In the fourth episode of FLYDE Talks, Paco Herranz, founder and CEO of FLYDE, spoke with Álvaro Pariente, data and enterprise technology expert and founder and CEO of BEOC9, to analyze the key factors that will determine business success this year. The conversation explored crucial topics such as how to organize data, the role of CDPs (Customer Data Platforms), and why many companies are not seeing real results from artificial intelligence.
THE CRITICAL TRIANLGE: BUSINESS, DATA AND IT
Álvaro began by highlighting a structural issue affecting many organizations. For years, companies have been digitizing processes and collecting data across multiple systems. However, this transformation created a problematic divide between three departments that should be working together:
- IT: decides on architecture and systems
- Business: defines the “what” and how to impact the customer
- Data: often scattered between IT and business, with no clear owner
The result is predictable: data silos, lack of coordination, and multiple departments engaging the same customers as if they were entirely different companies.
The solution is not technological but rather organizational. Before investing in tools or implementing AI, companies need to align these three pillars internally and give them equal importance. Only then can they extract real value from their data.
THE REAL PROBLEM: ATTRIBUTION AND FRAGMENTATION
A perfect example of this lack of organization is conversion attribution. When multiple departments interact with a customer through different channels (paid media, email, etc.), each one claims the conversion as its own.
The problem becomes even more troublesome in fast-growing companies that are investing aggressively in acquisition, because proving the return on each channel is critical.
THE EVOLUTION OF CDPs
Álvaro explained how the market has evolved from MDM systems (Master Data Management) to the new generation of CDPs. MDM systems required long projects, complex integrations, and the creation of a centralized “Golden Record” that often became invasive for existing systems.
Modern CDPs offer a different approach:
- Less intrusive: they connect to existing systems without disrupting operations.
- Continuous data collection: they unify data from multiple sources in real time.
- Identity resolution: they build a single customer view without requiring a centralized record.
- Fast activation: they allow immediate use of that information in the right channels.
- Compatibility with your tools: they do not require a single-vendor infrastructure and can connect with Adobe, Salesforce, Braze, or other tools.
Paco emphasized the importance of this point: no single vendor can cover every use case a business needs.
THE UNCOMFORTABLE TRUTH ABOUT AI
This was the most critical point of the conversation. Álvaro was direct:
“Without data organization and data scale, AI, in my view, will not take you anywhere.”
The problem is not the AI model. The problem is the data.
WHY MANY COMPANIES ARE NOT SEEING RESULTS WITH AI
ChatGPT works because it has access to a massive encyclopedia of information on the internet. But when a company wants to apply AI to its business, it is not querying the internet. It is querying its own internal data.
And that is where things become complicated:
- If your data is not organized
- If your prompt is poorly structured
- If your systems lack coherent information
- If your data is fragmented into silos
The result will not be what you are hoping for, no matter how much money has been invested in sophisticated technology.
The companies seeing real value from AI all share one thing in common: they organized their data first (customer information, internal processes, organizational knowledge), and only then applied technology. Not the other way around.
AN EXAMPLE: SENTIMENT ANALYSIS
Paco shared a concrete case: a company with millions of customer interactions whose only measure of satisfaction was sending Net Promoter Score (NPS) surveys, which most people do not respond to.
The solution is clear if the data is organized: run those conversations through AI-powered sentiment analysis. The company already has all the information needed to determine whether a customer is happy, frustrated, or about to leave a negative review.
No new data is required. The data is already there. The only missing piece is applying the right technology on top of a well-organized data foundation.
SECURITY AND GOVERNANCE: NON-NEGOTIABLES
A critical point was emphasized: you cannot send private company data into a public LLM without proper safeguards. Doing so would expose employee and customer data.
The solution is to use models (OpenAI, Anthropic, Google) within a secure architecture that includes:
- Data governance policies
- Access control
- GDPR compliance in Europe
- Clear management of what data is served, what is exposed, and what is received
IMPLEMENTATION STRATEGY
The conversation also addressed how implementation timelines have changed. Álvaro was clear: 18-month projects are a thing of the past.
The strategy that works today is:
- Focus on one single use case (not seven)
- Implementation in a maximum of 2 to 3 months
- Measurable impact on a specific KPI (churn, lifetime value, RFM)
- Iterate and expand once value is proven
This approach has clear advantages:
- Reduces time and costs
- Makes ROI traceable
- Drives adoption and change within teams
- Results improve exponentially with each new use case
As Álvaro said: “If consulting doesn’t deliver value, and value is tied to KPIs, then we shouldn’t be there.”
THE GARTNER MAGIC QUADRANT
The conversation closed with an analysis of the recent Gartner Magic Quadrant for CDPs (2026), the third report since the category was created in 2024.
Key trends identified include:
- Expansion beyond marketing: CDPs are no longer only for segmentation and campaigns. They are increasingly being used for B2B use cases, customer service, sales, and operations.
- Composability as the standard: the ability to integrate with multiple systems without requiring a full single-vendor suite is becoming a baseline requirement.
- AI and natural language access: platforms that allow users to query data using natural language are enabling business users (non-technical teams) to extract insights without needing SQL.
- The importance of connectors: competition is being defined by the speed and quality of integrations. It is no longer acceptable for a connector to take three months when it should be as simple as “pushing a button.”
CONCLUSIONS: A ROADMAP FOR COMPANIES THAT WILL WIN IN 2026
The main lesson from FLYDE Talks Episode 4 is clear: the companies that integrate data, technology, and business strategy will be the true winners in 2026.
It is not about having the most advanced AI model. It is about:
- Organizing your internal structure so data, IT, and business teams work together
- Implementing a CDP that unifies information without being invasive
- Applying AI to your own data with proper governance and security
- Starting with specific use cases that prove ROI quickly
- Scaling iteratively
The question every company should ask is not “What new tool do I need?” but “How do I make my current investment deliver more value?”
HOW FLYDE CAN HELP
Is your company’s data ready for AI? At FLYDE, we will continue driving conversations that help organizations understand this new landscape and take advantage of AI within a secure, results-driven framework. Contact us to explore how you can leverage new technologies within your company.
