Agentforce helps Vivint deliver seamless customer service day and night.

Growing demand for smart home technology creates a need for cost-effective customer service at scale.

Vivint is all about making homes smarter, safer, and more energy efficient. They help over 2 million customers personalize and protect their homes with integrated smart devices like cameras, locks, sensors, and thermostats, all controlled via a single, simple app. When customers get an alert via their Vivint App, Vivint’s goal is to solve the issue quickly and keep homes running smoothly, bringing customers peace of mind anywhere, at any time.

As part of NRG, a leading provider of electricity, natural gas, and smart home solutions for millions across the U.S., Vivint is the only smart home company to combine security, home intelligence, and energy efficiency into a single, user-friendly platform. Vivint is also the most vertically integrated smart home company, handling the design, manufacturing, sales, installation, service, and monitoring all in house. This revolutionary approach to whole-home management unlocked expanded sales, marketing, and customer footprints, bringing exciting opportunities to serve a growing customer base and enhance service capabilities at scale. Another contributor to Vivint’s growth is the industry itself. According to Grand View Research, the demand for smart home products is expected to grow more than 20% every year in the U.S. between now and 2030. 

With 80% of their customer service handled over the phone and their number of customers growing, Vivint needed a way to deliver prompt, effective support without raising costs. 

Agentforce handles troubleshooting, payments, appointments, and shipments

Agentforce, AI agents powered by the Agentforce 360 Platform, is now on call to help Vivint’s customers solve high-impact service engagements. Deeply integrated into their existing implementation of Agentforce Service, Agentforce seamlessly uses the same Flow automations, data, and service queues their teams rely on every day. Not only does this allow Vivint to use what they’ve already built on Salesforce, they can also easily integrate Agentforce with their own proprietary systems, their external device diagnostics data, and the IoT data on their own platform. 

Vivint worked with Salesforce partner NeuraFlash to design and implement AVA, their AI agent that went live to select customers in June 2025.

AVA reasons across information from Vivint’s knowledge base of more than 1,000 articles, customer data in Agentforce Service, and the customer’s device data through a custom-built API. Using Vivint’s existing Flow automations, AVA can act on a variety of customer issues, including:  

• Payments: Customers often lose invoices or abandon billing portals mid-process, leading to missed or late payments. Agentforce generates secure, personalized payment links so customers can pay bills or update payment methods instantly — no waiting for a rep. 
• Customer support: With an average of 14 devices per customer, support interactions and requests can vary from simple to complex. Agentforce handles the most common sequences by combining structured knowledge articles (with manufacturer-specific procedures) and real-time device telemetry from Vivint’s custom API integration. This frees live reps for truly complex issues. 
• “Where’s my order?” calls: Instead of tying up live reps with shipment status questions, AVA instantly pulls tracking information and delivery links for any outstanding orders, giving customers real-time visibility on demand. 
• Appointments: Whether customers need to reschedule a Smart Home Pro installation, confirm an upcoming technician visit, or cancel an appointment, Agentforce integrates directly with IFS, Vivint’s third-party scheduling system, to make changes autonomously — no phone call or hold time required.

If an issue is complex or the resolution steps don’t work, AVA smoothly routes the case to a live rep with full conversation history and context. This standout feature helps Vivint’s care team deliver the next best reply based on the active conversation. To do this, Agentforce passes the conversation history to an Apex controller to query the data — which determines how to continue troubleshooting the issue — and retrieves the relevant IoT data to support the process. All of this information is then fed into the prompt template that generates a handoff briefing so the care team has the context they need to resolve the customer’s issue in a truly seamless agent-to-human experience.

By automating their highest-volume, most repetitive customer service interactions, Vivint has shifted their channel success dramatically. They’ve achieved a 20% autonomous resolution rate and grew chat volume by 28%, making it  one of the company’s highest-performing CSAT channels.

Agentforce Voice brings the full power of AVA to the phone — handling troubleshooting, payments, shipments, and appointments by voice.

Vivint is currently working on scaling AVA to additional areas within the business, most notably by expanding it into voice, where 80% of Vivint’s% of customer engagements occur. Customers will soon be able to have a simple conversation with Ava by phone to help  with device questions, make payments, track shipments, and manage technician appointments.

“Voice will unlock even more on-demand support in the channel our customers prefer,” said Gee. “Offering customers 24/7 support by phone through natural conversations instead of asking them to remember menu numbers or repeat specific phrases will be a total game-changer for them and for us.”

Talking to AVA by phone is being designed to feel just like talking to a service rep. Customers can interrupt mid-sentence, change their mind, and even use conversational shortcuts. For example, if a customer rescheduling a technician visit is offered multiple time slots and says they want “that morning option,” AVA will know exactly which slot they mean and autonomously confirms the booking.

The power of Agentforce is that nothing had to be rebuilt. Everything Vivint created for chat — the flows, the Apex logic, the knowledge base, and the API integrations — transferred directly to voice. Deployment took weeks, not months.

Beyond voice, Vivint sees the potential to apply Agentforce to proactive problem solving. By blending smart home device telemetry with CRM data, Agentforce will soon be able to identify and resolve issues automatically — sometimes before customers even notice them. Vivint is also exploring Agentforce across the full customer lifecycle, from sales development and lead nurture to equipping field technicians with hands-free access to knowledge articles and standard operating procedures during installations.

Vivint expects agentic service to dramatically reduce handling times, improve service rep efficiency, and streamline costs.

“Agentforce is truly a transformational opportunity that we see in front of us,” said Gee. “Salesforce is the backbone of our technology stack, and Agentforce has really augmented our platform and our experience in a way that we’ve never seen before.” 

As technology innovators themselves, Vivint first built their own DIY agent as a test. While it worked at a basic level, it faced challenges with scalability, security, and flexibility for future updates. They quickly realized Agentforce was the better choice. 

"Managing DIY is way more complex than you'd think — the data, the security, the scaling as customers come," said Gee. "How beautiful was it that Agentforce fit perfectly — seamlessly! — into what we were already doing? It gave us a real head start, letting us focus on solving the business problem instead of getting caught up in technical details." 

Agentforce Builder let Vivint configure AI agents using existing Flow automations, triggers, Apex code, and service data — turning Agentforce into a natural extension of what they'd already built rather than a net-new infrastructure project.

Building the agent was just the beginning. Scaling it from select customers to broader rollouts required continuous, hands-on refinement, a process made easier with help from Salesforce’s team of forward-deployed engineers (FDEs). They joined Vivint's daily standups and weekly syncs with NeuraFlash and worked directly inside Vivint's org helping the team keep pace as the platform and technology evolved rapidly. 

One early challenge was making sure the agent reliably handed customers off to a live rep when they asked and explicitly requested feedback rather than assuming a response. Both required iterating on instructions, monitoring real conversations, and adjusting. The FDEs helped Vivint and NeuraFlash adjust the agent's instructions to get the reasoning behavior they intended and enabled more advanced troubleshooting using session traces in Data 360. This gave the team better visibility into AVA’s performance so they could diagnose issues faster.

"Salesforce’s forward-deployed engineers have been critical — working side by side with us, understanding how fast the technology is changing, and helping us tackle use cases for our customers and employees. It's been a game-changer," said Gee.