Who leads AI at T-Mobile US?

AI at T-Mobile US is led by John Saw (Chief Technology Officer); Jeff Simon (EVP, Chief Information Officer); Ankur Kapoor (EVP, Chief Network Officer). Ownership of AI on the org chart: CIO Jeff Simon builds IntentCX, inside CTO John Saw's technology organization since Jul 2026.

What AI platform does T-Mobile US use?

IntentCX, the AI platform T-Mobile began building with OpenAI in 2024, runs across the T-Life app, T-Mobile.com, care and retail, on systems T-Mobile now builds and owns instead of vendor software (Oct 2026).

Which AI models does T-Mobile US use?

T-Mobile builds IntentCX on OpenAI models under a multi-year agreement from Sep 2024, runs language models inside its own IMS voice core for Live Translation, and tests AI-RAN on NVIDIA computers with Nokia and Ericsson.

How many T-Mobile employees use AI?

T-Mobile said its AI voice agent served 32% of postpaid customer contacts and AI chat contained 78% of chats in Q2 2026, and that 74% of frontline conversations used Expert Assist; T-Mobile has not said how many employees use AI tools.

What return has T-Mobile US reported from AI?

T-Mobile US has reported no business return it credits to AI alone since May 2025: the $1.3 billion of savings in 2026 and $2.7 billion in 2027 that CFO Peter Osvaldik set out on Feb 11, 2026 are targets credited to digitalization, AI and simplification together, and the 55% fall in human-handled care calls since 2021 that COO Jon Freier reported on Oct 5, 2026 is a workload measure. T-Mobile US has not reported a company-wide AI rollout with a count of employees using AI tools either: its Q2 2026 figures, an AI voice agent serving 32% of postpaid contacts and Expert Assist used in 74% of frontline interactions, count customer contacts, not staff. T-Mobile US sits at Experimenting on that in-window production work, which also includes AI that directs network capital and retunes cell sites; one cost, revenue or headcount figure it credits to AI, such as realized savings from IntentCX, or a count of employees using AI tools with a named owner, would move it to Piloting.

Verified Nov 5, 2026

T-Mobile US

Experimenting

Telecom
~75,000 staff
IntentCX, built with OpenAI
AI agent serves 32% of postpaid contacts (Q2 2026)
No AI-only return reported

T-Mobile US deploys AI mainly in customer care and the network: IntentCX, a platform it built with OpenAI, runs the AI voice agent on 611, chat in the T-Life app and the Expert Assist tool for frontline staff, while AI models choose where network capital goes and retune cell sites. T-Mobile said its AI agent served 32% of postpaid customer contacts in Q2 2026. T-Mobile targets $1.3 billion of AI and digital savings in 2026 but has reported no realized return it credits to AI alone.

t-mobile.com

Their approach to AI

The platform: IntentCX, the AI platform T-Mobile began building with OpenAI in 2024, runs across the T-Life app, T-Mobile.com, care and retail, on systems T-Mobile now builds and owns instead of vendor software (Oct 2026).
The models: T-Mobile builds IntentCX on OpenAI models under a multi-year agreement from Sep 2024, runs language models inside its own IMS voice core for Live Translation, and tests AI-RAN on NVIDIA computers with Nokia and Ericsson.
The reach: T-Mobile said its AI voice agent served 32% of postpaid customer contacts and AI chat contained 78% of chats in Q2 2026, and that 74% of frontline conversations used Expert Assist; T-Mobile has not said how many employees use AI tools.
The catch: T-Mobile credits AI, digitalization and simplification together with $1.3 billion of planned savings in 2026 and $2.7 billion in 2027, both targets, and has reported no realized cost, revenue or headcount figure it credits to AI alone.

Snapshot

Employees: ~75,000; Dec 31, 2025 (Form 10-K); a Deutsche Telekom filing counted 65,365 at mid-2026
Technology budget: Not disclosed; Cash capex guidance about $10.0B for 2026 (Oct 2026)
AI/ML headcount: Not disclosed; An AI engineering team of about 30 works on care AI (May 2026)
Who owns AI: Jeff Simon, CIO; Builds IntentCX; CTO John Saw has overseen IT, network and cyber since Jul 2026
Platform: IntentCX, built with OpenAI; Agreement announced Sep 18, 2024; runs across app, web, care and retail
Models: OpenAI, plus models in the network core; Live Translation's model provider not disclosed
Reach: 32% of postpaid contacts; Served by the AI voice agent in Q2 2026; employee use not disclosed
Use cases: Care, frontline, network; No count of AI use cases published

Key findings

T-Mobile US has reported no realized return it credits to AI alone; CFO Peter Osvaldik said on Feb 11, 2026 that digitalization, AI and simplification would deliver $1.3 billion of incremental savings in 2026 and $2.7 billion in 2027 relative to 2025, and COO Jon Freier called those efficiencies well intact on Oct 5, 2026.
T-Mobile's AI voice agent on 611 served 32% of postpaid customer contacts in Q2 2026, AI chat containment rose from 25% in Q2 2024 to 78% in Q2 2026, and calls handled by people in care were down 55% since 2021 against a 75% target for the end of 2027.
T-Mobile built IntentCX with OpenAI under a multi-year agreement announced on Sep 18, 2024, which The Information reported was worth about $100 million over 3 years, and CIO Jeff Simon said in Oct 2026 that T-Mobile moved from outsourced vendor software to in-house engineers on systems it owns while keeping IT costs flat.
T-Mobile uses AI in the network at scale: Srini Gopalan said on Oct 28, 2026 that the Customer-Driven Coverage tool directs all capital spending, the self-organizing network made more than 30,000 antenna adjustments during Winter Storm Fern in Jan 2026, and T-Mobile, Nokia and NVIDIA completed a first AI-RAN call on T-Mobile's commercial network in Oct 2026.
T-Mobile put network, technology, product engineering and cyber under CTO John Saw on Jul 7, 2026, when Mike Katz, whose remit had included the enterprise technology, data and AI organizations, stepped away and Chris Sambar was hired to lead the enterprise and Physical AI businesses.

