Some recently claimed that with AI and agents, they can replicate previously assumed large and complex software products in weeks. If that’s true, our three-month or three-quarter projects are becoming more ambitious, with more components and interdependencies.

As projects scale in complexity, managing dependencies becomes incredibly challenging. Changing one component can ripple across the entire system. It’s nearly impossible for an unassisted human to track these relationships and predict how changes will affect each component. In our interviews with 1,000 knowledge workers, we found that current practices—status updates, documentation, and ad-hoc messaging—are ineffective at keeping teams aligned on complex projects.

In software product development, understanding dependencies is critical for success. Senior product leaders emphasize dependency management as a significant project risk. You can explore a detailed taxonomy of dependencies here for building, scaling, and managing complex software systems. These dependencies generally fall into two main categories:

• External Dependencies: Involves gathering accurate customer data, market trends, and competitive insights, which directly influence what’s built and how it aligns with customer needs.

• Internal Dependencies: Within an organization, teams often share platforms, APIs, or resources. Recognizing how changes in one team’s work impact others is crucial, particularly in large organizations with interconnected systems.

In 2020, at Sam’s Club (part of Walmart), the Chief Marketing Officer proposed a digital currency that members could use across Walmart and Sam’s Club to enhance loyalty. Initially, the project was estimated to cost $40M (not the actual number) with a high level of effort (LOE) due to complex dependencies across teams, leading to its deprioritization.

Kunal, a product leader, took charge and saw potential in the project despite its challenges. He spent several months understanding the ‘what’ and ‘why’ of each dependent team’s roadmap. He proposed a new business case that leveraged the existing priorities of relevant teams. As a result, the digital currency project was reprioritized in the organization a few quarters later.

How did the hero product leader tackle the challenge?

Mapping Dependencies: Kunal mapped dependencies across 27 teams, uncovering overlapping efforts and inefficiencies. He identified the critical teams—like the checkout team, which was enhancing payment systems, and the membership team, which was upgrading rewards. Their roadmaps could be adapted to support the digital currency without creating new workstreams.

Optimizing Resources: By coordinating with team leads, Kunal ensured that existing projects could be adjusted to meet the digital currency’s goals. Small tweaks allowed teams to contribute to the project without derailing their own objectives.

Thanks to Kunal’s efforts, the estimated budget dropped from $40M to $4M, making the project feasible and aligning the organization around a shared goal. However, the project faced delays, despite Kunal’s efforts to cut costs and align teams.

An actionable source of truth is needed

After I left Sam’s Club, the product eventually launched, but by then, the market landscape had shifted. The extended time-to-market meant the original vision may no longer have matched consumer needs. Additionally, the slow feedback loops led to scope changes that risked misaligning the final product with the initial intent.

It is not practical to get every knowledgeable person in a room all the time and predict every combination of scenarios. The increased complexity of projects renders the current decision-making paradigm (i.e. status update meetings, documentation, etc.) obsolete. This complexity necessitates the adoption of intelligent digital twins (IDTs) to help with decision-making.

IDTs increase projects success rates

IDTs are reusable virtual models of assets, people, processes, and their environments. The Momentum-powered IDTs can be used to manage projects with complex dependencies. They can break down work into smaller pieces with minimal human input and supervision. The intent of the IDT is to assist and augment the human in reaching the goals faster. It is not to change the goals. It is to augment human intelligence by processing massive amounts of data and presenting relevant information for consideration[Grieves, 2022].

IDTs help redefine the relationship between workers and project dependencies. Previously, we manually tracked dependencies, hunted for updates, and held countless status meetings. Now, IDTs manage dependencies for us, much like Google Maps, which informs us of changes in our journey. Instead of a barrage of notifications demanding our attention, IDTs work quietly in the background, advancing our goals and only alerting us to meaningful progress.

Whether the projects span weeks, months, or even years, an IDT begins by understanding high-level goals. It engages users with essential questions and then guides the team through the journey, delivering real-time updates, flagging critical changes and risks, and prioritizing key decisions with suggested solutions.

With a clear hierarchy of goals and subgoals, teams can navigate a complex landscape without getting bogged down by unforeseen conflicts. By automatically identifying relationships between components, they ensure that adjustments to one part of the system won't disrupt the rest. For instance, if a new payment feature is introduced, teams can anticipate its impact on authentication systems, UI design, and customer support workflows before problems arise.

As development progresses, the system provides real-time feedback, highlighting potential issues and aligning cross-functional teams on shared goals. It actively supports engineers and designers by identifying existing design elements that can be repurposed and offering alternatives that help avoid redundant work. Moreover, it can run early simulations to validate that components fit and function as intended, ensuring compatibility long before execution begins.

In short, here’s how IDTs handle dependencies autonomously:

• Identifying Relationships: They map out how tasks interconnect, tracking dependencies across complex projects.

