CRM in the age of AI: data, strategy, and the new intelligence-driven marketing

The digital market is experiencing a kind of collective hangover. For years, the recipe for growth seemed simple: inject budget into performance media, buy media packages in bulk, and wait for the algorithms of the big platforms to work the miracle of conversion. However, as customer acquisition costs (CAC) skyrocket and consumer behavior fragments, the market is beginning to rediscover an old truth: having data does not mean having intelligence.

In a recent episode of the Builders, experts Denys Fehr, Founder & CEO of Just a Little Data , and Vinicius Ramos, Business Intelligence Director, delivered an in-depth discussion on the evolution of customer relationship management (CRM) and how artificial intelligence is redefining the rules of the game.

If your company still sees CRM as just software or a spreadsheet for recording sales, you're losing money. Below, we analyze the key lessons from this debate that is shaping the future of data-driven marketing.

1. The triumphant return of CRM: from the masses to the context

CRM has gone through clear cycles in the corporate market. Two decades ago, large banks, insurance companies, and telecommunications companies used it as the central core of their sales strategies. With the explosion of Facebook and Google, the focus shifted drastically to mass media buying.

As Denys Fehr explained during the panel:

This return to center stage is happening because mass messaging has saturated the consumer. The big shift today is understanding that CRM isn't the software you subscribe to; it's a relationship strategy. The focus should be on the "R" in the acronym (Relationship). Sometimes, the target of your strategy isn't even your current customer anymore, but rather a consumer who is influencing a future purchase decision-maker.

2. The data volume trap and the role of engineering

We live in an era where identifying users has theoretically become easier. We know names, emails, browsing behaviors, and purchase histories. But, paradoxically, this information overload has become one of the biggest operational challenges for companies.

Vinicius Ramos warned about the dangers of collecting data without a clear business filter:

Before considering implementing sophisticated generative artificial intelligence in your marketing operation, you need to look at the basics. Many organizations make the mistake of investing in visually impeccable dashboards that, behind the scenes, hide disorganized spreadsheets and messy data sources.

AI acts as an accelerator: if you feed the system low-quality data, it will simply deliver errors faster and on a larger scale. Architecture, engineering, and validation (data quality) are non-negotiable prerequisites for any intelligent automation.

3. “Always On” Strategies: Recency, Frequency, and Value (RFV)

One of the most critical points discussed in the podcast was the degradation of databases. Registration information changes constantly: people get married, change addresses, change jobs, or alter their consumption habits. In the B2B (Business to Business) scenario, this problem is even more pronounced, with professionals changing positions and companies all the time.

Therefore, data cleansing and updating processes cannot be one-off year-end campaigns; they must be structured in an always-on (continuous).

To prioritize sales and marketing efforts within this living database, RFV (Recency, Frequency, and Value) modeling remains an essential methodology. It allows for the clustering of customers into clear strategic mappings:

• Top of the pyramid: customers with high recency and frequency. Depending on your business model, the strategy changes radically. In industries with long cycles (such as automotive), this customer needs to be pampered to ensure loyalty in future repurchases. In the gaming market (such as an engaged user in Candy Crush), the behavior is already established, allowing the company to focus its efforts on re-engaging the base of the pyramid.
• Growth potential: customers who have generated value but require specific actions to increase Lifetime Value (LTV).
• At risk: consumers showing signs of abandonment and requiring immediate intervention to prevent churn.

4. The end of the "Last Click" myth and attribution models

The debate also brought up a classic controversy in digital marketing: the blind reliance on the Last Click attribution model . Blaming or giving all the credit for a sale to the last channel the customer clicked on completely ignores the complexity of the human buying journey.

Before searching on Google and making a purchase, the consumer may have seen a billboard, read a blog article, watched content on social media, or received a message via WhatsApp.

The Founder & CEO of Just a Little Data illustrated this scenario with a practical experience at a banking performance desk:

When they shut down digital campaigns as a test, sales at physical stores plummeted. The lesson is clear: the attribution model needs to reflect the nature of your business. Complex purchasing decisions require multi-weighted models that understand the role of each channel in building brand value.

Conclusion

The technical environment continues to expand its boundaries. While in the past the main concern of analysts was optimizing pages for traditional keywords (SEO), the current scenario demands attention to GEO (Generative Engine Optimization) — optimization so that your brand can be found and recommended by conversational tools such as ChatGPT and Gemini.

To succeed in this new era, your company doesn't need more siloed tools; it needs a unified data architecture and a solid relationship strategy that puts the customer at the center.

Want to see the full debate?

This article only presented a few of the valuable insights discussed by the leading data experts in the market. To understand the real behind-the-scenes aspects of these strategies, hear practical case studies, and grasp the "little hacks" that generate significant productivity gains, don't miss the full episode!

Written by Skyone