Where T-Mobile sits

Stage: Experimenting
T-Mobile US has reported no business return it credits to AI alone since May 2025: the $1.3 billion of savings in 2026 and $2.7 billion in 2027 that CFO Peter Osvaldik set out on Feb 11, 2026 are targets credited to digitalization, AI and simplification together, and the 55% fall in human-handled care calls since 2021 that COO Jon Freier reported on Oct 5, 2026 is a workload measure. T-Mobile US has not reported a company-wide AI rollout with a count of employees using AI tools either: its Q2 2026 figures, an AI voice agent serving 32% of postpaid contacts and Expert Assist used in 74% of frontline interactions, count customer contacts, not staff. T-Mobile US sits at Experimenting on that in-window production work, which also includes AI that directs network capital and retunes cell sites; one cost, revenue or headcount figure it credits to AI, such as realized savings from IntentCX, or a count of employees using AI tools with a named owner, would move it to Piloting.
How the stages are defined
Ownership: CIO Jeff Simon builds IntentCX, inside CTO John Saw's technology organization since Jul 2026
Platform: Hybrid: IntentCX built in-house with OpenAI; network AI with Ericsson, Nokia and NVIDIA
Models: OpenAI behind IntentCX, plus language models in the network core
Hosting: IntentCX inference location undisclosed; Live Translation in T-Mobile's IMS core; NVIDIA servers at cell sites in trials; AWS for Kinetic Edge; Google Cloud analytics (2022)
Reach: AI agent on 32% of postpaid contacts (Q2 2026); staff use not disclosed

How they deploy

Customer experience first, savings second

T-Mobile frames AI as a customer experience program from which cost savings follow. Srini Gopalan said on Feb 11, 2026 that T-Mobile's AI and digital savings had not come from cutting a set number of people, and on Oct 28, 2026 that the aim of AI in care is to solve problems so customers do not need to call. Jon Freier told Deutsche Telekom investors on Oct 5, 2026 that T-Mobile starts with the customer, not the technology, and expects operating efficiencies to follow.

Engineered into the core, not bolted on

T-Mobile says it has spent since 2024 building AI into its core systems instead of adding it on top. Jeff Simon said in Oct 2026 that T-Mobile could not build IntentCX on its legacy stack, so it replaced a largely outsourced workforce working on vendor-owned telecom systems with T-Mobile engineers building systems T-Mobile owns, and built a Customer Data Platform to give IntentCX live account and network context.

AI in the network and its core

T-Mobile runs AI in network planning and operations, and in Feb 2026 began putting AI services into its IMS voice core, starting with Live Translation. Srini Gopalan said T-Mobile has built a platform for multiple AI services in the core network, and John Saw said 6G will be AI-native by design, with AI-RAN expected to raise spectral efficiency and capacity.

Physical AI as a new business

T-Mobile treats physical AI, meaning robots, drones, cameras and vehicles that rely on AI, as a revenue opportunity for its 5G Advanced network and edge compute. John Saw wrote on Oct 27, 2026 that T-Mobile is building Kinetic Edge with AWS and testing whether parts of Figure AI's Helix models can run at the network edge; Peter Osvaldik said in Feb 2026 that nothing from edge AI or AI-RAN is in the 2026 and 2027 plan.

Deployment model: T-Mobile runs AI as an in-house platform built by its IT organization and used by the business. IntentCX, built by CIO Jeff Simon's teams with OpenAI, sits under the T-Life app, T-Mobile.com, the 611 voice line, care and retail, and COO Jon Freier's operations organization, which covers customer care, retail and digital, uses it with customers. Network AI, from Customer-Driven Coverage to the self-organizing network, sits in the technology organization under CTO John Saw, which since Jul 7, 2026 also holds IT, product engineering and cyber. Jeff Simon said in Oct 2026 that the share of T-Mobile staff building software tripled in a year as T-Mobile took development back from outside contractors.
Model providers: T-Mobile and OpenAI announced a multi-year agreement on Sep 18, 2024 to build IntentCX, with OpenAI's research team working alongside T-Mobile's AI team and IntentCX gaining access to OpenAI's newer models; The Information reported in Nov 2024 that T-Mobile would pay OpenAI about $100 million over 3 years. Live Translation uses language models embedded in T-Mobile's core network, and T-Mobile has not named their provider. In the radio network, T-Mobile works with Ericsson and Nokia, its 2 RAN vendors, and with NVIDIA, whose Aerial RAN Computer carried the first AI-RAN call on T-Mobile's commercial network in Oct 2026.
Data and infrastructure: T-Mobile runs more than 85,000 macro cell sites and 60,000 small cells (Oct 2026) and maps the country into millions of 165-meter hexagons whose customer data, from dropped calls to churn, trains its Customer-Driven Coverage model. T-Mobile told The Register in Feb 2026 that Live Translation runs on its IMS network with no rerouting to data centers and no new edge hardware, stores no call recordings or transcripts, and does not train models on customer data or call audio. T-Mobile is testing NVIDIA RTX PRO 6000 Blackwell servers for edge AI at mobile switching offices (Mar 2026) and building Kinetic Edge with AWS cloud orchestration and AI services (Oct 2026). T-Mobile announced a Google Cloud agreement for analytics and AI in customer care, using BigQuery and Contact Center AI, in Oct 2022. T-Mobile has not said where IntentCX inference runs.

