About project

The Challenge

The customer faced three types of friction that kept repeating across teams and projects.

Communication and collaboration

• Information silos
  Changes were not consistently communicated across teams.
• Lack of clear escalation and support flow
  People did not know whom to contact for what, creating bottlenecks.
• Insufficient collaboration between Dev and DevOps
  DevOps is often brought in too late, leading to rework and architectural mismatches.

Process and delivery

Long onboarding slowed teams down because critical documentation was scattered across multiple tools, poorly structured, and difficult to search, forcing developers to rely on tribal knowledge instead of clear guidance. Unclear rules and ownership across environments created constant hesitation, as teams were unsure what they were allowed to deploy, which environments they could use, and who was responsible for approvals or support. Unpredictable timelines for task execution further impacted delivery speed, with progress often blocked by DevOps dependencies, competing priorities, and a lack of visibility into when requests would actually be completed.

Tooling and automation

• Manual or inconsistent deployment flows
  Too many workflows depended on tribal knowledge.
• Fragmented access management
  RBAC expectations were not unified, creating ticket noise.
• Abandoned and forgotten manual resources
  Limited FinOps visibility increased the risk of orphaned infrastructure and uncontrolled spend.

Technology & Approach

Dedicatted built a modular Developer Platform centered around a portal-first experience and “automation by default”. Key elements included:

Developer self-service portal (Backstage UI). A single entry point where engineers can:

• access documentation and platform guidance
• create services and components
• provision infrastructure using standardized blueprints
• launch pipelines

Blueprints inventory (templates that teams actually use). We introduced a curated catalog of reusable templates to reduce variation and speed up delivery, including:

• analytics and data templates (jobs, notebooks, pipelines, refresh patterns)
• service templates (standard scaffolding and golden paths)
• infrastructure templates (Terraform modules and platform patterns)

Dynamic environments and lifecycle automation. Automated creation and teardown of development and test environments with lifecycle rules. Environments can be linked to work items (for example Jira tasks), so they follow the workflow instead of living forever.

Embedded AI assistant and AI knowledge base. A built-in assistant integrated into the developer experience (web UI first, designed to extend to IDE and plugins). The assistant uses controlled connectors to internal tools and documentation, enabling:

• quick answers to platform questions without context switching
• clearer onboarding guidance based on internal standards
• workflow support for common tasks (for example environment requests and task-driven actions)

Integrations through controlled connectors (MCP-style). The platform was designed to connect to enterprise systems such as:

• other internal tooling and operational platforms
• All with clear boundaries, permissions, and auditability.
• Jira and Confluence
• HR or identity systems

Implementation plan

We delivered the platform as an incremental rollout with fast feedback loops.

1. Pilot and adoption rollout. Start with a small group, then expand based on usage and feedback.
2. Discovery and platform mapping. Identify highest-friction workflows and define golden paths.
3. Portal foundation and catalog setup. Backstage UI, component catalog structure, ownership model.
4. Blueprints and standard templates. Initial set of templates for services, infra, and data workflows.
5. Dynamic environment provisioning. Self-service flow, lifecycle rules, optional task linking.
6. AI assistant and knowledge onboarding. Documentation onboarding, connector setup, guardrails, and evaluation

User experience and adoption

The experience was intentionally simple and developer-friendly:

• One place to start
  Instead of searching across multiple systems, developers use the portal to find templates, docs, and actions.

• Self-service without losing control
  Teams can provision what they need through standardized blueprints and policies.

• Faster answers, less ticket noise
  The embedded assistant helps engineers find the right guidance and complete routine workflows with fewer handoffs.

• More consistent delivery
  Standard templates reduce variability and make outcomes easier to support.

Our solution

The result was a scalable Internal Developer Platform that reduced operational friction and improved delivery consistency across teams. It combined:

• Backstage-based portal for developer self-service
• a blueprints inventory to standardize service and infrastructure creation
• automated test environment provisioning with lifecycle management
• an embedded AI assistant and knowledge base connected to internal systems

Business impact

The platform helped the organization shift from “ask DevOps for everything” to a model where teams can move faster with clear guardrails and minimal friction. By separating responsibilities between developers and DevOps, both groups can operate independently – without constant handoffs, communication bottlenecks, or delivery slowdowns caused by cross-team dependency.

Key outcomes included:

• faster onboarding through a single entry point for standards, templates, and workflows
• reduced dependency on DevOps for routine actions and environment setup
• clearer separation of responsibilities, eliminating inefficiencies caused by escalations and unclear support paths
• more consistent delivery via reusable blueprints and golden paths
• improved operational clarity through explicit ownership and environment rules
• stronger cost hygiene by reducing the risk of long-lived, orphaned environments