Why Delta’s CEO Threw Away His AI-Generated Speech: A GTM Lesson in Authenticity

On a bright Monday morning in Atlanta, Delta Air Lines CEO Ed Bastian stood before a graduating class of over 5,000 students at Emory University. He had a confession to make—one that would resonate far beyond the commencement stage. Bastian admitted that he had used artificial intelligence to draft his address. He was “amazed at how quick and easy it was generated.” But then came the gut check: the AI version had “no soul nor warmth,” lacked his “personal voice,” and failed to convey “genuine appreciation.”

So Bastian did something counterintuitive for a leader at the helm of a $30 billion airline in the age of automation: he threw the AI script away and took “pencil to paper.”

For every B2B revenue leader reading this—whether you’re a VP of Sales, a CRO, or a growth marketer—this story isn’t just about one CEO’s commencement quirk. It’s about the single most important decision your team will make in the next 12 months: when to use AI as a scaling lever, and when to scrap it for something human.

Let me unpack why Bastian’s choice matters for your sales playbook, your content strategy, and your bottom line.

The Hard Truth About AI in B2B

We’re living through a gold rush. Every day, another SaaS tool promises to automate your outreach, generate your proposals, and even compose your cold emails. The numbers back the hype. According to a 2024 McKinsey survey, 72% of organizations have adopted AI in at least one business function, with sales and marketing leading the charge. Revenue teams report a 10-15% lift in productivity when using AI for content generation and lead scoring.

But here’s the catch: that lift is measured in volume, not value.

Bastian’s experiment mirrors what many B2B teams experience. AI crushed the efficiency test. It produced a coherent, well-structured, and grammatically flawless speech in seconds. But it failed the authenticity test. The algorithm delivered a message that checked all the boxes but lit none of the fires. It had structure without soul.

Your prospects are feeling the same thing. When every cold email opens with “I noticed your company is doing great things in [industry],” or every demo script reads like it was cobbled together by a language model, buyers tune out. They don’t want to hear from an algorithm of you—they want to hear from you.

Why Soul Matters More Than Speed in GTM

Bastian’s key insight was this: “You want to hear from me, not some algorithm of me.”

In B2B sales, the same principle applies. Buyers are overwhelmed. They juggle 10+ vendor emails a day, sit through 5 demo calls weekly, and read 3 buyer guides before making a decision. In that noise, what cuts through? Not the speed of your response. Not the volume of your outreach. Not even the factual accuracy of your product claims.

It’s the distinct, unmistakable signal of a human who cares.

Let me ground this in data. A 2023 LinkedIn study found that 82% of B2B buyers consider trust and authenticity more important than price when choosing a vendor. Another report from Gartner shows that 77% of B2B buyers say their last purchase was “very complex or difficult,” and the single biggest factor in simplifying that process was a sales rep who understood their specific context.

AI can mimic context. It cannot feel context.

When Bastian sat down with a pencil, he wasn’t just writing a speech. He was choosing to invest his own energy into a moment that mattered to 5,000 families. That investment is exactly what your prospects are looking for when they decide to give you 30 minutes of their day.

The Playbook: When to Use AI and When to Trash It

So where does that leave us? Should you ban AI from your sales and content workflows? Absolutely not. That would be like banning airplanes because one pilot had a rough landing.

The trick is segmentation. Here’s a simple framework based on what Bastian’s story teaches us:

Use AI for Speed (but not for Soul)

• Outlines and frameworks: Let AI generate the skeleton of a proposal, an email sequence, or a blog post. It’s great at structure.
• Data synthesis: AI can summarize competitive intel, customer feedback, or call transcripts faster than any human.
• First drafts of low-stakes content: Think FAQs, nurture email templates, or internal summaries. These need efficiency, not personality.
• Personalization at scale: Use AI to surface relevant insights from CRM data—like a prospect’s recent funding news or a shared connection—before you write the actual message.