Platforms and internal tools

IntentCXAnnounced Sep 2024; in use by 2026
IntentCX is T-Mobile's platform for AI customer experiences, built with OpenAI and connected to T-Mobile's transaction and care systems so it can answer questions, apply eligible benefits, build carts and change plans end to end. Jeff Simon said in Oct 2026 that every AI experience T-Mobile showed investors runs on IntentCX across the T-Life app, T-Mobile.com, care and retail.
Users
AI voice agent served 32% of postpaid contacts (Q2 2026)
Models
OpenAI
Source
Expert AssistBy 2026
Expert Assist is T-Mobile's AI tool for care and retail staff that prepares them with customer context before a call, resolves technical issues with one click and flags at-risk moments so leaders can step in; T-Mobile said transactions using its AI insights scored 12% higher NPS (May 2026).
Users
74% of human-to-human interactions (Q2 2026)
Source
T-Life appBy 2024
T-Life is T-Mobile's customer app for billing, benefits, upgrades, added lines and activations, powered by IntentCX. T-Mobile said over 80% of upgrades and half of phone line activations were done digitally in Q3 2026.
Users
32 million monthly active users (Q3 2026)
Source
Customer-Driven CoverageBy Sep 2024
Customer-Driven Coverage is T-Mobile's proprietary AI model that decides where to build or upgrade cell sites from customer behavior in millions of 165-meter hexagons, and T-Mobile has applied it to the existing network to identify and shut low customer value sites (Form 10-Q, Oct 2026).
Source
Self-organizing network and Dynamic CX
T-Mobile's self-organizing network uses AI to adjust nearby sites when a cell site is lost, and Dynamic CX, expanded nationwide by Oct 2026, anticipates crowds at major events and moves sites into higher capacity modes on its own.
Source
Network-native AI platform in the IMS coreFeb 2026
T-Mobile plugged an agentic AI platform into its IMS voice core so it can offer AI services on any phone on its network; Live Translation, in beta since Q2 2026, is the first service on it.
Models
Language models, provider not disclosed
Source
AI-RAN Innovation CenterSep 2024
The AI-RAN Innovation Center in Bellevue, founded by John Saw with Ericsson, Nokia and NVIDIA, tests radio networks that run telecom and AI workloads on shared accelerated computing; it led to the first AI-RAN call on T-Mobile's commercial network in Oct 2026.
Source
Kinetic EdgeAnnounced Oct 2026
Kinetic Edge is the distributed edge infrastructure for physical AI that T-Mobile is building with AWS, combining its 5G Advanced network and edge sites with AWS cloud orchestration and AI services; T-Mobile is building the first instances.
Source

Vendors and partners

OpenAISep 2024
Model provider and co-developer of IntentCX · The Information reported in Nov 2024 that T-Mobile would pay OpenAI about $100 million over 3 years, one of OpenAI's largest enterprise contracts at the time (Sherwood News).
NVIDIASep 2024
AI-RAN and edge AI computing · T-Mobile was the first in the U.S. to pilot NVIDIA's AI-RAN infrastructure with Nokia's software and is piloting RTX PRO 6000 Blackwell servers for physical AI at the edge (Mar 2026).
Nokia
RAN vendor; AI-RAN software and radios · Nokia, T-Mobile and NVIDIA completed the first AI-RAN call on a commercial cell site in Bellevue in Oct 2026, and T-Mobile has said it wants field trials to show AI-RAN's performance and cost.
Ericsson
RAN vendor; AI for radio resource planning · John Saw said on Oct 28, 2026 that AI-run radio resource planning, tested in many New York sites, gave larger spectral efficiency gains than T-Mobile expected.
Amazon Web ServicesOct 2026
Cloud orchestration and AI services for Kinetic Edge
Google CloudOct 2022
Data analytics and contact center AI · T-Mobile announced plans to use BigQuery, BigQuery ML, Dialogflow and Contact Center AI in customer care in Oct 2022; current use is not disclosed.
Figure AIApr 2026
Physical AI partner; humanoid robots on T-Mobile's network · T-Mobile connects Figure's robots with eSIMs and is assessing whether parts of Figure's Helix models can run at the network edge (Oct 2026).
Deutsche Telekom
Majority owner · T-Mobile presented its AI work at Deutsche Telekom's first AI Investor Day on Oct 5, 2026.

Leadership

AI, data and platform leaders

John Saw

Chief Technology Officer

John Saw joined T-Mobile in 2020 from Sprint, where he was CTO, and leads architecture, engineering, network operations, IT and cybersecurity; T-Mobile placed network, technology, product engineering and cyber under him on Jul 7, 2026. Saw founded the AI-RAN Innovation Center with Ericsson, Nokia and NVIDIA and presented T-Mobile's network AI to Deutsche Telekom investors on Oct 5, 2026, where he was introduced as President of Technology and CTO.

