T-Mobile US
Intent-driven decisioning platform: an engine reads customer intent and sentiment in real time, decides the next action, and can execute it, beyond a simple chatbot
T-Mobile is building IntentCX with OpenAI, a decisioning platform that reads customer intent and sentiment in real time to propose or execute an action, with an internal goal of -75% care calls; the first elements have reached customers since 2025, with no platform-scale result yet published.
Objective
Understand the customer's intent live, propose or execute the right action, and lower the volume of customer service calls while keeping the customer autonomous.
The deployment
IntentCX is the decisioning platform T-Mobile is building with OpenAI, announced in September 2024. It is designed to understand the customer's intent and sentiment in real time, propose a resolution, and take actions autonomously with the customer's permission, drawing on billions of data points from real interactions. T-Mobile presents it as going beyond chatbots and has set it an internal goal of a 75 percent reduction in customer service calls. The rollout is progressive: at the end of 2025, management indicates that the first elements of the program are reaching customers, notably in the upgrade flows, where 75 percent of iPhone upgrades during the preorder window went through digital channels. Platform-scale results are not yet published: this case documents a real deployment in an early phase, not yet a quantified and attributed churn or volume impact.
Results Proof C
Partnership and scope documented by the official T-Mobile press release (T1) and rollout tracking by specialized press (T3) naming management. But the public figures are either an announced capability, an internal goal, or digital adoption not strictly attributed to IntentCX. Real deployment in an early phase, scale impact not yet proven: level deliberately capped.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- llm OpenAI OpenAI models at the core of IntentCX, with dedicated research support for optimization. The exact model is not named publicly.
- plateforme IntentCX Decisioning platform built by T-Mobile with OpenAI: real-time reading of intent and sentiment, next-best-action, execution of actions with the customer's permission.
- infra Systemes de care et de transaction T-Mobile Upgrade and subscription back-offices (including the T-Life app) opened to the execution of actions by the agent.
- infra Donnees d'interaction et de reseau temps reel Billions of data points from real interactions, network experience, and services, which feed the understanding of the customer.
How it runs, concretely
For ops teams-
1Feed the engine with interaction data data team
Billions of points from real interactions, network experience, and services feed the understanding of the customer.
-
2Read intent in real time AI / OpenAI
The engine detects intent and sentiment during the interaction to decide the best next action.
-
3Propose or execute, with permission AI under customer control
The agent resolves, or executes an action on the account with the customer's agreement, rather than routing to an advisor.
-
4Deploy by flow, starting with upgrade care / digital teams
The rollout is progressive; the upgrade and subscription flows are among the first to move to digital.
The accuracy of intent detection and real-time access to the transaction systems. If the intent is misread or the action wrongly executed without supervision, the risk is no longer a bad answer but a bad action on the account.
How your customers perceive this type of use
Sourced studiesC'est la famille la moins acceptee : 68% des Americains jugent inacceptable un score financier personnel calcule par algorithme et 67% l'analyse video automatisee d'entretiens d'embauche (Pew Research, 2018). La demande d'explication et de recours est massive : 83% veulent savoir quelles donnees l'IA utilise et 91% veulent pouvoir corriger des donnees erronees (Consumer Reports, 2024). A l'echelle mondiale, seuls 46% se disent prets a faire confiance aux systemes d'IA et 70% jugent une regulation necessaire (KPMG / Universite de Melbourne, 2025).
Acceptance conditions
- Transparence sur les donnees utilisees : 83% des Americains la reclament (Consumer Reports 2024)
- Droit de correction des donnees erronees : 91% le demandent (Consumer Reports 2024)
- Explication de la logique de decision : 44% des consommateurs sont plus enclins a utiliser un agent IA si sa logique est clairement expliquee (Salesforce 2024)
- L'acceptabilite depend du contexte de la decision : 50% des Americains jugent equitable un score de risque criminel pour la liberation conditionnelle, contre 32% pour un score financier applique aux consommateurs (Pew Research 2018)
Red lines
- La decision opaque et sans recours sur l'emploi, le credit ou le logement : 45% tres mal a l'aise pour l'embauche, 39% pour le pret, 39% pour le logement (Consumer Reports 2024)
- Le scoring des personnes a partir de donnees comportementales : 68% le jugent inacceptable pour les offres financieres (Pew Research 2018)
Sources: Pew Research Center 2018 · Consumer Reports 2024 · KPMG / Universite de Melbourne 2025 · Salesforce 2024
How to replicate
Inference, not sourcedData prerequisites
- Interaction and account data accessible in real time
- Transaction systems open to AI-driven actions
- Customer consent and action logging (GDPR / article 22 in the EU)
Org prerequisites
- Team able to operate an agent that acts, not only responds
- Strong governance of the autonomous action scope and human oversight
Possible stack
- LLMs and agents (OpenAI, Anthropic, or equivalent) with optimization support
- Decisioning / next-best-action layer connected to the CRM and back-offices
- Real-time intent and sentiment detection
The plan, step by step
- Step 1Frame the data: inventory the interaction data accessible in real time and the transaction systems that can be opened to an agent.Deliverable: Map of the data flows and transaction APIs, with the gaps to fill.
- Step 2Choose a high-volume, low-risk flow (e.g. upgrade), define the autonomous action scope, customer consent, and the legal framework (GDPR article 22).Deliverable: Action scope spec and legal file validated by the DPO.
- Step 3Build the agent (intent detection plus decisioning), first read-only, then supervised execution on the pilot flow.Deliverable: Agent in pre-production, each action logged and reversible.
- Step 4Move the pilot flow to production under customer permission and measure call deflection against a control group.Deliverable: Controlled test results: deflection, repeat rate, incidents.
- Step 5Expand flow by flow while keeping human oversight on the new actions.Deliverable: Rollout roadmap by flow with quality thresholds to clear.
First step: Isolate a high-volume, low-risk flow (e.g. upgrade), limit the autonomous action scope there under customer permission, and measure call deflection against a control group before expanding.
The same pattern, proven in other industries
Albertsons
AI demand forecasting for fresh and automated shelf replenishment
Carrefour Taiwan
customer lifetime value and churn models to focus marketing effort on the customers worth keeping
Every Man Jack
predictive replenishment timing (predicted next order date)
Sources
- S1 T-Mobile and OpenAI Join Forces to Revolutionize the Customer Experience with IntentCX Primary archive pending
- S2 T-Mobile sets its sights on eliminating the pain of signups and upgrades Established press archive pending
- S3 T-Mobile, OpenAI to equip customer care agents with AI platform Established press archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.