice would you give to business and technology leaders who want to create a competitive advantage through data?
TT: Start now or get left behind.
We have realized the potential that AI can bring, but its potential can only be achieved with your organization’s first-party data. Without this input, your result will be the same as everyone else’s or, worse, inaccurate.
I encourage organizations to focus on preparing their digital core for AI. A modern digital core is the technological capability to drive data in AI-driven reinvention. It’s your organization’s combination of cloud infrastructure, data and AI capabilities, applications and platforms, with security designed at every level. Your data foundation, as an integral part of your digital core, is essential to hosting, cleansing and securing your data, ensuring it is high quality, governed and AI-ready.
Without a strong digital core, you don’t have the proverbial eyes to see, the brain to think, or the hands to act.
Your data is your competitive differentiator in the age of generative AI.
Teresa Tung, Ph.D. is the Global Head of Data Capabilities at Accenture. A prolific inventor with more than 225 patents, Tung specializes in meeting business needs with revolutionary technologies.
Learn more about how to prepare your data for AI:
- Learn how to develop an intelligent data strategy that endures in the AI era with the downloadable ebook.
- Watch this webinar on demand to hear Susan and Teresa dig deeper into how to extract the most value from data to differentiate yourself from the competition. Learn new ways to define data that will help drive your AI strategy, the importance of preparing your “digital core” before AI, and how to rethink data governance and security in the age of AI.
Visit Azure innovation insights to get a more executive perspective and advice on how to transform your business with the cloud.
