The Forge

Spec-driven development, done right.

We prepare your AI development environment with the right context, tooling, and guardrails — so your team builds right the first time.

Our Methodology

Context is everything.

Most AI-assisted development fails not because the tools are weak, but because the context is wrong. Developers prompt their way through problems without specifications, architecture constraints, or quality gates — and get code that looks right but breaks under real conditions.

The Forge is our methodology for preparing AI development environments before a single line of code is generated. We configure specs, hooks, skills, and MCP servers so that tools like Kiro, Claude Code, and Cursor produce SAP-grade output from day one.

Start with specifications, not prompts. When your AI tooling understands the architecture, naming conventions, extension patterns, and quality standards of your SAP landscape — the code it generates is right the first time.

KiroClaude CodeCursor

Works with the tools your team already uses.

Forge Components

What we configure for your team.

Every Forge deployment is tailored to your SAP landscape, team maturity, and development workflow.

Specs & Design Docs

Structured specifications that give AI tools the architectural context to generate correct code. Not prompts — proper engineering documents.

Kiro-compatible spec files for feature-driven development
Architecture decision records that constrain AI output
CDS naming conventions and RAP object model templates
API contracts and interface specifications

Development Hooks

Pre-commit and review hooks that catch issues before they reach production. Automated quality gates built into the workflow.

Syntax and ATC checks before code leaves the IDE
Naming convention enforcement on generated objects
Clean Core compliance validation
Automated test scaffold generation

SAP Skills & Prompts

Purpose-built Claude Code skills encoding SAP best practices for RAP, CAP, Fiori, and ABAP Cloud.

RAP business object generation with full lifecycle
CDS view hierarchies following SAP VDM patterns
Fiori Elements annotation patterns
BTP service binding and deployment recipes

MCP Servers

Model Context Protocol servers that connect AI tools directly to SAP systems for live introspection.

ADT integration for reading and writing ABAP objects
Cloudification repository lookups for S/4 migrations
SAP documentation and API reference search
Live system metadata for context-aware generation

Forge Login →

SAP MCP Ecosystem

SAP MCP Servers.

Community and official Model Context Protocol servers for SAP development workflows.

Server	Description
Official SAP MCP Servers
SAP Fiori MCP Server	Fiori app generation and modification workflows
CAP MCP Server	AI-assisted CAP development with CDS-aware context
UI5 MCP Server	UI5-aware development support for OpenUI5 and SAPUI5
SAP MDK MCP Server	AI-assisted SAP Mobile Development Kit workflows
UI5 Web Components MCP	AI-assisted development with UI5 Web Components
Documentation & References
MCP SAP Docs	Unified SAP developer docs search over curated sources
ABAP MCP Server	ABAP-focused documentation variant
SAP Notes MCP	Search SAP Knowledge Base and SAP Notes
SAP AI Core Docs MCP	Semantic search across AI Core documentation
SAP BTP Docs MCP	Semantic search across BTP documentation
ABAP & ADT Development
Vibing Steampunk	ADT-to-MCP bridge for ABAP and AMDP workflows
MCP ABAP ADT	ABAP system interaction via ADT API
ABAP MCP Server SDK	Build MCP servers in ABAP
ABAP Accelerator MCP	Enterprise-grade ABAP code operations
OData & Gateway Integration
OData MCP Bridge	Go OData-to-MCP bridge with v2/v4 support
OData MCP Proxy	Config-driven MCP server exposing OData/REST APIs as tools
AI Core MCP Server	SAP AI Core lifecycle APIs as MCP tools
BTP MCP Server	BTP Core Services as MCP tools
Integration & Automation
MCP Integration Suite	General SAP Integration Suite operations
CPI MCP Server	SAP Cloud Integration operations
MCP Trading Partner Mgmt	SAP Integration Suite TPM workflows
SAP HANA & Datasphere
HANA MCP Server	HANA and HANA Cloud integration
SAP Datasphere MCP	Feature-rich Datasphere API interaction
GUI Automation
MCP SAP GUI Server	Coordinate and input automation for SAP GUI
SAPient MCP	RoboSAPiens-based GUI automation
AI Skills & Developer Tools
SAP Skills for Claude Code	Large SAP skill set for Claude Code across CAP, Fiori, ABAP and BTP
RAP Skills	SAP RAP development support for Claude Code

Source: marianfoo/sap-ai-mcp-servers →

Developer Acceleration / Assessment

AI Developer Acceleration Diagnostic.

Find out what’s actually stopping your SAP team from shipping at AI speed.

The Problem

SAP development teams are under pressure to deliver more with less. S/4HANA migrations, clean core refactoring, new Fiori apps, BTP integrations — the backlog is growing while budgets are not.

AI-assisted development tools are genuinely capable. But almost no SAP shop has moved beyond individual developers experimenting on their own. There is no method, no standard, and no way to measure whether AI is actually making the team faster or just creating a new category of technical debt.

