If you are an organization, there is a reason you exist. An economist’s answer in the most fundamental way would be:
- You produce something (create value) and then exchange/trade that something (exchange value).
- You do this often within the context of your ecosystem domain (i.e. an industry) as part of a larger value chain.
- The ecosystem domain gets further mapped into industries and sectors across geographies with specific organizations as end points.
- The global financial systems provide an exchange framework for assigning and trading value globally across the ecosystems.
- Efficient markets strive for liquidity in representation and exchange of value.
If you zoom out, every organization is a link in a more complex and intertwined graph of value exchange which is concentrated around domains. If you zoom in, each link represents a value producer and consumer relationship. In the most fundamental way, a customer is the other end of the value chain link. It can take on the persona of a human being as an end consumer, an organization, a digital intermediary, or more complex, compounded entities.
It’s important to view this definition of a customer within the context of another undeniable fact – the global economy is increasingly becoming more digital. That means both ‘value creation’ and ‘value exchange’ are getting more digital. Hence the “customer” is getting more digital.
Now data is the lifeblood of the digital economy. Let’s discuss that next.
Data: The language of the digital world
In some sense it all starts and begins with data. Data is the footprint of doing digital business and can be considered as the liquid currency of value exchange in the digital economy.
There is no inherent meaning in data.Meaning emerges with interpretation and with meaning comes value. So the real value of data lies in its meaning (semantics). The rest is just an encoding to make it talk to the digital substrate (CPUs/GPUs). This is key to understanding the AI revolution.
Value (semantic) layer
The meaning of data emerges from the context provided by the ‘humans’. So the top end – at the system of engagement, where humans interact with domain-aware systems, is the value layer. It represents the digital encoding of the semantics we care about in the real world. So a simple number at this layer could be someone’s medical test result, something very personal and impactful or it could be a kids ‘Hello World’ playground.
Overhead (syntax) layer
On the other end, we have the overhead ( syntax layer), which is concerned about various forms of data/format manipulations so that eventually we are able to execute on the digital substrate. At the substrate level, data is just 0s and 1s with manipulating logic. We have to transform this encoding to allow various digital systems and humans to communicate effectively. This is essentially what IT does in an organization. Given the complexity of their IT and digital landscape, developers and IT teams perform the difficult overhead of making data jump though various systems, data formats, transport protocols, platform runtimes, and so on.
So from an economic perspective, the semantic (meaning) layer represents the business value of the data and the rest is largely the implementation overhead. There is a semantic leak (loss in meaning) at each of these junctions and an increased syntax overhead (coordination between multiple teams, technology vendors, and digital platforms).
First to semantics
The big promise of AI is to get to the meaning (semantics) of the data early on and enable autonomous agents to digitally negotiate and implement on behalf of their stakeholders.
And for that you need ‘context’ – the backdrop in which a task is performed. The business domain in and the customer lifecycle context (C 360) together define the complete context.
Context as intelligence
Intelligence here can be defined as the ability of a software component to be ‘business context’ aware in real-time. This real-time context is intelligence. This context and understanding of the semantics (early on) provides the basis of driving ‘agency’ in Autonomous Agents. This enables software assets to understand their goals and make autonomous decisions to chart a course to achieve them. As software systems and AI model the real world, the constraints and the relationships inherent in the real world entities are carried over to the digital world. This allows the customer context and domain context to provide the right semantic backdrop to model the AI enterprise.
Salesforce is the most natural platform to provide this customer and domain context. Let’s discuss that next.
Salesforce: Data and metadata for customer and domain context
With the previous discussion on customer and context (the backdrop that provides meaning), lets understand what it means for Salesforce to be a Customer Company along three key dimensions.
The Ecosystem Domain Context: Salesforce Industry Clouds
Salesforce industry clouds set the ecosystem domain context. They provide a system of engagement that connects the humans to business processes within the context of their domain and role (e.g. loan origination). They encapsulate the essence of the ecosystem: canonical representation of the data types, industry standards, compliance and regulations, innovations, customer expectations, channels of engagement, industry-specific processes, and a lot more.
The Customer Context: Salesforce Customer 360
The reason an organization exists is to be able to provide value to its customer. Now sales, service, marketing, and commerce represent the most fundamental areas of managing the entire customer lifecycle. The Core Salesforce clouds provide this functional context. They help organizations build a 360-degree view of their customers and help execute on strategic initiatives like launch a marketing campaign or a new sales and distribution channel.
