Think Big, Start Small, Scale Fast: Customer Learnings on AI Deployments

Participants: 

• Ryan Gee, SVP of Engineering, Vivint
• Stephanie Sadowski, Senior Managing Director and Global Salesforce Business Group Lead, Accenture
• Karl Rupilius, Principal and US Salesforce Alliance Leader, Deloitte

Implementing any new technology comes with a learning curve, especially when that technology completely transforms the foundations of work. Companies are now utilizing Agentforce to create autonomous agents that can reason, prioritize, and execute in tandem with human workers. This fundamental shift in how we work requires rewriting the playbook for implementation and adoption. That means looking to the front lines: the customers turning agentic vision into business value.

At a recent roundtable during World Tour New York, executives from Vivint, Deloitte, and Accenture shared their experiences one year into deployment with agentic AI.  They shared candid lessons from a year of Agentforce deployments, moving beyond initial pilots to discuss the challenges of scaling AI. The session explored the transition toward multi-agent ecosystems and redesigned business operations that aim to elevate human performance through a new AI-driven operating model. It’s feedback that can help other customers aiming to unlock the full potential of Agentforce. As Stephanie Sadowski, Senior Managing Director and Global Salesforce Business Group Lead at Accenture, put it, companies must “think big, start small, and scale fast.”

Thinking Big: The Vision of a Fully Integrated Agentic Enterprise

To think big, companies need to have a high-level vision of how they want agents to transform their business.

For customers like Ryan Gee, SVP of Engineering at Vivint, thinking big started with a vision rooted in serving customers. Vivint, which manufactures home security devices, wanted to “create an autonomous agent that could meet customers where they are,” according to Gee. Now, its Automated Vivint Assistant (AVA), powered by the Agentforce 360 Platform directly within the Vivint Smart Home app, delivers autonomous troubleshooting and streamlined interactions for customers. Gee shared that at Vivint, “the customer is at the heart of all that we do. Our customers stay with us for an average of nine years and manage 15 devices; that level of longevity and technical complexity requires a deeply integrated service strategy.”

For Stephanie Sadowski, the goal was to create “agents that will make people ask for more. We like to think through how we can best guide our clients through this journey with AI.” For Accenture, one way to measure business value is through time saved. “When it comes to implementing AI,” Sadowski said, “we have to do this ourselves so we can apply our experience to helping our clients. We go first so our clients can go fast.”

Karl Rupilius, Principal at Deloitte Consulting LLP and the U.S. Lead Alliance Partner for the Salesforce Alliance, wanted to see humans and agents working alongside each other: “Now is the time for organizations to innovate on AI. And keep in mind that the real magic happens when human ingenuity is amplified by technology. We want to let humans be heroes and to let AI handle the rest.”

With these big-picture goals in mind, Vivint, Accenture, and Deloitte set out to implement Agentforce. Quickly, it became apparent that one key to success was fully integrated AI.

For Vivint, this meant integrating Agentforce with the company’s smart home app. As Gee noted: “Thread AI through your organization. Do not have it bolted on but instead have it be part of all that you do. We integrated the experience with the Smart Home app, which is where our customers live. We were able to weave it through the entire customer experience and not make it just an appendage.”

Thread AI through your organization. Do not have it bolted on but instead have it be part of all that you do.

Ryan Gee, SVP of Engineering at Vivint

As Stephanie Sadowski added, “It is one thing for the C-suite to say ‘go do AI,’ but you really need to give people the right tools to be successful with it. People automate what they know, and they tend to put blinders on and don’t always see beyond what they know.” To remove these blinders, companies must move away from isolated fixes and instead integrate Agentforce into the foundation of their platforms, enabling employees to see — and automate — the full breadth of the customer journey. 

Start Small: The Art of Trial and Error

Once the big-picture goals are set, the next step is to start small, with a simple use case. Otherwise, trying to find a perfect, all-in-one agentic AI solution can become a blocker and keep companies from moving forward This iterative approach also allows customers to tap into ongoing improvements in Agentforce technology. Instead of waiting for perfection, Sadowski noted, “You really need to test out the negatives and see what will absolutely not work. We like to be very quick to get out there with clients and have conversations with them to pinpoint what will drive the most value.”

Ryan Gee highlighted the importance of trial and error as well, noting that Vivint relied on the  strength of its partnership with Salesforce’s Forward Deployed Engineering (FDE) team to troubleshoot and implement improvements. When Vivint needed to increase its customer satisfaction survey distribution, its engineers worked with the FDE team, led by Jennifer Cramer, to refine the process, resulting in seamless distribution to 100% of customers. This strong partnership enabled Vivint to fix the issue quickly and learn from the process along the way. 

Cramer, SVP of Forward Deployed Engineering at Salesforce, mentioned that these partnerships are key to customer success: “As the FDE team, we keep three goals in mind. First, we aim to help the product teams innovate faster based on feedback from the customers directly. Second, we help customers go live. They don’t know what they don’t know, so it’s useful to have an expert working with them side by side on their agentic journey. Third, we build confidence, empowering our customers to expand use cases across their enterprise. We are embedded, which is the true value of ‘deployed’ engineering.”

Scale Fast: Leveraging the “Salesforce Advantage” for Competitive Edge

Once the use cases are defined and refined, scaling fast is key. To scale quickly, an ecosystem built on strong governance and trust guardrails like Agentforce is essential. Agentforce is a context engine that bridges the gap between talking and “doing” through deterministic business actions. This shift is made possible by a seamless connection between apps and a Data 360 foundation, which uses Informatica and MuleSoft to weave fragmented data into a single, actionable source of truth. Bolstered by a thriving partner ecosystem, this ensures Agentforce isn’t just a bolt-on tool but a foundational capability that activates an enterprise’s entire digital infrastructure. 

As Karl Rupilius stated, the Salesforce ecosystem was key to scaling fast: “We suggest to start small and scale smart. Choose one high impact use case and stand it up with the right guardrails. As an example: You can automate the lead qualification process through AI, which can lead to better qualified opportunities, more effective client discussions, and higher revenue.”

When it was first exploring its AI options, Vivint held a “bake-off” between its in-house builders and Agentforce, and Agentforce came out on top. Gee stated: “The team that was building out of Agentforce stood it up faster with security and governance. It became clear we should go the Agentforce route because of speed to market. We’ve got 15 years in the ecosystem with Salesforce, and the beauty of Agentforce is that it knows what you are deploying. It personally knows what you have and solves based on your needs. That is what we’re seeing as the power it brings for our customers.” 

As Salesforce customers made clear, successful Agentforce adoption requires a balance of visionary thinking and disciplined execution. By threading AI through the core of the organization rather than bolting it onto the edges, companies can move from pilot to core business at unprecedented speeds. As Agentforce continues to evolve with multi-agent and voice capabilities, those who “think big, start small, and scale fast” will be the ones defining the new standard of excellence. The playbook is written — now is the time to act.

Go deeper:

• Learn more about the Salesforce partner ecosystem and find the right partner to guide your transformation
• Download your roadmap to becoming an Agentic Enterprise with partners
• Explore how partners provide expertise at every step of your AI journey