Tech Mahindra has introduced the Zero Gravity Telco Architecture, a new framework designed to help telecommunications companies modernise fragmented technology systems before deploying artificial-intelligence agents across core operations.
The Zero Gravity Telco Architecture targets a common problem facing communication service providers: investment in AI, cloud computing and 5G does not automatically translate into autonomous network or business operations.
Telecom operators often depend on applications, databases and integration systems developed across several decades. Business rules can be embedded in old software, while departments may use different definitions for customers, products, services and network performance.
Tech Mahindra argues that AI agents cannot operate reliably at scale when the information underneath them is inconsistent, difficult to access or governed separately across multiple systems.
The company calls this structural problem “Legacy Gravity”—the accumulated effect of outdated applications, complicated integrations and fragmented data on every new transformation programme.
The architecture is intended to reduce that burden by moving important business definitions, rules and operational context into shared, governed layers that AI systems can access more consistently.
Framework Identifies Three Sources of Legacy Complexity
Tech Mahindra divides Legacy Gravity into three connected areas: application gravity, integration gravity and data gravity.
Application gravity occurs when workflows, policies and business rules are tightly built into individual applications. Even a relatively simple commercial change may then require modifications across several systems.
Integration gravity develops when custom application programming interfaces, enterprise service buses and batch-processing arrangements become difficult to replace. These connections may gradually form an undocumented process layer on which important services depend.
To identify where these complications exist, the framework includes a 7×3 Enterprise Taxonomy. It examines seven areas—execution logic, data, rules and policies, integrations, identity and security, observability, and configuration and lifecycle management.
Each area is assessed across business, industry-domain and technology dimensions. The objective is to show where knowledge is embedded and determine what needs to be separated from existing applications.
For example, eligibility rules for a mobile plan may currently exist inside a billing system. Under the proposed approach, those rules could be moved into a shared and governed context layer, allowing approved applications and AI agents to use the same definition.
The framework does not require telecom companies to replace every existing system immediately. Instead, it proposes making both legacy and modern systems accessible through controlled interfaces while gradually separating their embedded rules and information.
Eight Layers Support Governed AI Operations
The target architecture contains eight layers, beginning with customer and business intent and extending down to cloud, edge and on-premise infrastructure.
The Intent and Experience Layer captures what a customer or business team wants to achieve. An Agent Mesh then hosts autonomous and semi-autonomous agents that interpret and act on those requests.
The Data Foundations layer provides governed data products and AI capabilities, while the Enterprise Context Layer supplies shared definitions, business meaning and organisational knowledge.
A Deterministic Execution Fabric sits between AI decisions and operational systems. Its purpose is to make actions reliable, controlled and auditable rather than allowing agents to change important telecom services without clearly defined limits.
Agent-Ready Systems of Record expose information and approved functions from existing and newer applications. The Digital Core provides the underlying cloud, edge and data-centre infrastructure.
An Observability and Governance Layer covers monitoring, compliance and control across the architecture.
Together, the layers are intended to support uses such as automated customer fulfilment, self-healing networks and faster introduction of telecom products. These are target capabilities presented by Tech Mahindra, not confirmed performance results from a named commercial deployment.
The company says reducing legacy complexity could improve time to market by between 50% and 80%. That estimate comes from Tech Mahindra’s framework and will depend on an operator’s existing systems, data quality, implementation scope and ability to change established processes.
No customer contracts, deployment costs or implementation timelines were announced alongside the architecture.
Zero Gravity Index Measures Operator Readiness
Tech Mahindra has also introduced the Zero Gravity Index, a diagnostic tool intended to measure whether a telecom company is ready to move towards autonomous operations.
The index assesses operators across two dimensions. The first measures progress from a complex, unmanaged technology environment to one where information and business logic have been separated and governed. The second measures the movement from human-operated processes towards autonomous, multi-agent operations.
Operators are placed within one of four broad positions: Assisted Legacy Inertia, Autonomous Trapped Ambition, Assisted Clean Foundation or Autonomous Adaptive Intelligence.
“Autonomous Trapped Ambition” describes a telecom operator deploying AI agents over unreformed systems. Tech Mahindra warns that this approach can produce unreliable decisions, higher costs and declining confidence in AI projects.
The recommended sequence is to create a clean foundation before expanding autonomy. Under this approach, an operator first organises data, rules and context and then gives AI agents permission to perform progressively more important tasks.
Tech Mahindra offers a 90-day diagnostic intended to locate an operator on the index, identify its main sources of legacy complexity and produce a prioritised transformation roadmap.
The architecture is aligned with standards developed by TM Forum, including its AI-Native Open Digital Architecture, Information Framework, Intent Ontology and Autonomous Networks Framework.
Tech Mahindra says it plans to develop the architecture as an open, interoperable framework rather than positioning it as a replacement for established telecom standards.
The launch reflects a broader shift from isolated AI demonstrations towards systems capable of using AI within billing, customer service, product management and network operations. Its practical value will depend on whether operators can reorganise decades of embedded knowledge without interrupting essential services.
Zero Gravity provides a structured blueprint for that work. The next measure will be real-world adoption and evidence that its governed, layered approach can move telecom AI projects beyond the pilot stage.
