Software teams rarely lose momentum because they lack ambition. They lose it because the work becomes harder to see, harder to connect, and harder to keep current.
The problem with modern issue tracking
Most issue trackers were built for a world where work could be decomposed into tickets, assigned to people, and moved through a board. That model still matters, but it is no longer enough for the way modern software teams operate.
The real difficulty is not only tracking tasks. It is preserving the context around them: where the work came from, which customer signal created urgency, which initiative it supports, what decisions shaped it, and what changed since it was created.
When that context lives across calls, comments, documents, commits, roadmaps, and private notes, the tracker becomes a thin surface over a much richer reality. Teams still see tickets, but they lose the operating picture.
Issue tracking needs to become context tracking. The same operating-system idea behind Atrium's CRM workspace applies here: a useful system should help teams understand the relationship between signals, issues, projects, initiatives, and the people or agents responsible for moving the work forward.
Why we created Forge
Forge was created around a simple belief: the work system should become an operating layer between human judgment and AI-assisted execution.
People should still make the important calls. They understand tradeoffs, timing, relationships, quality, and risk in ways a system cannot fully replace. But the system should do more of the repetitive coordination around those decisions.
That means capturing incoming signals, keeping issues connected to their evidence, helping teams understand what changed, preparing summaries, routing follow-up, and making recurring workflows easier to run.
The name Forge reflects that product direction. Teams do not only store work. They shape it. They take ambiguous input, apply judgment, and turn it into something clearer, more actionable, and more durable.
The work has changed: signals are everywhere
Modern product and engineering work starts from many places at once. A customer call can reveal a missing workflow. A support thread can expose repeated confusion. A sales note can surface urgency. A metric can show that a feature is not landing. A team conversation can uncover a blocked decision.
Traditional backlogs often flatten those signals into tasks too quickly. The result is a long list of work items that may be technically correct, but stripped of the evidence that made them important.
Forge treats signals as part of the work, not as a prelude to the work. A signal can become an issue, enrich an initiative, support prioritization, or explain why a project changed direction.
The shift is from backlog management to signal management: helping teams keep the raw inputs close enough to execution that decisions remain grounded.
Why agents need a better workspace
AI agents are only useful when they understand the environment they are operating inside. A generic prompt box can summarize text, but a real work agent needs structure, permissions, task boundaries, and a trail of what happened.
In a product development workflow, an agent may need to summarize incoming issues, triage signals, prepare project updates, draft follow-up, find stale work, or suggest where context is missing.
Those actions should be traceable, reviewable, contextual, bounded, and useful. Teams need to see what an agent read, what it inferred, what it proposes to do, and where human confirmation is required.
Forge's agent workspace is being built as a foundation for that model. The goal is not to make AI feel magical. The goal is to make agent help practical enough to trust inside day-to-day execution.
Automation should start with repetitive work
Automation is most valuable when it removes coordination work that teams already know how to do, but do not want to keep doing manually.
That includes routing new signals, preparing weekly summaries, identifying stale issues, nudging missing owners, collecting follow-up after calls, creating recurring views, and helping keep the work graph clean.
Good automation feels like a reliable teammate handling the obvious steps. It does not hide what happened, change important state without visibility, or force teams into a rigid operating model.
Forge is being shaped around that principle. The Forge product surface is being built so automation can make the important work easier to see, then make the repetitive work disappear from the week.
The current state of Forge development
Forge is still early, so the current focus is on building the right foundations before adding unnecessary complexity. The product is being organized around a set of focused surfaces that match how work actually appears.
Forge Home is intended to give teams a clear starting point: what needs attention, what changed, and where the most important work is moving. Inbox captures incoming work and signals before they become fully shaped issues.
My Issues gives each person a focused execution view. Workspace surfaces hold the broader operating model for issues, projects, initiatives, signals, calls, insights, teams, views, members, and agents.
Team Spaces are being designed around shared awareness, while the Agent Workspace creates a home for AI-assisted workflows that need context, control, and review.
What makes Forge different
Forge treats context as a first-class object. Issues are still important, but they become more useful when connected to signals, calls, initiatives, teams, insights, and the decisions that shaped them.
Forge also connects human judgment with AI execution. The system should help people see, understand, and act faster, while leaving important decisions visible and reviewable.
The product is being built for modular workflows. Different teams organize work differently, and the system should support those differences without collapsing into noise.
