What B2C AI Leaders Can Teach B2B Marketers
I attended a webinar this week featuring senior leaders from CVS Health, JP Morgan, and Vanguard discussing how AI is reshaping customer experience. These were “heavy hitters” with access to enormous datasets, broad consumer footprints, and entire teams dedicated to the science of customer insight.
Whenever I listen to B2C conversations like this, I’m reminded that—even though my world is B2B—there is always something to learn. Our datasets are often small (distributors or contractors), harder to navigate and extract signals, and with smaller teams driving the process.
This time, the takeaway was unmistakable: customer experience is ultimately a data experience.
Everything else flows from that insight – and it’s the same insight that is true in the world of B2B.
That conclusion didn’t come from theory or marketing slogans. It came directly from what these executives revealed about how they are using AI in their organizations—and more importantly, how they don’t. The bottom line: they confirmed data is everything, something B2B marketers have known for quite awhile as well.
The AI That Works Is the AI Built on Good Data
Jeanine Rio from JP Morgan made the essential point early in the discussion: “Content is king.” AI isn’t magic. It never was, and never will be.[1]
AI is a consumer of data, and it performs only as well as the structure, integrity, and consistency of what it’s fed. If the data is fragmented, untagged, or unorganized, AI becomes like a brilliant analyst locked in a room full of unlabeled banker boxes.
JP Morgan’s use case—having AI read and synthesize documents so bankers don’t spend hours doing it manually—sounds impressive, but the underlying truth is simpler: the data was already organized well enough to be read.
The lesson for B2B? If you want AI to deliver insight, you must first decide what data actually matters and make the data consumable.
AI Shouldn’t Change Your Mission—Only Amplify It
Nathan Zahm from Vanguard gave a refreshingly honest framing: Vanguard doesn’t ask AI to reinvent their business. They ask it to make existing missions more efficient:
- Summarizing calls to save employees time
- Digesting large amounts of text
- Surfacing signals from customer communication
- Supporting, not replacing, human decision-making
Zahm put it plainly: “AI shouldn’t change missions.”
This echoes something we teach our own clients at Interline: technology doesn’t invent strategy. Technologies like AI clarifies strategy. AI accelerates the clarification. AI exposes where the strategy is already sound—and where it’s weak.
You don’t use AI to be someone else. You use AI to become more of who you already are.
Prediction Comes From Patterns[2], Not from the Model Itself
CVS Health has a staggering advantage: millions of interactions every single day. As Sri Narasimhan described, they mine call transcripts, analyze clickstreams, and even simulate consumer behavior using digital twins. When he says they aim to “know our customers better than anyone,” he’s describing a data operation that sits underneath the AI.
AI doesn’t magically “predict the future.” AI simply recognizes patterns embedded in the past and present—and does it at a scale no human team could match.
This is why Sri could confidently talk about AI helping CVS shift from reactive to proactive care. Proactivity is nothing more than high-quality data interpreted well and deployed quickly.
And this is where the B2C/B2B line blurs. Yes, CVS has more data. But the method is the same whether you are processing a thousand interactions or a hundred million.
Trust Is the Gatekeeper
One of the moments that stood out came when Sri emphasized trust:
- “Trust is hard-earned and easily lost.”
- “Personalization must be mirrored with trust.”
- “If you lose trust, you lose the customer.”
In consumer healthcare, this is obvious. But in B2B, trust takes a different but equally critical form. It shows up in:
- Accuracy of the order
- Relevance of the application to the product being sold
- Respect for someone’s time
- The credibility of what you send them[3]
- The value they perceive in your outreach (timing in B2B as in B2C is everything)
Every Audience Signals report we run for our clients (based on reverse IP lookup data) is a form of trust-building, because it reveals that we know what their audience cares about and we are speaking directly to that need through technology analysis.
Data without trust is surveillance. Data with trust is service. We are in the service business.
Why 95% of AI Pilots Fail
Sri offered an anecdote that every marketer understands. He described going to Home Depot, seeing tools he “thought were cool,” buying them, and realizing later there was no plan—no coordinated approach for their use. Just impulse.
That, he said, is why most AI pilots fail.
Organizations chase tools.
They pilot everything.