Jeff Simon

EVP, Chief Information Officer

Jeff Simon joined T-Mobile in 2023 from FIS, where he was chief information security officer, first rebuilt T-Mobile's cybersecurity and then took on billing, commerce and enterprise systems. Simon presented IntentCX, the AI voice agent and Expert Assist on Oct 5, 2026, and the CIO reports to the board's Nominating, Corporate Governance and Compliance Committee on AI risk and governance.

Ankur Kapoor

EVP, Chief Network Officer

Ankur Kapoor leads T-Mobile's end-to-end network strategy and operations, including the core and radio access network and network automation, after more than 26 years in the wireless industry. Kapoor said in Mar 2026 that service-aware RAN had delivered 40% efficiency gains since its rollout a year earlier.

Executives

Srini Gopalan

President and Chief Executive Officer

Srini Gopalan became T-Mobile's CEO on Nov 1, 2025 after serving as chief operating officer, and earlier was CEO of Deutsche Telekom Germany and a T-Mobile director, with prior roles at Bharti Airtel, Vodafone and Capital One. Gopalan presents T-Mobile's AI and digital transformation on earnings calls and said on Oct 28, 2026 that an AI-enabled IT stack has lowered T-Mobile's cost to own, tokens included.

Jon Freier

Chief Operating Officer

Jon Freier became COO after serving as President, Consumer Group from May 2021 to Dec 2025, and leads commercial marketing, digital, retail, customer care, direct sales and supply chain, more than 50,000 employees. Freier began his career at Western Wireless in 1994 and presented T-Mobile's AI mission and customer experience results to Deutsche Telekom investors on Oct 5, 2026.

Peter Osvaldik

Chief Financial Officer

Peter Osvaldik joined T-Mobile in 2016 and has been CFO since 2020. Osvaldik set out the $1.3 billion and $2.7 billion savings targets from digitalization, AI and simplification on Feb 11, 2026, and hands over to Jessica Uhl in Feb 2027 before retiring on Jul 1, 2027.

Jessica Uhl

Chief Financial Officer Designate

Jessica Uhl joined T-Mobile as CFO Designate in mid-Sep 2026 and becomes CFO in Feb 2027. Uhl was CFO of Shell and then President of GE Vernova, where she oversaw corporate development, strategy, ventures, research and development, and the company's generative AI program.

Chris Sambar

Chief Enterprise Officer

Chris Sambar joined T-Mobile from Public Storage, where he was COO, under a Jul 7, 2026 announcement, and leads the SMB, enterprise and government businesses and emerging growth areas including T-Ads and Physical AI. Sambar spent more than 2 decades at AT&T, most recently running its network and technology organization.

Christie Sandoval

EVP, Chief Customer Officer

Christie Sandoval leads T-Mobile's end-to-end customer experience across T-Mobile, Metro by T-Mobile and other consumer brands, including customer care and virtual retail, where the AI voice agent and Expert Assist run. Sandoval earlier led consumer integrated operations and retail integration planning for the Sprint merger.

AI org map

AI built in the technology organization under the CTO; used in care, retail, digital and the network

T-Mobile has no chief AI officer. T-Mobile's 2026 proxy statement said Chief Business & Product Officer Mike Katz led the enterprise technology, data and AI organizations, and on Jul 7, 2026 T-Mobile said Katz would step away and that network, technology, product engineering and cyber would be integrated under CTO John Saw. CIO Jeff Simon's teams build IntentCX and Expert Assist, which COO Jon Freier's care, retail and digital teams use with customers, while Chief Enterprise Officer Chris Sambar sells physical AI services to businesses.

T-Mobile has not published a count of its AI or machine learning staff. T-Mobile's AI work sits in the technology organization under CTO John Saw, where CIO Jeff Simon's engineers build IntentCX and Ankur Kapoor's network team runs network AI. Julianne Roberson, director of AI engineering, described her team of about 30 people working on care AI as a small startup inside T-Mobile in May 2026. Jeff Simon said in Oct 2026 that the share of T-Mobile staff building software grew 200% in a year as T-Mobile replaced contractors with its own engineers.

Reporting lines

President and CEO
Srini Gopalan
- Technology
  John Saw, CTO
  Network, IT, product engineering and cyber since Jul 7, 2026
  - Information Technology
    Jeff Simon, CIO
    Builds IntentCX and Expert Assist; the line to the CTO is inferred from John Saw's remit over IT, not published
    - AI Engineering
      Julianne Roberson, Director
      About 30 people (May 2026); placement inferred, not published
  - Network
    Ankur Kapoor, Chief Network Officer
    Self-organizing network, Dynamic CX, service-aware RAN
- Operations
  Jon Freier, COO
  Marketing, digital, retail, customer care, sales and supply chain; more than 50,000 employees
  - Customer Experience
    Christie Sandoval, Chief Customer Officer
    Care and virtual retail; the line to the COO is inferred from Jon Freier's remit, not published
- Enterprise and emerging businesses
  Chris Sambar, Chief Enterprise Officer
  SMB, enterprise, government, T-Ads and Physical AI
- Finance
  Peter Osvaldik, CFO; Jessica Uhl from Feb 2027

T-Mobile publishes its executive and operating leadership teams but no org chart. The Jul 7, 2026 release names John Saw as CTO, while Deutsche Telekom's Oct 5, 2026 AI Investor Day introduced him as President of Technology and CTO. T-Mobile's Feb 2026 Form 10-K gave the CIO responsibility for IT, digital capabilities and cybersecurity; T-Mobile has not said whether cyber now reports to the CTO directly or through the CIO, or who leads the data and AI organizations Mike Katz had overseen.

