Raj Verma is the CEO of SingleStore. He brings more than 25 years of global experience in enterprise software and operating at scale.
getty
Silicon Valley Bank’s collapse in March 2023 made a major impact on the tech industry. The loss of a key financial institution and a valued source of funding for startups left some concerned about how the tech industry could access much-needed capital to continue innovating.
While venture capital funding declined in 2023, the past year was still big for generative artificial intelligence, which brought a wave of new investment into Silicon Valley. I believe the hype is well-deserved. For businesses in particular, the possibilities associated with AI are tremendous. This tech can help companies become more efficient, offer an increased range of services and create new insights that help them better understand their own data and customers’ needs.
But to help AI get there, we need to truly understand the role that data technology plays in its development. AI’s value is intrinsically tied to the quality and accessibility of the data it relies on. Without data, AI is useless. With the wrong data, slow data or inaccurate data, AI will be nothing more than the latest upgrade of a computer, or even dangerous in the worst-case scenarios. What, then, do data companies need to do to unleash AI’s full potential?
Being the CEO of a company that provides a data platform for AI applications, I’ve seen firsthand that successful AI data platforms need to possess three attributes: speed, simplicity and scale. Think of these attributes as key tenets of the data technology world in the age of AI. Without them, data platforms will fail in their mission to power this technology. Let’s break them down:
1. Speed
Speed refers to a platform that can process, analyze and contextualize information in milliseconds. AI systems that operate on a speedy platform will analyze data faster, therefore allowing for more accurate real-time predictions. Speed becomes especially critical in the modern world of AI-enriched applications and use cases, such as customer service chatbots, real-time recommendation engines, co-pilots and so on. Organizations need to be able to ingest, process, search and analyze diverse sets of structured, semi-structured and unstructured datasets in near real time to provide an engaging user experience.
However, developers and companies should balance their pursuit of speed with their security needs. Speed and security are often framed as being at odds with each other, with the common perception being that increased security measures can slow down a system. Instead, think of security as an accelerator—one that allows businesses to assess and mitigate risks that otherwise would lead to more serious problems.
Developers and businesses should also ensure they understand the type of risks their systems face and employ security measures that specifically address them. For example, a financial institution’s data platform might need authentication and encryption capabilities. By sticking to the necessary security features, companies building AI data platforms can maximize safety without slowing their systems—or business—down.
2. Scale
Scaling is about simplifying the complex. A scaled data platform is one that can manage increased demands in an organized and simple fashion. Not all data is created equal.
In our industry, we sometimes talk about data in temperature-related tiers: hot, warm and cold. Hot data is data that is used frequently and therefore has to be accessed quickly. Without this data, AI cannot produce an answer. Warm data is accessed less frequently but is still needed for analytics or reporting purposes. Cold data is historical data; this data is rarely accessed.
Know the type of data you are working with. This is key for scaling purposes. When you categorize your data and have your platform and AI system prioritize which data it retrieves—hot, warm or cold—you can scale it and therefore make it more efficient.
3. Simplicity
Simplicity refers to one data platform processing multiple types of data, transactional and analytical, and then adding on capabilities, such as vector similarity search and contextualization capacities. Not every data platform does this; I’ve found many focus on processing either transactional or analytical workloads. But, from my perspective, this does not always make sense for AI’s needs. Developers might consider creating a central data platform that allows AI to access all of the data necessary, no matter its type.
Lastly, remember that the platform itself needs to be simple to use. It shouldn’t require complex training or prior expertise to navigate, and its features and capabilities should be easily accessible to users.
Understanding these principles now is key because generative AI is in its infancy. I believe that in a few years, we will likely look at this era the same way we reflect on the first cell phones, which were as big as bricks and provided what felt like an hour of battery at most. I expect AI’s evolution to come about much faster than previous technologies, with data technology playing a crucial role in determining its pace. By embracing the three tenets of speed, scale and simplicity, data platforms can power AI to elevate businesses to new heights.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
