Why AI solutions aren’t moving to market as quickly as imagined

With all the buzz around ChatGPT and the rapid mainstreaming of generative AI, 2024 was predicted to be the year of AI. While the market certainly talks a lot about AI this year, we’ve yet to see much of it in production environments.

Events are a great chance for tech companies to showcase or announce new innovations to the market. At RSAC in May, we couldn’t help but notice the "AI-washing" that perpetuated across the showroom floor, even as we were happy to be one of the handful of companies able to give live demos of our new Sumo Logic Copilot capabilities.

So I had to ask myself, "Why so much talk about generative AI, yet so few live, demonstrable capabilities?" I gathered more insight into why this might have been the case as I spoke with industry experts in security, operations, and engineering.

This is by no means a complete list, but these conversations with peers shared similarities and common points.

AI needs to solve the right problem

Many technology vendors are still struggling to identify the valuable features that can be built from generative AI. With so much pressure from the market (and in certain cases from companies’ boards of directors), some vendors are compelled to get something, anything, out to market – a technology solution looking for a problem to solve.

If you have a value that can best be created from a System of Insight, it's the right fit for purpose. We certainly see our implementation of Sumo Logic Copilot in this light. We believe there is an entire class of Systems of Insights in observability and security that can help customers and go beyond a shiny but useless AI button.

Performance issues can slow the pace of AI innovation

Even companies intimately familiar with generative AI tools have seen unacceptable performance when implementing the technology on top of existing platforms. Obviously, a new set of "round trips" now needs to be considered, but these performance issues go beyond just adding a new layer of technology.

Other technology vendors have demonstrated their copilot-esque capabilities, but I’ve seen them take several minutes between questions. At Sumo Logic, the lesson here is that if you do not already have a performant, scalable environment to handle big data, you are starting on a shaky foundation when building these new, next-generation capabilities. Thankfully, the Sumo Logic Platform can handle hundreds and hundreds of TBs per day for single customers (adding up to a whopping 3.5 exabytes of data analyzed daily) without issues.

AI pricing models aren’t baked yet

I recently attended a conference full of CEOs, CTOs and CPOs, and the most common topic was monetization strategies for generative AI capabilities.

Implementing these new features costs money, which must be recouped. Most leaders rallied around one of two approaches: either offering a "Super" Enterprise version or SKU that includes these new capabilities or creating a whole new product centered around generative AI capabilities. Of course, there are upsides and downsides to either of these perspectives.

My perspective is a bit different. At Sumo Logic, we’ve seen firsthand that when you introduce consumption-based pricing, you worry less about recouping costs and focus on maximizing value for each feature to all customers. It is one of the reasons why we launched our $0 ingest pricing earlier this year. Customers can send all their data and pay for the value they receive (scans, searches, dashboards, alerts, analytics, etc.).

This includes our generative AI capabilities. Our product and engineering teams work with customers on features that truly add value, and customers pay if they see the value of the capabilities and use them. We don't have to limit AI to a a subset of customers (like a "Super Enterprise SKU") or wait until an entire end-to-end AI product is developed (with its own potential product-market-fit issues) before customers receive value.

Is 2024 still the year of AI?

Halfway through 2024, I wonder how other companies are progressing in addressing these problems. Events during the second half of this year will create lines in the sand for organizations hoping to capitalize on the AI hype this year. I’m excited to hopefully see more live demos and continue the conversation with peers about how to deliver true business value from AI innovation.

Learn more about the impact of AI on the market and the future of fifth-generation SIEM.