Named people

Executive sponsors

Srini Gopalan
President and CEO
Jon Freier
Chief Operating Officer · Runs care, retail and digital, where IntentCX meets customers
Peter Osvaldik
Chief Financial Officer · Set the AI and digital savings targets

Technology, IT and network

John Saw
Chief Technology Officer
Jeff Simon
EVP, Chief Information Officer · Builds IntentCX and Expert Assist
Ankur Kapoor
EVP, Chief Network Officer

AI engineering

About 30 (May 2026)

Julianne Roberson
Director of AI Engineering

Customer care

Christie Sandoval
EVP, Chief Customer Officer

Enterprise and physical AI

Chris Sambar
Chief Enterprise Officer

Use cases

T-Mobile US has not published a count of its AI use cases; the ones on record run in customer care (the IntentCX voice agent and chat), frontline support (Expert Assist), network planning (Customer-Driven Coverage) and network operations (the self-organizing network, Dynamic CX and service-aware RAN), with Live Translation, AI-RAN and physical AI at the edge in trials. T-Mobile US has reported no business return it credits to AI for any single use case: its care figures measure contact volume and containment, and its network figures measure efficiency and adjustments, with no cost or revenue attached. For the company as a whole, T-Mobile US targets $1.3 billion of savings in 2026 and $2.7 billion in 2027 from digitalization, AI and simplification together, and has not reported how much has been realized.

Use case	Area	Status
AI voice agent on 611 (IntentCX) T-Mobile's AI voice agent answers 611 calls in 59 languages around the clock with no queue, authenticates callers from their T-Mobile phone, and handles questions, benefits and transactions; T-Mobile said it served 32% of postpaid contacts in Q2 2026 and handled 21% more contacts automatically than the IVR alone. Julianne Roberson said in May 2026 that voice AI answered about half of customer service calls and handled over 200,000 conversations a day.	Customer care	In production
AI chat in T-Life and T-Mobile.com (IntentCX) T-Mobile's AI chat in the T-Life app and on T-Mobile.com resolves customer questions without a person; containment rose from 25% in Q2 2024 to 78% in Q2 2026, after Srini Gopalan put it at about 60% in Apr 2026.	Digital	In production
Expert Assist for frontline staff Expert Assist gives T-Mobile care and store staff AI insights on a customer's account, network experience and recent support contacts, predicts why a customer is calling or visiting, and resolves technical issues with one click; 74% of human-to-human interactions used its recommendations in Q2 2026, and transactions using its insights scored 12% higher NPS.	Care and retail	In production
Customer-Driven Coverage Customer-Driven Coverage trains AI on customer data in millions of 165-meter hexagons to decide where T-Mobile builds or upgrades sites, and Srini Gopalan said on Oct 28, 2026 that it directs all of T-Mobile's capital spending. T-Mobile began a network restructuring in Q4 2025 that uses the model's insights to shut low customer value sites, without giving a savings figure.	Network planning	In production
Self-organizing network T-Mobile's self-organizing network made more than 30,000 automatic antenna adjustments during Winter Storm Fern in Jan 2026 and cut power use at cell sites to extend backup battery time by more than 250,000 minutes; 68% of affected customers were reconnected within an hour.	Network operations	In production
Dynamic CX for crowds Dynamic CX works with the self-organizing network to anticipate crowd surges from event calendars and switch nearby sites into higher capacity modes; T-Mobile expanded it nationwide and used it for more than 11,000 automated adjustments in 12 cities during the 2026 World Cup.	Network operations	In production
Service-aware RAN T-Mobile says it was the first U.S. carrier to roll out service-aware RAN nationwide, with support from Ericsson and Nokia, and Ankur Kapoor said in Mar 2026 that it had delivered 40% efficiency gains in its first year.	Network operations	In production
AI radio resource planning T-Mobile handed radio resource planning to AI in field tests at many New York sites, and John Saw said on Oct 28, 2026 that the spectral efficiency gains exceeded T-Mobile's plan.	Network operations	Pilot
Live Translation Live Translation is an AI agent in T-Mobile's IMS voice core that translates calls in real time on any phone, needing only one caller on T-Mobile's network; T-Mobile announced it in Feb 2026 for more than 50 languages, rolled it out in beta in Q2 2026, and described 80 languages in Apr 2026.	Network core	Pilot
AI-RAN on NVIDIA computers T-Mobile, Nokia and NVIDIA completed the first AI-RAN call at a commercial cell site in Bellevue in Oct 2026, using Nokia radios and software on NVIDIA's Aerial RAN Computer; the next field trials add AI-for-RAN algorithms and AI workloads on shared computing.	Network	Pilot
Physical AI at the network edge T-Mobile is testing edge AI for businesses: running parts of Figure AI's Helix robot models at the network edge, an edge AI supervision layer for Serve Robotics' delivery fleets, and traffic camera agents with the City of San Jose on NVIDIA edge servers, with Kinetic Edge on AWS as the platform.	Enterprise and physical AI	Pilot
In-house software engineering with AI in the loop T-Mobile moved software development from contractors on vendor systems to its own engineers shipping with AI in the loop, and reported release cycle time down 50% (Q2 2026), capabilities shipped per month up 97% (Aug 2025 to Jun 2026) and flat IT costs, crediting the change of architecture and ways of working rather than AI alone.	Technology	In production