Data and AI: The Salesforce Intelligence Platform
The Salesforce platform provides the capabilities and the architecture to manage the entire lifecycle of the digital assets needed to run your organization, including Agents. These digital building blocks power the system of engagements across various clouds. Data Cloud has access to all the data to build a coherent view of your business, and Agentforce infuses intelligence at every touchpoint.
So Salesforce not only provides the integrated AI platform but is unique in its ability to also provide the critical data and the metadata that establishes the customer and the industry domain context.
The AI enterprise
Having established a perspective on the foundational elements, lets see how this all comes together as an AI enterprise powered by Salesforce. This framework provides the complementary value proposition of technical capabilities and the impact they can have on an organization’s future.
Human-machine interaction
This layer interfaces humans with systems within the context of their role. Humans interact with one another in natural language. For the first time in human history, generative AI – through the use of LLMs – has enabled humans and machines to communicate in a natural language. With vectored embeddings, this can be extended to pictures, text, and videos as well. This is where the Agentforce provides a conversational copilot and the platform to templatize and customize actions and prompts. It caters to the fact that organizations often have their foundational models, custom models, and data spread across multiple domains and vendors.
Computable enterprise
It is critical to note that AI is conversational right now, but increasingly it will become computational, meaning we should be able to ask it to do stuff as well. This leap from conversational to computational is the basis for Agentic Execution.
We are headed toward an autonomous agent-based software architecture that drives the digital economy. In this architecture, software entities within a domain can discover and automate the orchestration of the relevant digital assets, cooperate and coordinate with each other, track and manage the value exchange, and design processes to model its incentives (reward functions).
This agent-based discovery, negotiation, and process execution – by orchestrating your enterprise’s intelligent digital building blocks – is the defining feature of a computable enterprise.
This layer provides the action framework necessary to orchestrate AI. It can discover and automate the orchestration of the relevant digital assets made available to it. This layer relies on the composable enterprise for its execution. More specifically, the concept of the intelligent digital building blocks (iDBB) that represent the “digital vocabulary,” or software representation of the “nouns and verbs” of your business.
For instance, a sentence written in a natural language at the system of engagement that has access to the iDBB can drive automation. This context provided by the iDBB is what is needed to ground LLMs (with trust) to make generative AI work for any business. The effectiveness of the orchestration layer is very closely tied to the quality of the composable enterprise that drives it.
The composable enterprise
This is the layer responsible for converting the siloed and complex enterprise into a composable enterprise that can allow an organization to drive AI. MuleSoft and Data Cloud are critical in building the composable enterprise. Together they provide the iDBBs’ data product and the associated actions, events, and triggers that the computable enterprise can use. iDBBs implement the interface for business context, business value, APIs, Governance, Data, and data ports. Data Cloud has native access to the customer context in Salesforce and MuleSoft helps extend the scope of this ecosystem by integrating any data from any system from anywhere creating a true Customer 360 and augmenting the domain context. This layer provides the digital vocabulary that can be used to enable Agent based execution in the enterprise.
The complex enterprise
This layer represents the reality of the complex and heterogeneous digital and IT landscape that most organizations find themselves in. With nearly 1000 different applications running on different platforms and rarely communicating with one another, this is where IT teams end up with the overhead layer. They spend a lot of time just making data and apps jump over multiple barriers. A dirty enterprise with trapped data cannot drive the AI orchestration layer.
MuleSoft and the AI enterprise
MuleSoft harnesses the power of AI through APIs, delivering value wherever it can be effectively applied: whether in traditional applications, autonomous agents, or Salesforce’s system of engagement. It can:
- Solve the backend complexity of an enterprise through the use of domain context-aware agents
- Infuse intelligence in the applications by providing the customer and domain context from the enterprise
Through its integration, universal API management, and automation capabilities, it helps organizations become composable, connected, and automated in their ecosystems. MuleSoft helps build a composable enterprise and drive automation by agentic execution in the computable enterprise.
This composable, flexible, plug and play structure of the enterprise built from the digital building blocks is what makes the enterprise agile and AI ready.
The AI revolution can be a disruptive threat, but also presents incredible opportunities for growth and transformation. The AI enterprise in an AI economy can allow organizations to find new forms of efficiency in their operations, scale their productivity, pioneer new business models, and unlock the art of the possible.
