In what could be a peek at the future of AI integration in enterprises, MicroStrategy on Tuesday announced a new addition to its platform that simplifies access to business analytical data within organizations.
MicroStrategy Auto is a customizable AI bot that the company said offers a faster, simpler way to deliver business intelligence to anyone in an organization. Auto is the latest enhancement to MicroStrategy AI, released in October 2023, a solution for rapidly building AI applications on trusted data.
Auto can be deployed as a standalone app or embedded into third-party applications, the company noted, and offers complete customization. Its appearance, language style, and level of detail can all be tailored to a user’s specifications.
Because generative AI powers Auto, users can interact with the bot using natural language.
“We use GPT4 for the backend — for figuring out what the user is asking for and how to answer the question,” explained MicroStrategy Executive Vice President and Chief Product Officer Saurabh Abhyankar.
“The difference between MicroStrategy and a general purpose large language model is that in addition to the cognitive skill the LLM has, we add an analytic data structure,” he told TechNewsWorld. “So if you ask how many hats do I have at store X, the LLM figures out what the user is asking, and the MicroStrategy layer executes the query, brings the data back, and applies security and rules for calculating inventory.”
“You need both things in an enterprise analytics scenario because a chatbot like ChatGPT doesn’t have the context, business knowledge, security, and governance required to answer a question like that,” he added.
Unlocking User Value
Empowered by AI, Auto can remove barriers to fast, effective decision-making by making applications smarter and putting enterprise analytics in the hands of users no matter what skill level or application they’re using, the company maintained.
There’s no need to use a complex dashboard to get insights, and users can ask for information in ordinary language, making it effortless to incorporate business intelligence into business decision-making, it added.
“We think using MicroStrategy AI will unlock huge value by providing a variety of users with deeper insights that previously required more clicks and more granularity to understand. It’s powerful for user self-service,” Nena Pidskalny, director of supply chain strategy and planning for Federated Co-operatives Limited, said in a statement.
“Giving more employees access to business intelligence data can benefit a company by fostering informed decision-making across departments, enabling agility in responding to market changes, and promoting a culture of data-driven decision-making,” added Mark N. Vena, president and principal analyst at SmartTech Research in San Jose, Calif.
“However, easier access to business intelligence data may lead to potential harms such as data breaches, misuse of sensitive information, and compromising competitive advantage if not properly managed and secured,” he told TechNewsWorld.
Customized generative AI bots have some advantages over general-purpose bots like ChatGPT, Gemini, and Claude, noted Rob Enderle, president and principal analyst at the Enderle Group, an advisory services firm in Bend, Ore. “Generally, they are more focused and able to do one or a few things well and potentially better,” he told TechNewsWorld. “They’re also able to run locally because they use smaller libraries.”
Enderle added that customized enterprise bots can also be safer than general-purpose bots. “They generally are derivatives of large LLMs,” he explained, “but because they are reduced and more focused, in theory, they are less likely to do things you don’t want done.”
Tackling Concerns About AI
Custom generative AI bots can also address businesses’ concerns about data sharing with large chatbots. “There’s always anxiety if you’re offering up your or your customers’ proprietary information to a tool that’s going to iterate on that data and may re-present it in some way down the road,” said Will Duffield, a policy analyst at the Cato Institute, a Washington, D.C. think tank.
“Consumer-centric bots are allowing the firms behind them to use your conversations to make the bots better,” he told TechNewsWorld. “That wouldn’t be the case with a lot of these business tools because how the information can be used will be contractually specified.”
“Enterprises don’t want to send all their data to a general-purpose LLM,” Abhyankar added. “They don’t want to train the LLM with their data because of the risk of that data leaking.”
With MicroStrategy, he explained, data is stored in the customer’s environment. Only bits of metadata get sent to our LLM and the LLM isn’t trained with that data. “We can do that because MicroStrategy runs the calculations, and because the LLM doesn’t need to do that, it doesn’t need all the data,” he explained.
For that same reason, the LLM can be prevented from hallucinating. “LLMs, by their nature, are probabilistic,” Abhyankar said. “You can ask it questions, but you can get different answers for the same question. That’s not ideal for a business scenario.”
By running calculations in the MicroStrategy layer and doing them based on the business logic that the customer has encoded in our platform we can avoid probabilistic problems, he maintained.
“So challenges of data sharing and hallucinations are largely removed because of the way we use the LLM only for cognitive skills, and we use the customer’s data in the MicroStrategy layer in a trusted fashion,” he added.
Pumping Up Productivity
Making business intelligence more accessible to enterprise personnel can have productivity benefits. “It should allow decision makers to make better and more timely decisions, resulting in greater operational success,” Enderle said.
Data analysts, in particular, should see productivity gains from the self-service aspect of MicroStrategy Auto. “It makes data analysts more productive because they can do more in the same amount of time,” Abhyankar said.” It’s a productivity boost for them.”
“When the end-user can serve themselves, it affords the analyst key benefits,” he continued, “They get freed up to focus on higher value things because they are dealing from fewer questions and requests from end-users.”
Sharad Varshney, CEO of OvalEdge, a data governance consultancy and end-to-end data catalog solutions provider in Alpharetta, Ga., noted that generative AI technologies are profoundly impacting data analytics across the board. “They remove the complexity of data discovery, enabling teams like marketing or HR that aren’t traditionally analytics-focused to use company data assets easily,” he told TechNewsWorld.
“However,” he said, “the data received must be accurately governed. While a generative AI tool can quickly find and contextualize data, it doesn’t account for data quality, lineage, or access.”
“Once data is discovered, policies must be in place that ensure the user requesting the data has the relevant access permissions to extract it,” he continued. “Then, it needs to undergo various quality measurements for duplication, inconsistency, and other factors before being classified and cataloged. Only then will it be suitable for analysis.”
“Thankfully,” he added, “tools are available that can automate these governance processes and others that make data analysis and visualization very straightforward.”