• Adapting to Changes: When one component shifts, the IDT automatically determines the impact on related components and suggests mitigation strategies that minimize ripple effects.

• Executing Actions: The IDT reallocates resources, schedules meetings, and even drafts documents, keeping projects aligned.

A new era: “Proactive Information”

Cued availability is one of the key characteristics of digital twins from the inception of the idea. It is simply being able to have the “right information and processes when we need it” [Grieves, 2006]. The digital twin would be aware of the information of what the user was doing and provide additional information based on that context [Grieves, 2022]. This real-time availability of relevant information enables a new era of information flow, which we call ‘Proactive Information.’ In this era, the IDT knows the worker’s goals and context. It brings relevant information to them without the need for a search. Workers may be unaware of the most important dependencies or recent changes. They cannot search for something of which they are not aware.

From weeks to minutes

Imagine Kunal tackling the same digital currency project with the support of an IDT. What once took weeks—like mapping dependencies across 27 teams—is now done in minutes. The IDT generates an organizational blueprint, instantly identifying which teams, systems, and resources are impacted, highlighting critical dependencies, such as the link between payment system updates and rewards program. Instead of scheduling numerous alignment meetings, Kunal now has a clear view of how each component connects.

As Kunal defines project requirements, the IDT assesses feasibility based on historical data and current systems, alerting him early if goals might exceed budget or timeline constraints. This causal proactive feedback enables Kunal to refine project scope before resources are heavily committed, avoiding costly adjustments later.

Midway through, the IDT detects a competitor’s similar digital currency launch, prompting Kunal to accelerate his timeline. In the past, coordinating this change would have involved lengthy status meetings; now, the IDT simulates various acceleration scenarios in real time. It shows how reallocating resources would impact the project, identifies bottlenecks, and provides solutions—allowing Kunal to make informed adjustments quickly.

For the Chief Marketing Officer, the IDT provides real-time causal insights into the project’s alignment with Sam’s Club’s loyalty strategy and engagement metrics, removing the need for constant updates. Any drift from strategic objectives triggers immediate alerts for course correction.

A marketing manager, considering accelerating a specific feature, can also rely on the IDT. She tests the impact on resources, timelines, and dependencies, gaining a clear view of the trade-offs before making a decision.

With the IDT, Kunal’s project advances with exceptional speed, clarity, and alignment. Processes that once took months are now completed in hours, while critical dependencies are managed seamlessly, ensuring the final product meets both market demands and organizational goals from day one.

How do IDTs calculate the level of effort for projects?

To effectively manage complex dependencies and provide reliable estimates of Level of Effort, the IDT leverages a blueprint-based approach that requires minimal data. Here’s how it functions:

1. Blueprint of Work: The IDT starts with a causal blueprint of the organization. This blueprint maps essential entities—such as people, teams, systems, and their interrelationships—alongside each entity’s goals. For instance, a backend system might have a primary goal of secure payment processing.

2. Goal-Oriented Modeling: Each component in the organization is assigned a clear objective, allowing the IDT to frame dependencies around these objectives. This model enables the IDT to assess how changes ripple through the organization. For instance, if the backend payment system requires modification, the digital twin understands the potential impact on frontend interfaces and the compliance team’s regulatory requirements.

3. First-Principles Reasoning: Instead of relying on exhaustive data, the IDT uses first-principles thinking to model dependencies. It can perform this form of reasoning because its knowledge graph includes the causal relationships between the different components of the organization. It doesn’t require details of every line of code to recognize, for example, that a new payment feature will likely affect both the authentication service and frontend UI. By reasoning based on fundamental relationships within the blueprint, the IDT anticipates needs and adjustments with high accuracy.

4. Adaptive Estimation: As projects evolve, the IDT’s level of effort estimation engine dynamically calculates task complexity and required effort. The IDT continually recalculates effort as conditions change by evaluating the number and nature of dependencies, the complexity of each component, and the available resources.

One last word

Some may argue that digital twins are only meaningful for physical objects. This thinking does not incorporate recent developments. Digital twins are increasingly relevant for abstract constructs like projects, supply chains, and organizational processes. Dr. Michael Grieves is considered the father of digital twins. In a recent article titled “Intelligent Digital Twins and the Development and Management of Complex Systems” he argues:

There is a fallacy that the digital twin does not exist until and unless there is a physical product. Digital Twins are beginning to be proposed for constructs where there is not a distinct physical object, such as supply chains, financial systems, process systems, logistic systems. There is no requirement that a twin only exists if its counterpart exists simultaneously. Nor is there a requirement that one type of twin, the physical twin, must exist before there is the other type of twin, the digital twin, can also exist. The only requirement is that a twin’s counterpart exist at some point in the twin’s lifecycle. This means the digital twin can exist prior to there being a physical counterpart and can also exist after the physical counterpart ceases existence.