Application UI is overrated. As people, we’re visual creatures– we ascribe value to what we can see, because it’s what we can comprehend. What goes into serving that UI, and what lies underneath, are largely incomprehensible, and thus undervalued.
At the same time, the UI of today’s application is the application. The UI is an instantiation of the underlying data model of the application, made consumable for people. The UI controls the nouns and the verbs of the business.
As I’ve written extensively before, enterprise applications, to over simplify, provide opinionated frameworks for how a specific business function or business process should be done, and enable you to do (some portion of) those tasks. In order to accomplish this, these applications provide a data model, which again provides an opinionated view into how that function or process should be viewed and structured.
The application UI thus controls the business. Because today, people are the ones that largely conduct business. They’re the ones with hands on keyboards, senders of emails, maestros of excel macros. People are the engine that makes everything work. In this world, the UI is the way an organization sets the guardrails for thousands of interrelated workflows that make a business run. But it’s ultimately a facade for underlying data and workflow.
Very simply, an app today looks like this.
At the risk of reductionism, a business is really just a bunch of workflows. And a workflow is really just a series of actions that causes one field in a database to change to another. Open Support Ticket to Closed Support Ticket. Lead to Closed/Won Opportunity. Open Job Rec to Hired Employee.
If you investigate any valuable business, you’ll realize that enterprise value rises as the Objects owned are more critical to running a business. It’s why Salesforce, Workday, and ServiceNow are huge, and most other businesses by comparison are not.
Because Salesforce owns the customer, Workday owns the employee, and ServiceNow owns your technology. And what is a business if not employees, customers, and technology?
Salesforce has built an empire with this ownership of the data model. You don’t touch the customer without talking to Salesforce.
As you’ll commonly hear, "If it's not in Salesforce it didn't happen." Salesforce’s data model, for all its perceived shortcomings, is the lingua franca of business. If your application doesn’t talk Salesforce, it doesn’t make money. It’s why the AppExchange is what it is. You can see this in the enterprise value of all related sales and marketing technology. The 2nd best player doesn’t even come close. It’s the same story for Workday and ServiceNow. You might call this objective imperialism. Own the business objects, own the ecosystem, own the value.
In today’s world, the application UI is how a business turns ownership of this data into value. And so as users we ascribe value to the frontend, it’s really a facade for the data and the process. Users experience the UI, businesses reap the benefits of the process.
And because users generally don’t like the UI of most applications¹, the “what does Salesforce even do?” has become a common meme, while the literal Tower that dominates the San Francisco skyline reminds astute observers that Salesforce is probably doing something pretty valuable.
To put all of this simply… the application UI is both an overrated but necessary abstraction over the workflows to be done within an organization. The UI is how an organization makes a prescribed and opinionated process human-comprehensible, such that they can force adherence to it. After all, a business really is just a process machine, allocating resources as efficiently as possible. Iterating on and adhering to sales, marketing or product development frameworks are how enterprise value is created and protected.
The largest software businesses in the world have spent the last three decades riding ownership of these opinionated workflows to riches. And while consumers went a bit crazy when Prometheus arrived to give humanity fire in November 2022, the enterprise titans barely flinched.
But things have begun to change. First slowly, and now seemingly all at once. Over the last few quarters, Satya has declared that SaaS is dead, Benioff pivoted the entire company towards Agentforce mere weeks before the most recognizable conference in software, and McDermott is taking repeated shots across the bow at Salesforce’s crown jewel, CRM².
These headlines would be unthinkable just a few years ago. So what’s going on?
"Those who were seen dancing were thought to be insane by those who could not hear the music"
The enterprise giants have heard the music, and now suddenly everyone’s dancing. Almost two years ago Brad Gerstner said that “AI is going to be bigger than the Internet, bigger than Mobile” and while slightly hyperbolic, he’s probably more right than he is wrong. The titans of enterprise have decided that they too want to be more right than they are wrong. They see the crescendo is coming– and recognize that missing it represents an extinction level event.