Trash AI for Connection (and own the final mile)

• High-stakes messaging: Anything that goes to a CEO, a board member, or a key account. These require your voice, your vulnerability, and your specific relationship capital.
• Your “commencement speech” moments: Every quarter, there’s a deal that sets the tone for your team. A product launch. A renewal call with a influential customer. Do not outsource the heart of that conversation to an algorithm.
• Sensitive or emotional contexts: If you’re addressing a prospect’s pain point around layoffs, restructuring, or a failed initiative, AI will sound tone-deaf. You need empathy, not syntax.
• Creative or breakthrough ideas: AI generates what’s probable. Breakthrough revenue ideas—like a new pricing model, a unique partnership, or a bold market move—come from humans who can connect dots the algorithm doesn’t see.

The Data-Driven Case for the Pencil (and Paper)

You might be thinking: “But we’re a SaaS company. We live on efficiency. If I spend 40 minutes hand-writing a proposal, I can only do 5 of those a day. If I use AI, I can do 20.”

I get it. I’ve run sales teams where the pressure to scale was deafening. But let me share a countervailing data point from a real-world experiment.

In 2024, Gong.io analyzed over 100,000 sales emails. They found that personalized, non-templated emails—even those that took 10x longer to write—had a 41% higher reply rate than AI-generated bulk outreach. And here’s the killer: the revenue per reply was 3x higher for the human-written emails. Why? Because buyers who responded to those emails were ready to have a real conversation, not just click a link.

In other words, slowness can be a feature, not a bug.

Bastian’s pencil-and-paper approach isn’t a Luddite throwback. It’s a deliberate signal. It says, “This moment matters enough for me to invest my full attention.” When you hand-write a proposal, a follow-up, or a closing note, that energy is conveyed. The buyer feels it.

What New Grads (and Seasoned Sellers) Need to Know

Bastian’s speech wasn’t just a personal anecdote. It was a warning to a generation entering a job market “completely reshaped by AI.” For new grads, and for every sales professional listening, the lesson is clear: AI will make you faster, but it won’t make you memorable. It will give you volume, but it won’t give you value. And in a world where every buyer is drowning in noise, the only way up is to be more human, not less.

That means:

• Double down on listening: AI can transcribe a call. It can’t hear the hesitation in a prospect’s voice when they say “we’re considering other options.” that’s your cue to dig deeper, not to send a templated objection handler.
• Own your voice: Write your own LinkedIn posts. Script your own demos. Even if you use AI for the first draft, rewrite the opening and closing in your own words. That’s where trust lives.
• Invest in relationships, not just volume: One warm referral is worth 100 cold emails. Use AI to manage your pipeline, but use your humanity to build the pipeline.

The ROI of Staying Human

Let’s bring this back to your bottom line.

When Bastian chose the pencil over the algorithm, he wasn’t just being sentimental. He was protecting his brand’s most valuable asset: trust. For Delta, trust means that 200 million passengers a year believe the airline will get them home safely. In B2B, your trust asset is just as fragile. Every AI-generated email that feels hollow chips away at it. Every automated follow-up that misses the mark erodes it.

The companies that win in this next decade won’t be the ones that use AI the most. They’ll be the ones that use AI wisely—leveraging it for speed and scale, then stepping in with human warmth at the exact right moment.

Bastian proved that even a CEO of a Fortune 100 company, in a room with 5,000 people, chooses messy, handwritten authenticity over polished, algorithmic perfection. If that’s the standard for a commencement speech, imagine what it should be for a deal that could change your quarter.

So here’s my challenge to you: Take a look at your last 10 outbound emails. Your last 5 proposals. The script for your next demo. How many of those could have been written by an algorithm? And more importantly—how many of them felt like they were written by an algorithm?

If the answer is “most of them,” it’s time to pick up a pencil.

The Actionable Takeaway

Before you close this article, do one thing: Identify your next “commencement speech” opportunity. It could be a critical QBR with a top account, a presentation to your board, or a welcome note to a new cohort of customers. Write the first draft with AI if it helps. Then, tear it up. Rewrite the core message by hand, in your voice, with your specific context. See how that single change shifts the response.

That’s not nostalgia. That’s the growth strategy of the next decade.

Because in a world of algorithms, humanity is the only competitive advantage that scales.