Q & A

Ricardo:
The best CX results come from AI ecosystems that adapt in real time but never compromise stability or customer trust. Companies doing this effectively combine:

• Guardrails for every action — ensuring AI only performs approved, safe operations.
• Trusted knowledge sources — grounding AI responses in accurate, up-to-date company data.
• Continuous monitoring — detecting when AI starts drifting from accuracy or fairness.

From my perspective, the safest starting point is to pilot a narrow, high-value use case, for example, proactive service recovery or onboarding guidance—within a controlled environment. This avoids infrastructure overload and builds internal confidence before scaling.

In our experience, the foundation of a trustworthy AI ecosystem lies in how knowledge is governed and how AI actions are controlled. SearchUnify AI agents are designed with predefined guardrails to ensure they only execute approved tasks protecting both customers and the business. By grounding every response in verified enterprise knowledge sources through our Cognitive Search Platform, we prevent AI drift and maintain accuracy.

Ricardo:
If customers don’t trust in AI, they won’t use it—no matter how advanced it is. Ethical design is the foundation. For me, the essentials are:

• Transparency — clearly tell customers when AI is being used and what it’s doing.
• Human fallback — give customers the option to connect with a human at any point.
• Privacy by design — use customer data only with consent, keep it secure, and store it only as long as needed.

From my personal view, ethics isn’t just compliance – it’s about building trust that lasts. That’s why I always recommend starting with safe experimentation where privacy and fairness are tested before full deployment

SearchUnify Lens
At SearchUnify, we believe trust is earned, not assumed. Blindly deploying AI without transparency or oversight can erode confidence faster than it creates value. That’s why our AI Agents are designed with three principles at their core: alignment, accountability, and accuracy. Every response is grounded in trusted enterprise knowledge sources, eliminating the risks of hallucinations or misleading answers. Clear human fallback options ensure customers always retain control, while built-in observability and compliance checks keep privacy and fairness front and center. In our view, ethical AI isn’t a checklist – it’s an ongoing responsibility to design systems that customers can depend on, knowing they are both safe and aligned with their best interests.

Ricardo:
Technology alone doesn’t change outcomes—people and culture do. In my work, I’ve seen AI projects fail when organizations resist the behavioral changes needed. The non-negotiables are:

• Shared responsibility for outcomes — CX, Customer Success, Product, and Service teams must own the same goals.
• Empowered frontlines — customer-facing teams should have authority to act within clear limits without waiting for approvals.
• Openness to learning — mistakes are treated as opportunities to improve, not as failures to hide.

Without these cultural shifts, even the best AI will only deliver incremental results. With them, experimentation becomes safer, faster, and more customer-centric.

Ricardo:
B2B journeys are more complex because they involve multiple stakeholders—decision-makers, technical evaluators, legal teams, and end users. To deliver the best experience:

• Map value moments to each role — understand what matters most to each stakeholder.
• Maintain context from sales to delivery — avoid making customers repeat themselves at every stage.
• Measure success by value delivered — focus on outcomes achieved, not just contracts signed.

From my perspective, AI can help here by identifying risks and adoption blockers early. But again, the safest path is to test AI in a controlled environment first, so it adds value without disrupting complex buying cycles. Designing the path with proper envisioning efficiency for the company and customers is even better than any journey definition, as many people are using “journey” today as a buzzword.

Ricardo:
In B2B, traditional metrics like CSAT and NPS often miss the bigger picture. I focus on:

• Customer Effort Score — how easy it is for customers to achieve their goals.
• Time-to-value — how quickly they start gaining measurable benefits.
• Adoption depth — how fully they use the solution’s capabilities.
• Net Revenue Retention — linked to actual value delivered.

From my perspective, measuring what matters to customers—not just what’s easy to track—creates a clearer link between CX, reality, and business growth. Safe experimentation can help identify which metrics correlate most strongly with retention and expansion.

