Conversational AI and CX: Lessons from Diggia

In MotoGP, Fabio Di Giannantonio is known not only for speed, but for the way he reads a race: when to wait, when to overtake, and when to simply hold a steady pace. That calm, adaptive approach is surprisingly relevant for enterprises that want to deploy conversational AI to enhance their digital customer experience.

Many brands rush into chatbots and AI, yet customers keep reporting the same issues: rigid scripts, broken flows, and answers that ignore context. Just like a rider who only trusts data without feel, poorly implemented conversational AI can make customer experience worse instead of better.

This article looks at how banks, fintechs, e-commerce players and service providers in Southeast Asia can design conversational AI that feels more human, responsive, and measurable – taking inspiration from how Diggia approaches a race: focused on rhythm, consistency, and timely communication.

What Conversational AI Really Means for Enterprises

Conversational AI refers to technologies that enable machines to understand, process, and respond to human conversations across digital channels: chat, voice, and text. In a typical enterprise context, this usually appears as:

• Chatbots on WhatsApp, web, and mobile apps
• Voice bots for call centers and IVR
• Virtual assistants embedded into an omnichannel communication stack

If you think in MotoGP terms, conversational AI is like combining the rider, the data engineer, and the race strategist into one system that answers three key customer questions:

1. What’s happening? (order status, payment, service tickets)
2. What should I do next? (steps to complete, confirmations, product choices)
3. Who can help if this is not standard? (clear escalation to human agents)

The big difference between a generic chatbot and a mature conversational AI layer is its ability to understand context, remember history, and respond consistently across channels – from enterprise SMS to WhatsApp and voice.

Why Fabio Di Giannantonio’s Style Matters Here

Diggia may not be the biggest name on the grid, but his style offers a useful lens for CX leaders:

• He’s not the kind of rider who attacks every corner; he prefers to build rhythm and lap time consistency.
• He and his team rely heavily on telemetry to understand grip, engine behavior, and weather – and adjust strategy over race distance.
• He emphasizes two-way communication via pitwall radio: real-time feedback so the team can fine-tune the bike.

Translated into digital customer experience, Diggia’s approach suggests three design principles for conversational AI:

1. Rhythm over raw speed
   Customers need more than instant replies; they need a predictable, coherent flow every time they interact with your brand – across WhatsApp, webchat, and SMS.
2. Data as a guide, not a cage
   Conversational AI should learn from interaction data, but remain flexible enough to handle the unexpected.
3. Honest, two-way communication
   Like radio chatter, customers must be able to give feedback easily, and the system must know when to say “I don’t know, let me connect you to a person.”

Why Conversational AI Is Becoming Core to Digital CX

Over the past few years, customer behavior in Southeast Asia has shifted fast:

• Customers are used to transacting over chat, especially WhatsApp.
• Response-time expectations have moved from “business hours” to always-on.
• Customer journeys are now multi-channel: social ads, WhatsApp, SMS OTPs, app notifications, and follow-up calls.

In this context, conversational AI delivers three crucial benefits:

1. Scaling Responses Without Killing Quality

For enterprises serving millions of customers, it’s simply impossible for humans to handle every message. Conversational AI can take the lead on:

• High-frequency questions (store hours, delivery status, how-to guides)
• Transactional flows (top-ups, bill payments, plan changes)
• Proactive messaging and notifications via WhatsApp API and SMS

Human agents remain essential, but are reserved for complex, high-value, or emotionally charged cases.

2. Consistency Across Every Channel

Without an integrated conversational layer, customers often experience:

• Broken context between chat, email, and phone support
• Endless repetition of identity and account details
• Promotional messages that ignore their history and preferences

With an omnichannel approach, conversational AI can:

• Unify conversation history from WhatsApp, SMS, webchat, and other channels
• Respect channel preferences (e.g. WhatsApp for updates, SMS for OTP)
• Adapt tone and language to different segments and past interactions

3. Turning Conversation Data Into Business Insight

Telemetry helps race teams fine-tune the bike; conversation logs can help enterprises refine products and journeys. An effective conversational AI strategy lets you answer:

• Which questions typically appear right before a customer upgrades or churns?
• Where exactly do new users get stuck during onboarding?
• Which words or topics are sensitive for certain segments (fees, interest, risk)?

A good system doesn’t just answer questions; it learns continuously from them.