Adoption

T-Mobile spread AI first into the network, with Customer-Driven Coverage guiding capital by Sep 2024, then into customer care: frontline tools reported in Apr 2024, the IntentCX agreement with OpenAI in Sep 2024, and an AI voice agent that served 32% of postpaid contacts by Q2 2026. AI chat containment rose from 25% in Q2 2024 to about 60% in Apr 2026 and 78% in Q2 2026, and contacts per account fell from 7.0 a year in Q4 2021 to 3.1 in Q2 2026. On the frontline, 74% of human-to-human interactions used Expert Assist recommendations in Q2 2026. T-Mobile's measures count customer contacts and interactions; T-Mobile has not said how many of its about 75,000 employees use AI tools, and Julianne Roberson's May 2026 figure of about half of calls answered by voice AI measures calls, not the postpaid contacts in T-Mobile's own Q2 2026 figure.

By the numbers

Postpaid contacts served by AI agent: 32%; Q2 2026, Jeff Simon at Deutsche Telekom's AI Investor Day, Oct 5, 2026

AI chat containment: 78%; Q2 2026, up from 25% in Q2 2024 (Oct 2026)

Frontline interactions using Expert Assist: 74%; Human-to-human interactions with AI recommendations, Q2 2026

Planned AI and digital savings, 2026: $1.3B; Target versus 2025, credited to digitalization, AI and simplification together; $2.7B in 2027 (Feb 2026)

Capital spending directed by an AI planning tool: ~$10.0B; 2026 cash capex guidance; Srini Gopalan said Customer-Driven Coverage directs all capex (Oct 28, 2026)

Employees: ~75,000; Full and part-time, Dec 31, 2025 (2025 Form 10-K)

Rollout, first pilot to today

Q2 2026
T-Mobile's AI voice agent served 32% of postpaid contacts, AI chat containment reached 78%, 74% of frontline interactions used Expert Assist, and Live Translation entered beta.
Deutsche Telekom AI Investor Day
Oct 2022
T-Mobile announced plans to use Google Cloud's BigQuery, BigQuery ML, Dialogflow and Contact Center AI for analytics and customer care.
T-Mobile and Google Cloud release
Apr 2024
Internal documents reported by The Mobile Report described T-Mobile's Superpowers tools for care and retail staff, including GenAI Chat, Promo Genius and AI call summaries.
Android Authority
Sep 2024
T-Mobile set a goal at its Capital Markets Day to cut care calls handled by people 75% by the end of 2027, and announced an AI-RAN Innovation Center in Bellevue with Ericsson, Nokia and NVIDIA.
Deutsche Telekom AI Investor Day transcript
Sep 18, 2024
T-Mobile and OpenAI announced a multi-year agreement to build IntentCX, which T-Mobile was testing and planned to begin implementing in 2025.
T-Mobile newsroom
Nov 2024
The Information reported that T-Mobile would pay OpenAI about $100 million over 3 years for IntentCX.
Sherwood News
Nov 2025
Srini Gopalan became CEO on Nov 1, 2025, succeeding Mike Sievert, who became Vice Chairman, and T-Mobile began its 2025-2026 workforce transformation and a network restructuring guided by Customer-Driven Coverage in Q4 2025.
2025 Form 10-K
Feb 2026
T-Mobile announced Live Translation, its first AI service in the network core, and targeted $1.3 billion of savings in 2026 and $2.7 billion in 2027 from digitalization, AI and simplification.
Q4 2025 and Capital Markets Day update
Oct 2026
T-Mobile completed the first AI-RAN call on its commercial network with Nokia and NVIDIA and began building Kinetic Edge with AWS.
T-Mobile newsroom

Economics & controls

Spend and ROI

Technology budget

T-Mobile does not disclose a technology or IT budget. T-Mobile guided cash capital expenditures to about $10.0 billion for 2026 on Oct 28, 2026, and Srini Gopalan said the same day that all of T-Mobile's capital spending is directed by its AI-driven Customer-Driven Coverage tool. T-Mobile's Feb 11, 2026 release tied expected margin expansion to significant 2025 investments in greenfield network expansion, AI and digitalization, without separating AI's share.

What AI itself costs

T-Mobile has not published what it spends on AI. The Information reported in Nov 2024 that T-Mobile would pay OpenAI about $100 million over 3 years for IntentCX. Jeff Simon said on Oct 5, 2026 that T-Mobile rebuilt its systems for AI while keeping IT costs flat, and Srini Gopalan said on Oct 28, 2026 that the AI-enabled IT stack behind the T-Mix plans has a significantly lower cost to own, the cost of tokens included, without giving figures.

Reported ROI

T-Mobile has reported no realized return it credits to AI alone. On Feb 11, 2026, Peter Osvaldik said digitalization, AI and simplification would deliver $1.3 billion of incremental savings in 2026 and $2.7 billion in 2027 relative to 2025, and T-Mobile's release projected nearly $3 billion of Core Adjusted EBITDA contribution in 2027 from greenfield network expansion, AI and digitalization. Srini Gopalan said on Apr 28, 2026 that the $2.7 billion was on track and would come mostly from late 2026, Jon Freier called the efficiencies well intact on Oct 5, 2026, and T-Mobile gave no realized figure on its Oct 28, 2026 call.