But what is it and why does it matter?
It’s about the data, and it’s about AI. But more specifically, it’s about how AI fundamentally changes the way we can gather, understand, and act on data. It changes the nature of the abstraction between the data and the workflow. Because with AI, agents can act on data. At infinite scale and zero marginal cost.
Humans are no longer the only player in the workflow paradigm. This means that the total amount of work done within an organization will dramatically increase, but decoupled from cost and headcount. More code will be shipped, more agreements redlined, more vendor reviews conducted, more transactions audited.
People interacting with an application UI has always been the weakest link in getting work done. A manual and necessary evil. But with agents that can act on data to drive workflow, the idea that work can only be done by people via an application UI is blown to shreds. In the near future, UIs will likely be dynamically built by AI so that people can quickly step in and be fed exactly what’s needed to move the workflow along3.
There’s a new abstraction for work, and that abstraction is agents. The frenzy that you see in the market is because like the previous shift from on premises to the cloud, no one really has the incumbent right to win this market.
It’s an entirely new layer of software that has never existed. Crucially, it sits on top of existing layers of software, and is the layer at which the lion’s share of value will accrue in the future. Someone will win this layer, and with it build a software business of significantly more value than we’ve ever seen before.
With this layer we move from a world where people interact with application interfaces to get work done, to one where (1) people act with an agentic interface on top of the application to get work done, and (2) an agentic layer on top of these existing applications that actually does increasingly more of the work.
This shift relies on two major trends, which are recursively enabling of each other.
The first is agents, which have many definitions. I happen to like this one from Anthropic, which is that [Agents are] systems where LLMs dynamically direct their own processes and tool usage, maintaining control over how they accomplish tasks.
The second is data, the definition of which might be obvious– but a major breakthrough with AI is that we can now gather, process, and make sense of data in ways that were previously unimaginable. We also don’t need to specifically know what we’re going to do with the data before we gather it.
Historically, applications have been confined largely to the realm of structured data, for a number of reasons. First, is that these applications need to be human usable, which requires a simplification of everything. Very specific states, generally computed by people and adjusted in the UI, which then persists to the database. There really isn’t room for nuance.
You can visualize this with sales stages (though it applies equally to almost any business process in the application of your choosing). To move a prospect between sales, a lot happens. But the way it’s reflected in application is quite simple, necessarily. The seller will do a ton of work (research, meeting, emails, powerpoints, etc) that ultimately gets recorded in the system in a fairly simple way… from Stage 1 to Stage 2… Prospecting to Qualification… with a few mandatory fields to fill out.
But the texture, the granularity of what happened, is largely lost because there’s no way for the system as designed to comprehend or make use of it.
AI changes this fundamentally. Those call transcripts, those emails, those notes, those powerpoints– all crucial parts of the process with rich telemetry about interactions– can now be utilized in real time to paint a far richer picture of the relationship being built. Because AI, unlike people, can draw meaning from large bodies of unstructured information near instantaneously. And it can then write it back to systems in the format it’s needed.
This unstructured data doesn’t fit into the existing construct of the application, and so it’s largely discarded. The same problem exists across nearly every workflow– from sales to hiring to support to marketing. We lose the richness and texture of data, because it has to be fed to and utilized by structured systems operated by humans. We resort to a lowest common denominator of language to describe these processes.
And because agentic systems both create, and make use of this data, they create increasingly large data flywheels (which some might call moats).
Gone are the false promises of Data Lakes, and platitudes of “Data is the new Oil”... gone is the era that vendors (un?)knowingly insisted that customers gather as much data as possible, in anticipation of one day finding a use for it. This is real.
You can look at the meteoric and continued rises of Snowflake and Databricks. Or Data Cloud as Salesforce’s fastest growing product ever. They’re taking that a step further in bolstering it with their largest acquisition since Slack5.
These data platforms are a fundamental piece of infrastructure that enable agents, providing them with the context they need to act, as well as the corpus of data on which they act. The broader market is coming to understand that.
