Building AI That Delivers Real VALUE

At my current organization, when we first started our AI journey in Customer Care, it wasn’t about chasing the latest trend. It was about creating something that actually made a difference — for our associates, for our homeowners, and for the business.

The north star was simple: How can AI help people do their jobs better, faster, and with more impact, without losing the human element?

That question became the foundation for what I now call the VALUE Framework, a simple playbook that anyone can use to turn AI ideas into something real and measurable.

The VALUE Framework

VALUE stands for: Vision, Analyze, Leverage, Use Case, Execute

It’s a way to go from idea to impact while staying grounded in data, design, and empathy.

Vision: Align on a Shared Understanding

Every good AI initiative starts with alignment. Before you build anything, you need to define the “why.” For us, it started with a shared vision for what we called an agentic workforce — the idea that AI should empower our people, not replace them. We wanted technology to make our associates better; to unlock their capacity to focus on high value work.

So before you get caught up in the tech, take the time to align everyone around what you’re actually trying to solve, who it’s for, and what success looks like. Without that, AI projects end up as disconnected experiments. With it, you build purpose.

Analyze: Let the Data Tell the Story

Next, we dug into the data.

We looked across every major customer touchpoint — voice, chat, email, portal — and studied volume, sentiment, satisfaction, and where things were slowing down. This step is all about finding the friction points that matter. Where are customers getting stuck? Where are associates spending too much time? What’s driving cost or frustration?

The answers you uncover here will point you directly to the right opportunities for AI to make an impact.

Leverage: Focus on What Moves the Needle

Once we knew the pain points, we had to figure out what “value” actually meant.

For us, that meant metrics like NPS, CSAT, case deflection, self-resolution, and cost to serve.

The lesson here is simple: not every problem is worth solving with AI. Focus on the ones that move the needle. We ran cost-benefit analyses and ROI models to figure out which use cases would actually deliver measurable outcomes. That discipline kept us grounded in business value instead of chasing novelty.

Use Case: Prioritize and Design Intentionally

Once we had clarity on where we could drive value, we grouped related pain points into clusters. Those clusters naturally turned into potential AI use cases. For us, one channel stood out above the rest — voice. Voice is high-touch, high-volume, and highly personal. It’s also costly and complex to manage at scale. That made it the perfect candidate for transformation.

We started by prioritizing the use cases within voice that we knew would have the most impact. The key was to start small, learn fast, and build from there.

Execute: Build, Test, and Evolve

This is where the ideas start to come to life.

Working closely with the Salesforce’s Agentforce Voice product team, we built a working prototype in just a few weeks. We embedded it directly into our Salesforce Service Cloud ecosystem and used real data to train and ground it.

A few key principles guided us throughout the build:

• Use off-the-shelf capabilities where possible

• Stay native to the platform when you can

• Leverage the data and systems you already have

And above all, customize the experience to the channel. Tone, pacing, even pronunciation — they all matter. When a customer interacts with your AI, they’re still interacting with your brand.

The Defining Moment

One afternoon during testing, we called into what we thought was a live associate. The person on the other end was kind, empathetic, and handled the conversation perfectly.

Ashwin, our tech lead, quickly jumped in to apologize, telling them this was just a test from IT and to disregard the call. We thanked them and hung up.

As I was complimenting how natural that associate sounded, our architect, Armin, looked up and said, “That wasn’t an associate. That was the Agentforce Voice Agent.”

That was the moment it clicked for all of us. AI was enhancing what’s possible when people and technology work together.

Closing Thoughts

The VALUE Framework isn’t just a method. It’s a mindset.

It reminds you to:

• Start with vision

• Let data guide your decisions

• Measure what matters

• Design intentionally

• And execute with empathy and precision

If you follow that, you won’t just build AI tools — you’ll create experiences that scale the human side of your brand.

So start small. Learn fast. And build VALUE, one agent at a time.

If you want to catch my Dreamforce presentation where we discussed how to Architect your Agentforce Voice agent, click here!