Digital Twins Move from the Asset to the Enterprise

Every quarter, finance teams model disaster scenarios. What if the company loses its biggest customer? What if a competitor slashes prices? Or how — the marketing team wonders — might customers respond to a 10% price hike?

These exercises help guide decisions. But the process is based on a flattened version of reality that reduces a living, adaptive organization to a handful of static assumptions and outdated information on a page. It’s too abstract to gauge how customers actually behave, how teams respond under pressure, and how second- and third-order effects — like supplier delays or ripple effects across departments — can cascade throughout a business.

Now imagine you could simulate the enterprise in a far richer way — customer journeys, sales strategies, organizational design, even competitive responses. What if leadership could war-game a radical move, like reorganizing around customer relationships instead of internal functions, without putting real revenue, reputations, or jobs at risk?

This is the promise of the enterprise digital twin: a virtual replica of the business that lets executives pilot strategies, stress-test assumptions, and explore alternatives before going live. In the physical world, digital twins have transformed how factories, refineries, and supply chains are designed and optimized. Formula One teams, for instance, have been using digital twins for years to simulate race strategy in real time, testing thousands of pit stops and tire decisions before a single lap is run. 

But now, powered by agentic AI, the enterprise digital twin could bring that same transformation to knowledge work itself.

A workable enterprise digital twin would give leaders a way to explore strategic tradeoffs, anticipate unintended consequences, and learn faster — without paying the traditional real-world costs.

Building an enterprise digital twin is different from modeling machines. Simulating knowledge work means capturing why decisions get made — company values, policies, workflows, and decision traces that shape organizational behavior. Today, a CMO typically sees the world through marketing data, while a CFO typically sees the world through a financial lens. A working enterprise digital twin changes that, giving each function a view of the whole system, not just its own corner of it. (eVerse, created by Salesforce AI Research, lets teams simulate large volumes of customer interaction and explore how systems behave under realistic but controlled conditions. It offers an early illustration of what’s becoming possible.)

The concept is still nascent, but the implications are profound. A workable enterprise digital twin would give leaders a way to explore strategic tradeoffs, anticipate unintended consequences, and learn faster — without paying the traditional real-world costs. The companies that benefit most will be the ones that start laying the groundwork now.

The Integration Breakthrough

For years, this leap toward full-scale enterprise modeling remained mostly theoretical. The problem was data integration. Enterprise data lives in dozens, sometimes hundreds, of disconnected systems: CRM platforms, supply chain databases, payroll systems, and product usage logs. Critical information can be buried in legal contracts, spreadsheets, and internal strategy documents.

Stitching these fragments together into a coherent, living model has traditionally required massive, bespoke engineering efforts so expensive that only organizations like the CIA or the Pentagon could afford it.

For example, before launching Agentforce Voice, Salesforce’s AI voice platform, engineers stress-tested the system inside eVerse, revealing failure modes traditional testing would have missed.

That barrier has begun to fall: AI coding agents can now orchestrate data integration in days. These agents can handle schemas, APIs, permissions, and business logic well enough to connect systems with far less need for custom code than before.

What’s emerging is still not a finished model of the enterprise but a practical way to experiment with parts of it.

Enter the Sandbox

For example, before launching Agentforce Voice, Salesforce’s AI voice platform, engineers stress-tested the system inside eVerse, revealing failure modes traditional testing would have missed. The agents struggled with regional dialects. They misinterpreted overlapping speakers. They broke down when customers shifted tone mid-conversation. These weren’t hypothetical risks but scenarios the system would face immediately in production. Finding them in simulation meant fixing them before customers had to experience such shortcomings.

UCSF Health is now using eVerse to train AI billing agents in healthcare, where only 60%–70% of inquiries follow documented procedures. The rest require judgment, institutional knowledge, and pattern recognition across messy, incomplete records. In eVerse’s simulated environment, clinical experts provide feedback based on the answers the AI agent offers, correcting mistakes and validating strong answers. The upfront work may be tedious for those clinicians, but early results show that trained AI agents can handle up to 88% of cases, freeing human experts from answering the same question over and over.

These are specific use cases and single workflows. But they point toward the kinds of architectures companies will need as they begin building toward a full enterprise digital twin — systems that can coordinate data, simulations, and AI agents across functions rather than in isolation. As enterprises integrate more data sources, these sandboxes grow richer. Together, these components lay the groundwork for moving from isolated simulations to something closer to a simplified model of organizational behavior itself. Salesforce AI Research calls this Enterprise General Intelligence, or EGI. 

Early Implementations 

The path from narrow training environments to enterprisewide simulation is already underway, and the results can be striking, even at the process level. The restaurant chain Wendy’s, for instance, built a digital twin that integrated its 3,500 trucks, 34 distribution centers, and 6,450 restaurants. When the company faced a syrup shortage, the system identified the problem and simulated solutions in five minutes. In the past, such a task would require more than a dozen people working a full day. The drugstore chain Walgreens scaled a similar pilot from 10 stores to 4,000 in eight months. 

But simulations are not crystal balls, and they’re based on data that is sometimes incomplete, siloed, or simply not in a form that models can use. Designing interfaces that allow companies to meaningfully explore scenarios with dozens of variables remains difficult. The human brain reaches its limits when it tries to comprehend multidimensional models, and some forces that impact real-world behavior, like the attribution problem in marketing (Which of the many possible marketing leverage points was decisive in someone buying that product?), sit partially outside the reach of any model. The value lies not in achieving certainty but in exploring possibilities more rigorously than intuition alone allows.

Former Secretary of State Colin Powell, who rose to four-star general and Chairman of the Joint Chiefs before his diplomatic career,  served on Salesforce’s board until his death in 2021. He forged his “40-70 rule” through decades of high-stakes military decision-making. Most decisions must be made with incomplete information, but below 40%, you’re guessing; above 70%, you’ve likely waited too long.

Strategic simulations don’t push that 70% to 100. But they help leaders get more value from the information they do have — surfacing hidden assumptions, stress-testing them across multiple futures, and sharpening the questions that matter most. 

Your Move

Even if the full enterprise digital twin isn’t yet here, organizations don’t need to wait to start preparing. The path forward begins with small, bounded simulations — modeling specific workflows, decision-making trees, or customer interactions where outcomes can be tested safely before being deployed in the real world.

As the technology matures, the opportunity moves beyond tools to organizational capability. The advantage will go to organizations that learn fastest by running more simulations, testing more alternatives, and surfacing failure modes in synthetic environments rather than in business operations that could damage the company.

Start experimenting now, then expand as capabilities mature. Enterprise digital twins won’t arrive all at once. They’ll be assembled piece by piece, by organizations that treat simulation not as a one-off experiment but as a core way of making decisions.

More stories by the Futures team: 

• How the Agentic Wave is transforming enterprises
• Why imagination must shape the future of work in the agentic era
• Top predictions and trends to watch in 2026 for AI agents