These data platforms will also likely argue that, because they already have that all-important data, they have as much right as any to build the agent layer on top.
The byproduct of this shift is that as agents do more work, and bring real time, deeper context across all relevant data to both people and agents doing work, the entirety of the existing application stack collapses to be little more than a data source and (for now) the keeper of workflow state (eg, the scoreboard– closed customers, new employee hired, support tickets closed).
In the CRM example, the data within the CRM becomes a small fraction of the overall picture of the customer. This has been, of course, always the case6. Most of the work happens outside of the CRM, and the data inside is mostly manually recorded by people. But by utilizing agents, you can get the full picture. And while a structured description of that full picture can be sent to the CRM to provide state, the real evolving picture of the relationship lives elsewhere– in agent memory.
Consider the following simplification.
A far more straightforward picture emerges, where the entirety of the existing application layer becomes merely an input to the data layer. On top of raw data, agentic systems bring context tailored specifically for the organization using it, creating an always-on layer of intelligent state, on top of which lives an interaction layer by which agents and people perform workflows on the data. The actions update the state, and the process continues.
The value is in the work. AI presents a new abstraction for work, and the entire existing software-industrial complex gets relegated to a data source feeding the data layer. And you still wonder why the SaaS giants are dancing?
But it doesn’t stop there. Today AI is largely used in an “agent in the loop” manner. That is, workflows are owned by existing software systems and agents are used by people to augment and amplify their ability to do the work prescribed to them.
But as we feed these systems increasingly large amounts of data, the logical next step is to move planning and orchestration from people to the system itself. After all, once we’ve determined the reward function– to close a customer, or hire a candidate, the system has far more information with which to plan and act to make that process a reality. If there’s one thing AI is good at, it’s bringing deep context to unfathomably large amounts of data.
If your goal is to hire 10 people, or close 10 customers, the system should identify who those people or candidates are. Given an unlimited universe of customers or candidates to pursue, a properly set up agentic system will be superior at identifying the best 10. And it will be able to provide to the user, with specificity, the reasons they are the best to pursue.
It will also be able to begin the pursuit, through an AI SDR or AI Recruiter, making us of the rich and dynamic context that led to it being identified as a prime target.
This moves business process from agent in the loop, to human in the loop, over time abstracting more and more of the work from people to agents. A business is just a process machine, dynamically allocating resources towards their most efficient use… and agents are a far more efficient and scalable resource to allocate.
This new interface to work becomes infinitely scalable and self-optimizing. Everything becomes a loop. And agents are great at loops.
¹ To be fair, most users don’t like the UI of any business applications. Because the application is essentially a proxy for their boss, lording over them and telling them what to do. The application UI represents work… and in the oft repeated words of a prominent poet, work sucks, I know. Your trusty enterprise application is just caught in the crossfire.
² While hiring a comical number of ex-Salesforce folks to build, market, and sell it.
³ Someone should write an essay on this.
4 Assuming, and it’s a big assumption, that you’ve designed your agentic system properly. If not, it’s going to regurgitate slop all throughout the gears of your beautifully constructed business process.
⁵ The market usually gets mad when Salesforce makes an acquisition, but I think this one is smart. They see that the future is agents, and agents need data. While Salesforce has a lot of data across products, and increasingly more in data cloud, the majority of enterprise data is still in legacy systems, both in the cloud and on premises. That data is governed and moved with Informatica. Owning Informatica takes them closer to that data, which gives them pole position to get into the places they can act on it— from Sales to Service to Agentforce. It also doesn’t hurt that they got it at a pretty big discount, and it’s generating cash. The market watched them walk through glass as they transitioned from license to subscription, and then scooped them up when they saw light at the end of the tunnel. Smart for the acquirer, but maybe a little bittersweet for the acquired.
6 Salesforce knows this, and it’s why Data Cloud has been and will continue to be such a high priority.
