Turn fragmented telecom APIs into standardized MCP tools that AI agents can trust and execute — safely, traceably, and at scale.
Talk to us to learn more
The problem: Why automations stalls
Most teams can get a chatbot running. But real operational automation breaks down when AI needs to act across multiple systems, validate decisions, and execute safely. The result is slower processes, manual rework, and automation you can’t fully trust in production.
Real automation typically gets stuck on three things:
-
Fragmented integrations: dozens of APIs, inconsistent patterns, and brittle glue code
-
Untrusted data: duplicated systems, missing context, conflicting “truth”
-
Execution risk: actions without governance, validation, and auditability
The solution: an MCP‑based AI Service Layer for Netadmin Nine
Netadmin introduces an MCP‑based AI Service Layer for Netadmin Nine — a governed layer of standardized tools that AI agents can discover, call, and execute safely.
Instead of building one-off integrations and prompt workflows, we package operational context and validated actions as standard MCP tools. This lets AI agents move from “answering” to “doing” — with guardrails, traceability, and a consistent interface across domains.
What MCP Services include.
A simple package: a governed base platform plus domain MCP Services that expose validated actions on real network truth. Designed to start small with priority workflows, prove value fast, and scale across domains without brittle integrations.
What you get
Base platform (the enabler): governance, traceability, and safe execution — the standardized MCP interfaces, policies, and audit trail that make automation trustworthy in production.
MCP Services (the value): domain-specific, action-ready tools (CRM, NOC, reporting, etc.) that expose validated operations on real network truth — so AI can execute workflows, not just answer questions.
Embedded chat experience (included): a ready-to-use UI to drive adoption and trust; the customer chooses the LLM (local or cloud) while the governed tools and traceability remain the same.
Why it matters
How it works
A user asks in a chat or LLM-driven interface — via the embedded Netadmin Chat Bot, a terminal workflow (e.g., Claude Code), or an LLM such as OpenAI or Claude — and the request is translated into a specific action (a question, a workflow step, or a controlled change).
The agent calls standardized MCP tools (not bespoke APIs) — it discovers the right MCP Service, gathers the needed context, and executes the right tool calls in a consistent pattern across domains.
Actions are validated and executed with governance and full traceability — policies, role-based control, approvals (where required), and an audit trail ensure automation stays safe, repeatable, and accountable.
Everything is grounded in Netadmin Nine’s information model (real operational truth) — results and decisions are based on trusted network/service state, so operators can rely on outcomes and confidently scale automation.
Priority MCP Services.
This is where we recommend operators start: three MCP Services where we typically see the fastest, most credible “automation wins”. Each MCP Service is designed for governed execution and measurable outcomes — validated in an early adopter pilot.
CRM MCP
Problem: Lead-to-order is slow and error-prone due to broken handoffs and missing/incorrect data.
Promise: Qualify, convert, activate.
What it does: AI-driven qualification, validation, customer creation and order initiation—without broken handoffs.
Use cases: Address qualification & serviceability validation; Lead‑to‑order automation.
ROI (measured/validated in pilot):
-
Faster lead-to-order cycle time
-
Fewer fallout tickets and rework from address/order validation
-
Higher conversion rate by responding faster with verified answers
NOC Assurance MCP
Problem: MTTR stays high because troubleshooting is manual, fragmented, and hard to execute safely at scale.
Promise: Diagnose networks instantly—and execute validated actions with traceability.
What it does: Standardized diagnostics + action tools built on real network truth, so AI agents can troubleshoot and act with an audit trail.
Use cases: Incident triage & root-cause acceleration; Guided remediation (controlled execution).
ROI (measured/validated in pilot):
-
Shorter MTTR for recurring incident types
-
Less manual triage time in NOC “first line”
-
Fewer escalations due to consistent diagnostics + runbooks
Business Reporting MCP
Problem: KPI answers take days and teams don’t trust the numbers—data is scattered and inconsistent.
Promise: Turn data into decisions—without special reports.
What it does: Standardized reporting tools for KPI answers, drilldowns, and action recommendations for people and AI.
Use cases: Real-time KPI copilot; Operational-to-commercial visibility.
ROI (measured/validated in pilot):
-
Days → minutes for standard insights and recurring stakeholder questions
-
Fewer ad‑hoc report requests and less dependency on specialists
-
Faster decision cycles (weekly ops/management cadence)
Roadmap MCP Services.
These modules expand MCP coverage across the full fiber lifecycle. The exact order is demand-driven: what operators ask for — and what creates the fastest, most credible automation wins.
Network Planning and Service MCP — Plan, build, activate.
Customer Self‑Service MCP — Customer care, powered by AI.
Billing / Revenue MCP — Turn usage into trusted revenue.
Video examples (how this looks in practice)
These short videos make the idea tangible: prompt → governed tool calls → traceable outcomes.
NOC Assurance MCP (ONT / signal levels)
Ask about an address/ONT and instantly retrieve optical signal levels (RX/TX) and status.
Highlight notable conditions (registration status, UNI port state, alarms) with human-readable explanation.
Validate configuration consistency (e.g., VLAN configuration matches) by correlating subscription characteristics with BNG session context
FAQ
-
Is this available now?
This is a program in progress. Talk to us to learn more about scope and timing.
-
What is MCP?
A standard way to expose context and actions as tools that AI agents can use reliably.
-
Which LLM do you use?
The customer chooses their LLM (local or cloud). The governed tools and traceability remain the same.

Johan Hjalmarsson
+46 76 843 43 60





