6:35am PST • December 9, 2023
Photo credit: Arthur Debate / Getty Images
Jonathan Golden Contributor
Jonathan Golden is a partner at NEA and former product director at Airbnb. More posts from this contributor
- Changing consumer behavior is the key to unlocking multi-billion dollar businesses
Generative AI is a paradigm shift in technology and will drive a massive shift in business spending over the next decade and beyond. Transformations of this magnitude can feel fast on the surface, especially when they make a big splash like generative AI has in recent months, but it’s a steep and steady climb to penetrate the layers of the enterprise technology stack.
The infrastructure layer captures the initial spend as companies put together the building blocks for power and performance – the capital flowing into Nvidia and GPU aggregators today shows this is well underway. As adoption (and dollars) grow, development focus will shift to the new experiences and products that will reshape each subsequent tier.
We’re just getting a taste of how this change will play out at the application level, and early signals suggest that the disruption will be profound.
Long before generative AI, enterprise applications began to offer more consumer-like experiences by improving user interfaces and introducing interactive elements that would engage everyday users and speed up workflows. This led to a shift from “system of record” apps like Salesforce and Workday to “system of engagement” apps like Slack and Notion.
As generative AI shapes the next generation of application products, we can expect to see even more widespread development.
Collaboration has been a defining feature of this new generation of business tools, with features such as multiplayer, annotation, version history, and metadata. These apps also leveraged consumer-owned viral components to drive adoption and enable seamless content sharing within and across organizations. The core data set retained its intrinsic value within these interaction systems and served as the foundation for the growing amount of information created at the interaction level.
As generative AI shapes the next generation of application products, we can expect to see even more widespread development. The early players are similar to ChatGPT integrators who build lightweight tools directly on top of generative models that deliver immediate but fleeting value. We have already seen a variety of generative AI products on the market that initially show explosive growth, but also extremely high churn rates due to limited workflows or lack of additional functionality. These applications typically produce a generative output, which is a single-use content or media type (i.e. not embedded in a user’s everyday workflow), and their value relies on standard generative models that are widely used by others in the world are available market.
The second wave of generative AI applications, which is just beginning to take shape, will leverage generative models to combine the structured data that resides in system-of-record applications and the unstructured data that resides in system-of-engagement applications. Applications lie to integrate.
Developers of these products have more potential to create lasting businesses than new entrants, but only if they find a way to “own” the layer on top of the system-of-engagement and system-of-record applications – no easy task though Established companies like Salesforce are already scrambling to implement generative AI to create a protective moat around their underlying layers.
This leads to the third wave, in which participants create their own defensible “intelligence system” layer. Startups will initially introduce novel product offerings that add value by leveraging existing system-of-record and system-of-engagement capabilities. Once a strong use case is established, they will develop workflows that can ultimately function on their own as a true enterprise application.
This does not necessarily mean replacing the existing interactive or database layers. Instead, they will create new structured and unstructured data, with generative models leveraging these new data sets to improve product experiences – creating a new class of “super data sets.”
A key focus of these products should be integrations with the ability to ingest, clean, and label data. For example, to build a new customer support experience, it is not enough to simply inherit the knowledge base of existing customer support tickets. A truly compelling product should also include bug tracking, product documentation, internal team communications, and much more. It will know how to extract, highlight and weigh the relevant information to gain new insights. It has a feedback loop that allows it to improve training and usage, not only within an organization but also across multiple organizations.
If a product does all of this, switching to a competitor becomes very difficult – the weighted, cleaned data is very valuable and it would take too long to achieve the same quality with a new product.
At this point, the intelligence lies not only in the product or model, but also in the hierarchy, labeling and weighting associated with it. Delivering insights will take minutes rather than days, with a focus on actions and decisions rather than just synthesizing information. These will be the true system of intelligence products that leverage generative AI and have the following key features:
- They have deep integration into company processes and can capture newly created structured and unstructured data.
- Be adept at characterizing and processing data through hierarchy, labels, and weights.
- Create data feedback loops within and between customers to improve product experiences.
An important question I like to ask clients is: “How does a new product stack compare to the other tools you use?” Typically, the system-of-record product is the most important, followed by the system-of -Engagement product, with additional tools at the end of the list.
The least important product is the first to be cut when the budget is tight. Therefore, new systems of intelligence must provide lasting value in order to survive. They will also face strong competition from established companies that will integrate generative AI-powered intelligence capabilities into their products. It will be up to the new wave of systems of intelligence to combine their offerings with high-quality workflows, collaboration and the introduction of super data sets to endure.
The transformation of the AI space has accelerated over the last 12 months and the industry is learning quickly. Open source models are increasing and closed proprietary models are also developing unusually quickly. Now it’s up to founders to build lasting systems of intelligence products on this rapidly changing landscape – and if done right, the impact on businesses will be extraordinary.