Artificial Intelligence (AI) will play a growing role in unlocking the value of enterprise data, according to Google Cloud’s senior data analytics manager.
Gerrit Kazmaier, vice president and general manager of databases, data analytics and Looker at Google Cloudtold Computer Weekly that the cloud and search giant’s customers are already combining AI with more conventional business intelligence tools.
This is because AI helps bring together structured and unstructured data, Kazmaier said. AI systems are beginning to perform increasingly complex analyses, but they can do so at much faster speeds and with much greater volumes of information than human experts.
Google supports its customers in this process by drawing on its research experience, as well as its cloud resources and its experience in the development of the Transformer model, one of the foundations of generative AI systems.
“We’re reinventing, say, Google Search for enterprise data,” Kazmaier said.. This is largely about combining the potential of AI tools, including generative AI, trained largely on public data, with the domain- and business-specific insights contained in AI applications. enterprise and enterprise data lakes.
“So far, Google Search is active primarily in the public domain or on the public web,” he said. “Ultimately, there was a big opportunity to bring this into the enterprise domain, essentially giving every data point that exists in enterprises, which is not part of the worldwide web, a similar interface.
“Everyone knows how to use Google. I’m sure every CEO on the planet knows how to use Google to search the public web. I also believe that only a very small number of people on this planet, and certainly a small number of CEOs, would be able to use a dashboard tool themselves to find information about their own company.
“With Generative AI (GenAI), we have the ability to communicate with your business data, just like you can communicate with public data through Google Search. »
Google “recovers” data
Google has a “cultural understanding” of the need to make information more accessible, according to Kazmaier. It is at the heart of its mission to bring together AI and conventional analytics.
“From the technologist’s perspective, it starts with researching the world’s information and making relevant information universally accessible and useful. This is necessary to develop technology that is widely used in generative AI today,” he continued.
“There is a reason why Google was the original inventor of the transformer model, which is now the underlying architecture for all of these models. Gemini (formerly Bard of Google)or ChatGPT, (Meta’s) Llama and so on.
“There is first a deep understanding, when we say we want to map someone’s question to a meaningful answer, of the technology we need to build to understand the semantics necessary to effectively address that question and restore it in a form. factor that a human can work with.
Google has established a roadmap for integrating AI into its analytics tools, integrating BigQuery with Vertex AI, enabling data feeds to AI in BigQuery Studio, and allowing users to create analytics models. machine learning in BigQuery ML and exporting to Vertex AI, as well as adding features to Looker and Looker Studio.
According to Google, one of the most promising applications of generative AI in business is helping non-specialists interact with business data.
Rather than learning coding or analysis skills, or writing queries and designing dashboards, GenAI should enable business users to interact with a database, data warehouse, or data application. data lake using natural language – and also get a natural language response.
This has two key advantages, besides ease of use.
It removes the need to filter data to fit the format and functionality of a dashboard. This inevitably means that some information will be truncated or deleted. And only a minority of enterprise users have the skills to explore analytics tools on their own.
An AI-based system has the potential to be more accurate because it can process larger volumes of data and a wider range of data sources. Kazmaier called this “expansive data.”
The other advantage is that users can interact with Systems driven by AI in a more iterative way. They can refine and refine their queries, asking more questions until they find the information they need.
Kazmaier cites the example of Camanchaca, a Chilean seafood company that uses a suite of standard BI tools, including BigQuery, Vertex AI and Looker. He created an AI agent to allow all employees to access company data.
“This opens the door to data and analytics for professionals who are not specialized in data analysis. Everyone has a question to ask. Not everyone has an analyst to answer that question,” he said.
“New use cases are emerging for generative AI capabilities that bring us much more than dashboards and traditional data analysis. The consumer is evolving from data analysts to every knowledge worker with access to meaningful data analysis.
This allows business intelligence to move from simply visualizing data to interpreting information, as a human analyst would, according to Kazmaier.
“When you look at data, you want someone knowledgeable, like an analytics specialist, to help you interpret it. What does this represent conceptually, or how does it compare? he said.
“It’s not a question that needs to be answered by the data point itself, but you need someone who is really calibrated if you will, who understands how to interpret: ‘Is this a good or a bad margin? Is it a good or bad day of exceptional sales? “.
“This can be trained and coded and is generated by the agents we introduce into our BI offering. So basically, you’re collaborating with an analyst who can help you understand and interpret the data you’ll see. One of the main problems we have with traditional BI is that we need to compress information to a level that becomes human understandable.
According to Kazmaier, data consumers are evolving. More and more users want access to data, and AI – particularly generative AI – offers a way to unlock that access in a way that conventional BI cannot.
But the integration of AI into business intelligence and Google’s roadmap is not limited to providing a better interface. AI offers businesses a way to stay ahead of the seemingly endless growth of enterprise data – and hopefully derive some business value from it at the same time.
Kazmaier talks about “broad” rather than big data: not just having more data, but adding more data points to the analysis. AI systems are well-positioned to decide whether it’s worth considering additional factors, he said, and they have the processing power to do it quickly enough, so they don’t delay decision-making.
“One of the biggest changes we’ve seen is the use of unstructured data,” he said. “If you think about it, unstructured data represents roughly 90% of the world’s data.. Traditionally, this data has not been used in data analysis. There were specialized applications for documents or to automate certain processes like paying invoices, but they were not considered part of an enterprise data landscape that we actively use, explore and analyze, like you do it with structured data.
“With generative AI, working with unstructured data, people understanding it and extracting insights from it, becomes extremely flexible and available,” he continued.
And AI tools allow business users to dig deeper into their data and better understand trends in their organization: moving from questions of “what, when and where” to, ultimately, “why.”
“You have big models trained on public data, and you can ask them about questions in the public domain and it’s amazing what that can do,” Kazmaier added.
“But these models aren’t trained to use a company’s data, and that’s quite interesting. How do we deploy these big (language) models with enterprise data so that you can open up all the information you have about your data, so that it’s all useful in the enterprise?
AI agents, he says, already provide these answers.