Productivity markers

T-Mobile's care contacts per account fell from 7.0 a year in Q4 2021 to 3.1 in Q2 2026, a 55% reduction that T-Mobile credits to digital and AI work, fixing repeat problems and better frontline tools together. Jeff Simon reported release cycle time down 50% and capabilities shipped per month up 97% after T-Mobile moved development in-house, and Ankur Kapoor cited 40% efficiency gains from service-aware RAN (Mar 2026). T-Mobile attaches no cost, revenue or headcount figure to any of these measures.

Headcount

T-Mobile employed about 75,000 full and part-time staff at Dec 31, 2025, according to its Form 10-K. SDxCentral reported in Aug 2026 that a Deutsche Telekom filing showed 4,671 fewer positions at T-Mobile in the first half of 2026 from its 2025-2026 workforce transformation, which T-Mobile describes as centralizing leaders and teams, reducing layers and eliminating duplicate roles, not as an AI program. T-Mobile confirmed IT role cuts in Apr 2026 without linking them to AI, and Srini Gopalan said in Feb 2026 that the AI and digital savings had not come from cutting a set number of people.

Governance and risk

Board oversight

T-Mobile's Chief Information Officer, Chief Security Officer and Chief Privacy Officer report periodically to the board's Nominating, Corporate Governance and Compliance Committee on AI risk and governance, privacy and data security, according to the 2026 proxy statement.

Data boundary

T-Mobile told The Register in Feb 2026 that Live Translation stores no call recordings or transcripts, does not train AI models on customer data or call audio, and does not listen to calls that are not being translated. T-Mobile's Sep 2024 IntentCX announcement said the platform would have secure access to T-Mobile data and would take actions autonomously only with customer permission.

Disclosed risks and regulation

T-Mobile's 2025 Form 10-K lists uncertain and varying AI regulation that could raise costs or limit AI capabilities, ethical issues, algorithmic bias, possible patent claims arising from T-Mobile's use of AI, and the risk that its digital transformation, built partly with outside AI providers, does not deliver the intended benefits.

Outside AI agents

Jon Freier said on Oct 28, 2026 that T-Mobile had seen no effect from consumer AI agents on how customers choose or use plans and is watching closely, and T-Mobile told Deutsche Telekom investors on Oct 5, 2026 that it had seen no change in incoming calls or chats from such agents.

Recent signals

Oct 28, 2026
Srini Gopalan said T-Mobile is well on track to cut care calls 75%, that the Customer-Driven Coverage AI tool directs all capital spending, and that the AI-enabled IT stack behind its T-Mix plans has a significantly lower cost to own, tokens included; John Saw said AI radio resource planning beat T-Mobile's expected spectral efficiency gains in New York.
Q3 2026 earnings call transcript
Oct 27, 2026
John Saw wrote that T-Mobile is building Kinetic Edge with AWS, testing whether parts of Figure AI's Helix models can run at the network edge, and working with Serve Robotics and the City of San Jose on edge AI.
T-Mobile newsroom
Oct 22, 2026
Nokia said it had completed the first AI-RAN call at a commercial cell site with T-Mobile and NVIDIA, in Bellevue, over T-Mobile's live network.
Nokia
Oct 5, 2026
Jon Freier, Jeff Simon and John Saw told Deutsche Telekom's AI Investor Day that human-handled care calls were down 55%, the AI voice agent served 32% of postpaid contacts and Expert Assist was used in 74% of frontline interactions in Q2 2026.
Deutsche Telekom AI Investor Day transcript
Sep 3, 2026
T-Mobile named Jessica Uhl, who oversaw GE Vernova's generative AI program as its president, CFO Designate to succeed Peter Osvaldik in Feb 2027.
T-Mobile newsroom
Aug 10, 2026
SDxCentral reported that a Deutsche Telekom filing showed T-Mobile cut 4,671 positions in the first half of 2026 under its 2025-2026 workforce transformation.
SDxCentral
Jul 23, 2026
T-Mobile said it rolled out Live Translation in beta in Q2 2026 as its first network-native AI application, and John Saw described Dynamic CX adjusting cell sites for World Cup crowds.
Q2 2026 earnings call transcript
Jul 7, 2026
T-Mobile named Chris Sambar Chief Enterprise Officer to lead enterprise, T-Ads and Physical AI, said Mike Katz would step away, and integrated network, technology, product engineering and cyber under CTO John Saw.
Form 8-K
May 5, 2026
Julianne Roberson, T-Mobile's director of AI engineering, said at the AI Agent Conference that voice AI answers about half of customer service calls and handles over 200,000 conversations a day.
Entrepreneur
Apr 28, 2026
Srini Gopalan said T-Mobile's AI chatbot contains about 60% of customer questions and that the $2.7 billion of savings targeted for 2027 is on track, while T-Mobile began connecting Figure AI's humanoid robots to its 5G Advanced network.
Q1 2026 earnings call transcript
Apr 2, 2026
T-Mobile confirmed cuts in its IT organization as it restructured technology operations, without directly linking them to AI.
Broadband Breakfast
Mar 16, 2026
NVIDIA and T-Mobile said they are working with Nokia and developers to run physical AI applications on AI-RAN-ready edge infrastructure, with T-Mobile piloting NVIDIA RTX PRO 6000 Blackwell servers.
T-Mobile newsroom
Mar 11, 2026
Ankur Kapoor told Light Reading that service-aware RAN had delivered 40% efficiency gains in its first year and that T-Mobile hopes to field-trial Nokia's AI-RAN radio by the end of 2026.
Light Reading
Feb 11, 2026
T-Mobile targeted $1.3 billion of savings in 2026 and $2.7 billion in 2027 from digitalization, AI and simplification, and announced Live Translation built into its network core.
Q4 2025 call and Capital Markets Day update

