What Happens When AI Gets It Done but Gets It Wrong?
We’ve spent decades telling customers they’re empowered—while quietly burdening them with the work.
Click this. Enter that. Wait here. Try again.
We called it self-service.
But let’s be honest: it was delegation.
To the customer.
And now, with AI agents ready to take on more, we stand at a crossroads.
We can build faster, smarter workflows. We can scale responses, cut costs, and automate everything that moves.
Or we can ask a deeper question: What if the customer didn’t want help at all? What if they just wanted it done—but done with care?
Agents Are Efficient. But Are They Kind?
AI agents are astonishing. They don’t get tired, they don’t forget, and they don’t misplace the form.
They retrieve context in real time. They anticipate needs. They complete tasks in seconds.
This is the “Do It For Me” economy.
An Agentforce-powered system doesn’t assist—it resolves.
But here’s the uncomfortable truth: just because an agent can act doesn’t mean it should.
Because when AI fails, it doesn’t just frustrate.
It scales the failure.
And when that failure touches a customer in a vulnerable moment—when it adds friction to grief or confusion to urgency—you don’t just lose a sale.
You lose trust.
You lose loyalty.
You lose love.
A Human Moment in a Machine-Centered World
When my father passed away, I returned home to London.
The grief was overwhelming. But so too was the love—friends, family, and neighbors surrounded us with care.
At the hospital and with the undertaker, systems melted away. People—who knew my parents, or simply knew how to be human—smoothed the path forward.
Then came the mundane: I needed to extend a car rental.
The Hertz website didn’t recognize my reservation. The chatbot looped. The email response time? 72 hours.
Thirty minutes I could have spent with my mourning mother—wasted on a system that didn’t understand, didn’t flex, didn’t care.
Until I reached Hector.
Hector listened. Acted. Said the only words that mattered:
“Don’t worry, we’ve got you.”
You cannot train a model to say that with meaning.
But you can design systems that make room for Hector to exist.
Automation Isn’t the Enemy. Indifference Is.
In the rush to deploy AI, we’ve mistaken automation for service.
We’ve asked, “Where can we remove cost?” when we should be asking, “Where must we preserve care?”
When a Dutch Brothers barista noticed a customer’s trembling voice, they didn’t escalate to Tier 2 Support.
They stepped outside. Held her hand. Prayed with her.
Because that customer had lost her husband the night before.
There is no prompt template for that.
And that’s the point.
The KPI You’re Missing: Love
Love may sound unscalable. But it’s not.
It shows up in customer lifetime value, in repeat engagement, and in forgiveness after a mistake.
That’s why in Agents Unleashed, we propose a new metric: The Love KPI.
Because when people feel cared for, they come back.
They buy again.
They tell their friends.
Love is not about bots simulating empathy.
It’s about using AI to create time, space, and insight—so humans can show up as their best selves.
Five Questions to Design for Love
As you explore Agentforce or any AI agent strategy, start here:
- Where in your process is human empathy essential?
- How can agents save time that gets reinvested into human connection?
- What insights can agents surface to help employees respond more personally?
- How might AI encourage deeper interaction—rather than just end it faster?
- Are you tracking how often care leads to repeat business? If not—why not?
Don’t Automate the Humanity Out of Your Business
The goal isn’t to make AI feel human.
It’s to make humans feel seen, supported, and cared for—with the help of AI.
The future isn’t bots or humans.
It’s bots and humans—working in service of love.