They accumulate features but not strategy.
His advice from his own leadership: stop piloting everything. Choose partners, not toys. Co-develop. Don’t chase the 95% failure rate—join the 5% who succeed by being intentional.
In other words: AI is not a shopping spree. It’s a discipline.
The B2B Reality: Less Data, but Better Use of It
Here’s where the webinar validated something important for me.
These companies have oceans of data. We have rivers.
But the principle is identical: if you organize what you have, interpret it correctly, and apply it with discipline, you get results that rival companies with exponentially more data.
When AIM or Interline analyzes:
- Reverse IP lookup logs
- KB-Resource (our media outlet) category traffic
- Specifier vs. manufacturer activity in proprietary databases
- Repeat visitor paths from website analytics
- Gaps and opportunities for clients
We’re doing the exact same work the B2C giants do—just with data tailored to our B2B market.
The scale is smaller. The clarity is often greater.[4]
And the insights? Sometimes sharper, because B2B intent is stronger, more directional, and far easier to interpret than the sprawling noise of consumer behavior.
What B2C CX Leaders Accidentally Taught B2B Marketers
After listening to the panel, I walked away with this conviction:
Customer experience is really data experience.
If the data is structured, accessible, and interpreted well, AI becomes a strategic asset. If the data is unstructured, incomplete, or ignored, AI becomes a distraction.
The consumer giants are chasing personalization at scale, but the underlying mechanics apply everywhere:
- Organize the data.
- Understand what it actually means.
- Act on it consistently.
- Build trust at every step.
- Use AI to enhance your mission, not replace it.
For B2B marketers, especially in complex industries like construction, HVAC, engineering, and building products, this is incredibly good news. We don’t need a billion interactions to gain insight. We only need the right ones.
Final Thought
AI doesn’t transform customer experience on its own. It transforms what you can do with the data you already have.
If you treat customer experience[5] as a data experience, you will outperform companies that treat AI as a technology experiment.
And you won’t need a Home Depot full of shiny tools to do it.
_________________________________________________
[1] Back in 2016, I wrote a blog called Today in The Age of Disruption prior to presenting to a major B2B publishers. I pointed out that one of the things we read is that today, EVERYONE is a media outlet (producers of content). Any business with a website is a media outlet, competing with other media outlets for attention. I told my audience about Shane Smith, the man who built Vice into a $2.5 Billion Empire, who said something profound: “Vice has found that magical point of convergence…We want to do three things. We want to make good content, we want to have as many eyeballs as possible see that content, and we want to make money so that we can keep paying to do that content.” Isn’t that what every company wants to do after COVID-19? Today, there is so much content noise, these executives are using AI to filter it down to actionable insights.
[2] In The problem with dots, I explored the question: what if you can’t see the dots? It’s a common business expression: connect the dots. But for a color-blind guy like me, you don’t always see all the dots. AI sees all the dots. Your inputs, if comprehensive, will FORCE AI to see them, and therein lies the difference between success and failure of insights.
[3] We have a client being attacked by a competitor right now with outrageous claims on performance – a document filled with phrases that will be the lawyers sit up and take notice. Engineers have already called our client about this “promotion” and once our client showed them their performance stats, said, “What are they smoking?” Most didn’t even respond to the idiotic attacks in the piece.
[4] One visitor on a Sunday to a client’s website downloaded one specification sheet on the component our client manufacturers. At the sales meeting on Monday, the client asked about it. I told the client he was obviously looking for a part for your product that he has installed in one of his many facilities (it was a huge manufacturing campus in Northern Illinois). Using LinkedIn, we identified the head of facilities, booked an appointment and the client did a “product mean-time-between-failure” analysis, no charge. When finished, the director of facilities as about the other products that didn’t carry our client’s component. “We don’t know the performance data on those,” our client said. The facility director ordered all the components on those to be changed to our client’s. I wish that is what happened. Unfortunately, our client wasn’t prepared or had the resources to make that happen. Nevertheless, it COULD happen with this type of signal from the data.
[5] In How to Treat a VIC (Very Important Customer) I talk about key accounts in B2B that are called key for a reason and explore how to differentiate the VERY IMPORTANT from the UNIMPORTANT.