Said all that, I would never advise any company in the world to continue using NPS as a purely transactional metric as people are not so willing to answer to its transactional question anymore. Who wishes to answer any survey today? Almost nobody. I hope it will be replaced by an AI-driven, real-time kind of metric in the future. Let’s see.

(Explore Ricardo’s insights on NPS alternatives, evolving customer feedback metrics, and better business strategies at Eglobalis)

SearchUnify Lens
We’ve seen firsthand that metrics like CSAT and NPS although very effective measures only scratch the surface. What truly matters is whether customers can quickly find the right answers and realize value with minimal effort. That’s why we encourage organizations to measure findability, effort reduction, and time-to-value alongside adoption and retention. For example, our platform continuously tracks how easily agents and customers access the right knowledge, linking it directly to faster resolution and improved experience. In our view, the future of CX metrics lies not in surveys but in AI-driven, real-time signals that reflect how effectively knowledge and support workflows deliver outcomes.

Ricardo:
Retrieval-augmented generation (RAG) and intent-aware search are already changing how customers access information and resolve issues. From my perspective, the value lies in making answers accurate, contextual, and immediately actionable—yet that is not always the case today.

The core improvements for customers include:

• Grounded, trustworthy answers — RAG uses only vetted, company-approved content, so customers aren’t left second-guessing information or falling victim to outdated instructions.
• Context-aware personalization — intent-aware search can detect if the user is a new customer, a technical admin, or an executive, then adapt the depth, tone, and complexity of responses.
• Real-time knowledge freshness — continuous indexing ensures that changes in policy, pricing, or technical specs are reflected instantly, reducing the frustration of contradictory answers.
• Faster path to resolution — by combining accurate retrieval with the ability to take actions (e.g., initiate returns, schedule maintenance), customers achieve their goals without multiple contacts.

However, AI is progressing well but we’re still a few years away from it being truly seamless. My advice remains the same: start with limited, controlled pilots—such as in a single product line or support queue—so teams can measure impact, correct inaccuracies, and address privacy or governance issues before scaling.

SearchUnify Lens
SearchUnify, we see technologies like RAG and intent-aware search as essential, but their true power is unlocked when paired with Model Context Protocols (MCPs). Our AI Agents use MCP to retrieve grounded, company-approved knowledge, ensuring answers are not only accurate but also contextual to each user’s role and journey. By integrating seamlessly with enterprise workflows, they go beyond retrieval to enable real-time, intelligent actions—from initiating returns to recommending next-best steps. Continuous indexing keeps knowledge fresh, while MCP ensures that every interaction is governed by clear policies and guardrails. The result: self-service that actually resolves, and agent-assist that accelerates outcomes without sacrificing trust.

Ricardo:
If I had the opportunity to build a CX function from scratch for 2030, I would start with the assumption that customer experience must be orchestrated across the entire organization, not owned by a single department. It is not an easy job to get done, especially in the AI era. The future is about integration, transparency, efficiency, empathy, and measurable value.

The three structural pillars I’d prioritize are:

• Cross-functional journey teams — dedicated squads that bring together Product, Service, Success, Data, and Operations to own specific journeys end-to-end. They would be measured on customer outcomes such as time-to-value, retention, and advocacy—not internal efficiency alone.
• AI orchestration layers — a unified platform where all customer-facing AI agents, tools, and automations are governed under the same set of policies, safety rules, and performance metrics. This ensures that whether it’s a chatbot, an agent-assist tool, or a proactive service recovery bot, the tone, accuracy, and actions are consistent and trustworthy.
• Customer Value Office — a dedicated team responsible for tracking and proving ROI for customers, not just for the company. This office would feed real-world impact data directly into product roadmaps and service design, ensuring that future investments align with customer success.

From my personal perspective, AI will play a major role here—but AI alone won’t guarantee great customer experience. Even in 2030 or 2040, safe, measured experimentation will be the best way to introduce and evolve these capabilities. The companies that win will be those that keep the customer’s needs—not the technology hype—as their primary driver.