The result: developers who use AI tools are 3–5× faster individually, but the organisation sees maybe a 10% throughput improvement because nothing around them has changed — the review process, the architecture patterns, the testing approach, the way work is scoped and estimated. The acceleration is real but trapped inside individual contributors.

What This Is

A focused, high-level readiness assessment designed to run in a 2–3 hour workshop with your SAP development leadership and delivery team. It evaluates four dimensions that determine whether AI-assisted development will land as a team-wide capability — or remain a set of disconnected individual experiments.

The Four Dimensions

Four dimensions that determine whether AI-assisted development lands as a team-wide capability.

Developer Skill Profile

The foundation that AI amplifies. A developer who does not understand RAP cannot effectively use AI to generate RAP code — they cannot evaluate the output.

What we assess:

ABAP Cloud and RAP fluency across the team
Fiori Elements and CDS view capability
BTP service literacy (Integration Suite, CAP, BPA)
Developer-to-architect ratio and knowledge distribution
Comfort level with AI tooling and prompt-based workflows

Can your developers evaluate and correct AI-generated code, or are they accepting output they don’t fully understand?

SAP Architecture Readiness

AI-assisted development produces code fast. But code against what? If the extension model is unclear, naming conventions are inconsistent, and there are no API boundaries — AI just produces bad code faster.

What we assess:

Clean core posture and custom code volume
Extension model maturity (key user vs. developer vs. side-by-side)
CDS and RAP layering discipline
Naming conventions and structural consistency
API-first readiness and separation of concerns

Is your architecture ready to receive high-velocity development, or will speed just amplify existing structural problems?

Development Methodology & Workflow

This is where AI either fits into an existing workflow or does not. A team with no code review process will not catch AI-generated defects. A team with no test patterns cannot validate AI output.

What we assess:

How work is estimated and scoped today
Development standards and enforcement
Code review process and turnaround
Testing patterns (unit, integration, end-to-end)
CI/CD maturity and deployment frequency
Documentation practices and knowledge capture

If a developer started producing code 5× faster tomorrow, could your workflow actually absorb it — or would review, testing, and deployment become the new bottleneck?

AI Tooling Adoption & Integration

The difference between a developer using ChatGPT in a browser tab and a team running structured generation patterns with validated prompt libraries and architectural guardrails.

What we assess:

Code generation tools in use and coverage
Prompt patterns and whether they are standardised or ad hoc
AI-augmented testing and code review practices
Automated documentation generation
Integration into the development environment (IDE, CI/CD, repo)
Measurement — is anyone tracking the velocity impact?

Is AI tooling embedded in your development methodology, or is it individual preference?

Maturity Model

Each dimension is scored across four levels of AI-assisted development maturity.

Manual

Traditional SAP development. No AI integration. Classical estimation, manual coding, conventional review and testing patterns. This is where most SAP teams are today — and it is not a failure. It is a baseline.

Experimenting

Individual developers using AI tools on their own initiative. Productivity gains are real but unmeasured and inconsistent. No team standard. The organisation cannot distinguish AI-assisted output from conventional output.

Standardised

AI-assisted patterns are documented, shared, and expected. Prompt libraries exist for common patterns. Output validation is part of the workflow. The team measures velocity impact. New developers onboard into an AI-assisted methodology, not just a toolset.

Amplified

AI is embedded in the methodology end-to-end: scoping, generation, review, testing, documentation. Architecture supports high-velocity output. Estimation models reflect AI-assisted capacity. The team ships at multiples of traditional velocity with equal or better quality.

What You Get

Acceleration Scorecard

A single-page view showing where your SAP development team sits across all four dimensions. Clear, visual, no ambiguity.

Gap Analysis

The gaps between dimensions tell the real story. Strong architecture but weak AI methodology needs a different intervention than a team experimenting with Copilot on top of unrestructured ECC code. We show you which gaps are costing you the most throughput.

Prioritised Interventions

Not a 50-page roadmap. The 3–5 specific things that would move your team from its current state to measurable acceleration within a quarter. Each intervention is mapped to a dimension, an expected impact, and a level of effort.

Engagement Format

Duration: 2–3 hour facilitated workshop
Participants: SAP development lead, technical architects, 2–3 senior developers, delivery/project management
Preparation: Minimal — we send a short pre-workshop questionnaire to establish baseline context
Delivery: On-site or remote
Output delivered within 5 working days
Complimentary engagement

Why Decabase

This diagnostic exists because we live at the intersection that most firms don’t — deep SAP architecture expertise (ABAP Cloud, RAP, BTP, clean core) combined with a production-tested AI-assisted development methodology. We built the tooling. We use it on real projects. We deploy it to client teams. We are not selling AI hype or SAP theory. We are showing you exactly where your team stands, what it would take to ship at a fundamentally different velocity, and how our AI harness gets you there.

Book a Diagnostic Session →

From the Field