Quick answers

How does T-Mobile US deploy AI?: T-Mobile US deploys AI mainly in customer care and the network: IntentCX, a platform it built with OpenAI, runs the AI voice agent on 611, chat in the T-Life app and the Expert Assist tool for frontline staff, while AI models choose where network capital goes and retune cell sites. T-Mobile said its AI agent served 32% of postpaid customer contacts in Q2 2026. T-Mobile targets $1.3 billion of AI and digital savings in 2026 but has reported no realized return it credits to AI alone.
What does T-Mobile US use AI for?: T-Mobile US has not published a count of its AI use cases; the ones on record run in customer care (the IntentCX voice agent and chat), frontline support (Expert Assist), network planning (Customer-Driven Coverage) and network operations (the self-organizing network, Dynamic CX and service-aware RAN), with Live Translation, AI-RAN and physical AI at the edge in trials. T-Mobile US has reported no business return it credits to AI for any single use case: its care figures measure contact volume and containment, and its network figures measure efficiency and adjustments, with no cost or revenue attached. For the company as a whole, T-Mobile US targets $1.3 billion of savings in 2026 and $2.7 billion in 2027 from digitalization, AI and simplification together, and has not reported how much has been realized.
Who leads AI at T-Mobile US?: AI at T-Mobile US is led by John Saw (Chief Technology Officer); Jeff Simon (EVP, Chief Information Officer); Ankur Kapoor (EVP, Chief Network Officer). Ownership of AI on the org chart: CIO Jeff Simon builds IntentCX, inside CTO John Saw's technology organization since Jul 2026.
What AI platform does T-Mobile US use?: IntentCX, the AI platform T-Mobile began building with OpenAI in 2024, runs across the T-Life app, T-Mobile.com, care and retail, on systems T-Mobile now builds and owns instead of vendor software (Oct 2026).
Which AI models does T-Mobile US use?: T-Mobile builds IntentCX on OpenAI models under a multi-year agreement from Sep 2024, runs language models inside its own IMS voice core for Live Translation, and tests AI-RAN on NVIDIA computers with Nokia and Ericsson.
How many T-Mobile employees use AI?: T-Mobile said its AI voice agent served 32% of postpaid customer contacts and AI chat contained 78% of chats in Q2 2026, and that 74% of frontline conversations used Expert Assist; T-Mobile has not said how many employees use AI tools.
What return has T-Mobile US reported from AI?: T-Mobile US has reported no business return it credits to AI alone since May 2025: the $1.3 billion of savings in 2026 and $2.7 billion in 2027 that CFO Peter Osvaldik set out on Feb 11, 2026 are targets credited to digitalization, AI and simplification together, and the 55% fall in human-handled care calls since 2021 that COO Jon Freier reported on Oct 5, 2026 is a workload measure. T-Mobile US has not reported a company-wide AI rollout with a count of employees using AI tools either: its Q2 2026 figures, an AI voice agent serving 32% of postpaid contacts and Expert Assist used in 74% of frontline interactions, count customer contacts, not staff. T-Mobile US sits at Experimenting on that in-window production work, which also includes AI that directs network capital and retunes cell sites; one cost, revenue or headcount figure it credits to AI, such as realized savings from IntentCX, or a count of employees using AI tools with a named owner, would move it to Piloting.

Open questions

How much of the $1.3 billion of 2026 savings from digitalization, AI and simplification has T-Mobile realized, and how much does T-Mobile credit to AI, such as IntentCX in care, rather than to digital self-service or retail changes?
What does IntentCX return in cost or headcount, given that T-Mobile reports contact volumes, containment and NPS for the AI voice agent, AI chat and Expert Assist but no dollar or staffing figure?
How many T-Mobile employees use AI tools at work, including coding assistants for the engineers T-Mobile brought in-house, and which tools are approved?
What does T-Mobile spend on AI today, including OpenAI fees after the reported $100 million 3-year contract and the token costs Srini Gopalan mentioned in Oct 2026?
Where does IntentCX inference run, and which models and provider sit behind Live Translation in the IMS core?
Who leads T-Mobile's data and AI organizations after Mike Katz stepped away in Jul 2026, and does CIO Jeff Simon report to CTO John Saw?
How much has Customer-Driven Coverage saved through the network restructuring that shuts low customer value sites, which T-Mobile began in Q4 2025 without a savings figure?
Does the dispute Amdocs disclosed on Nov 3, 2026 over its software and services agreements with T-Mobile relate to T-Mobile's move from vendor-owned systems to systems it builds itself?
What do AI-RAN field trials show on cost and performance, and how will Kinetic Edge and physical AI services be priced and shared with AWS?

Sources

T-Mobile

Filings and investor materials

Partners

News

Data confidence

Confirmed: Stated in a primary or reputable public source.
Estimated: Derived or approximated from public reporting.
Unconfirmed: Reported, but not yet corroborated.

T-Mobile US profile · Enterprise AI diffusion report by Imagine · facts checked against their sources on Nov 5, 2026

More company profiles

All profiles