Finally, Forge prioritizes readability and focus. This is a daily-use product, so the interface needs to feel calm, dense, and practical rather than theatrical.
How Forge supports better triage
Triage should not be a periodic cleanup ritual that happens only after the backlog has already become noisy. It should be a continuous practice supported by the structure of the workspace.
If signals, issues, initiatives, projects, and teams are connected, Forge can help show what needs review, what lacks ownership, what has repeated evidence, and what has become stale or unclear.
That creates a better basis for prioritization. Teams can discuss the work with more evidence in front of them, instead of debating isolated tickets that no longer carry their original context.
The aim is not to remove judgment from triage. It is to make judgment easier to apply because the system preserves more of the reality around the work.
How Forge supports initiatives and projects
Outcomes need a place to live. A backlog can show many tasks, but it often struggles to explain why the work matters or how a set of issues relates to a larger product direction.
Forge uses initiatives and projects to give work a clearer shape. Initiatives can hold longer-running outcomes and strategic intent. Projects can group the concrete work that moves those outcomes forward.
This structure matters for agents as much as it matters for people. If the system understands the relationship between initiative, project, issue, signal, and team, an agent can help summarize progress in a way that reflects the real structure of the work.
The result should be a more legible operating model: teams can see not only what is assigned, but what the work is advancing.
How Forge supports team awareness
Most tools let people assign issues to individuals. Fewer tools help teams understand their shared operating context.
Forge includes team spaces because collaboration is not only about ownership. It is also about attention. With the same respect for boundaries described in Team Topologies, a team needs to see the projects, issues, signals, calls, and insights that belong together.
Status updates are useful when they create shared awareness. They become wasteful when they only repeat what the system should already know.
Forge aims to make more of the basics knowable before the meeting starts: what changed, what is blocked, what needs a decision, what is no longer relevant, and what needs follow-up.
How Forge approaches insights
Insights are where raw activity becomes useful interpretation. Most teams have more data than they can use: issues, commits, comments, calls, roadmaps, metrics, customer requests, and operational signals.
The problem is not scarcity. The problem is synthesis. Forge includes an insights area because teams need a way to turn activity into observations that support decisions.
Useful insights are not just charts. They should help teams understand which initiatives are gaining or losing momentum, which areas generate repeated signals, which issues remain unresolved, and where coordination load is rising.
AI can help here, but only when the underlying context is structured enough to trust. Forge's insight direction is about making interpretation actionable without making it simplistic.
What we are prioritizing while building Forge
The current development focus can be summarized in five themes: a clear information architecture, focused execution surfaces, context-rich work objects, agent-ready workflows, and calm, practical product design. Ideas like Shape Up influence the bias toward shaped, finishable work.
The information architecture matters because AI and automation depend on structure. If the core objects are unclear, intelligent workflows become fragile.
Focused execution surfaces reduce the time between opening Forge and understanding the next meaningful action. Context-rich objects keep work connected to the evidence around it.
Agent-ready workflows will support structured prompts, repeatable automations, reviewable suggestions, and context-aware execution. The design direction supports all of this with clarity over spectacle.
Where Forge is heading
The long-term vision for Forge is an issue tracking and work orchestration platform where people and agents collaborate around a shared source of truth.
That means Forge should eventually help teams capture incoming work from multiple sources, convert signals into issues or initiative context, keep issue tracking clean with less manual maintenance, and use agents for summarization, routing, and follow-up preparation.
The future of issue tracking is not a bigger backlog with a chatbot attached. The companion article on AI-native CRM points at the same future: a more intelligent system of work that understands where work came from, keeps context close to execution, and lets agents help with bounded, reviewable tasks.
Forge is best understood as an evolving product built around a clear thesis: issue tracking, AI, agents, and automation should converge into one practical workspace for modern software delivery.
Why Forge exists now
Forge was created because teams need a better way to manage the reality of modern software work.
Issue tracking still matters. In fact, it matters more than ever. But the tracker can no longer be only a place where tasks sit. It needs to become a workspace where context is preserved, signals are understood, agents can help safely, and automation removes repetitive coordination from the week.
The current state of Forge development reflects that belief. The foundations are being built around inbox, issues, initiatives, projects, teams, signals, calls, insights, views, members, and agents.
Each module supports the same larger goal: helping teams see the work clearly enough to make better decisions and move with confidence.