|
|
|
Dave Zilberman is a general partner at Norwest focused on early to late stage investments in enterprise and infrastructure. NORWEST |
|
Good morning. Everyone knows that the sentiment around AI usage in the enterprise has shifted in recent months, but the nature of that change and its consequences are a matter of ongoing investigation. |
|
Dave Zilberman, a general partner at Norwest, begins with the now widely held premise that the corporate world’s embrace of artificial intelligence remains very much intact, albeit with a more pragmatic approach that gives priority to controlling expenditures and driving a return on investment. The next question revolves around how companies define a meaningful return on AI. |
|
To keep close to the markets, he relies on insights drawn from the Norwest Operator Collective, a community of about 90 to 95 Fortune 500 tech leaders, including chief information officers and chief information security officers. Zilberman, who focuses on early to late-stage investments in enterprise and infrastructure, said members of the NOC are uncompensated to avoid conflict of interest. |
|
“That provides us with invaluable insights about the ground truth as to what’s really going on with customers versus what the entrepreneurs may aspire to, or what the press may say. What are the actual consumption patterns?” says Zilberman, who during a varied career also worked at Comcast Ventures and Lehman Brothers, among other places. |
|
Zilberman says companies moved beyond an initial “promiscuous” era of AI experimentation. But he also thinks they have skipped the dreaded Trough of Disillusionment, a phase that is part of research firm Gartner’s Tech Hype Cycle. He thinks that companies have skipped to the next phase of the cycle, the Slope of Enlightenment, which is the step that occurs before the Plateau of Productivity. (Nota bene: the aforementioned “promiscuous” stage isn’t part of the Gartner framework.) |
|
“How do we operationalize AI? That is the big focus we’re seeing from the NOC community,” he said. “All the noise in the capital markets doesn’t really impact them. They are focused on their business, they are focused on how they capitalize on the benefits of AI.” |
|
Cost cutting vs. growth. Members of the NOC aren’t primarily focused on using AI to cut jobs and reduce costs, according to Zilberman. “It is much more about efficiency and scale to bring products to market. It is how do we launch new products quicker into the market? It is the commercial aspect of the ROI rather than the cost savings side of ROI,” he said. |
|
Even so, companies are increasingly focused on understanding and controlling their spending on AI, just as they came to develop tools and processes for controlling their spending on cloud computing, according to Zilberman. “Enterprises don’t really today report on AI spend. But they will. They absolutely will,” he said. |
|
There will need to be continued innovation around the cost curve, according to Zilberman. On one level, that can mean building tools to help companies measure the ROI on their AI spending. Orchestration, coordination and security of AI and AI agents need to be addressed as well. He perceives growing demand for small language models, too. “I think there is a big opportunity… for very discreet edge-based applications…where you don’t need a large language model that has trillions of parameters. A few hundred million parameters is sufficient. It is obviously lower cost, there is greater efficiency,” he said. |
|
How has your company’s approach to deploying and managing AI evolved in recent months? Let us know. |
|
|
|