The Technology Stack Behind Enterprise-Grade Conversational AI

Discussions around AI often stop at high-level concepts. For a production-grade rollout in Southeast Asia, several technical building blocks really matter:

1. Local Language NLP, Not Just English

Real customer messages include:

• Abbreviations and slang (“gmn ya”, “tp blm masuk”, “lah kok?”)
• Code-switching across Indonesian, English, Thai, Vietnamese, etc.
• Typos, emojis, and non-standard grammar

Your NLP engine needs to:

• Understand informal local languages, including regional nuances
• Handle spelling errors and mixed-language sentences
• Classify intents and entities robustly in Bahasa Indonesia and other SEA languages

2. Integration With Core Messaging Channels

Conversational AI has to be where your customers actually are. For most Southeast Asian markets, that means prioritizing three channels:

1. WhatsApp Business API
   As the dominant messaging platform, integrating with an official provider such as WhatsApp Business API lets you:

• Deploy chatbots to serve queries 24/7
• Send personalised transactional and promotional notifications
• Hand over smoothly to human agents when needed

2. SMS Masking
   Still essential for mission-critical alerts and wide coverage, especially when:

• Delivering OTPs and login verifications
• Sending due-date reminders, service outages, or appointment alerts
• Reaching customers with poor data connectivity or inactive WhatsApp

Providers like SMSMasking.id offer local direct SMS connectivity to mobile operators, ensuring high deliverability and brand-safe sender IDs.

3. Omnichannel Platform
   Instead of building one-off integrations, enterprises can leverage an omnichannel messaging platform to:

• Manage WhatsApp, SMS and other channels in one console
• Share conversation context and history across teams
• Run unified chatbot journeys regardless of channel

3. Dialogue Orchestration and Backend Integration

Without proper backend integration – to CRM, core banking, billing, or order systems – your chatbot will act as a glorified FAQ. Dialogue orchestration should be able to:

• Pull customer data securely in real time (balance, order status, plan details)
• Trigger actions (change address, modify plan, reset credentials) with proper authorization
• Log conversations back into analytics and CRM platforms

The ideal architecture treats conversational AI as a layer on top of APIs, not a direct database client.

Designing Human-Like Conversational AI: Borrowing from Diggia’s Race Craft

Diggia often spends the opening laps feeling out grip before pushing hard. For enterprises, that translates into:

1. Start With Critical, High-Impact Use Cases

Instead of trying to automate everything, focus your first wave of conversational AI on a handful of journeys that strongly influence CX and cost:

• New customer onboarding (registration, activation, verification)
• Order and transaction status
• Billing, payments, and collections
• Basic account support (password reset, PIN help, usage tips)

For each use case, design the end-to-end conversation: trigger, identity verification, data fetching, action, and confirmation via WhatsApp or SMS.

2. Tone of Voice: Consistent, Concise, Empathetic

Great conversational AI has a distinctive yet subtle “voice.” Consider:

• Persona: How formal should your bot sound for banking, telco, or retail?
• Sentence design: Short, clear, and free of jargon whenever possible
• Empathy: Acknowledge inconvenience first, then move to resolution (“I’m sorry your payment is delayed. Let me check it for you.”)

And always make escalation easy: after two or three failed attempts, offer a clear path to speak with a human agent.

3. The Rhythm of Interaction: Don’t Flood Customers

Sending too many notifications is the messaging equivalent of over-braking on every corner. Shape your outreach strategy carefully:

• Control frequency (daily, weekly, monthly caps)
• Respect local time zones, holidays, and religious times
• Ensure relevance: only send offers that align with previous behavior and preferences

A platform like SMSMasking.id can help orchestrate WhatsApp and SMS campaigns so your messages feel like a helpful team radio, not spam.

Conceptual Case Study: A Digital Bank Finding Its Pace

Imagine a Southeast Asian digital bank aiming to lower contact-center costs, improve NPS, and accelerate onboarding. It designs a conversational AI roadmap in three “stints”:

Stint 1: Onboarding and Verification

• The customer fills in a sign-up form on web or app.
• The system sends an OTP via masked SMS through a direct operator gateway to verify their phone number.
• Once verified, a WhatsApp chatbot greets the user and offers step-by-step help completing additional KYC details.
• If the user drops off mid-way, a gentle WhatsApp reminder nudges them to continue, instead of blasting generic reminders.

Stint 2: Everyday Banking and Financial Education

• The chatbot instantly answers balance, recent transactions, and credit card limit queries.
• Users who repeatedly ask about savings or investments are proactively invited to receive tailored educational content via WhatsApp (with clear consent).
• All conversations are stored in the omnichannel platform, giving human agents full context when they step in.

Stint 3: Incident Handling and Failed Transactions

• When there’s a system disruption, the bank sends proactive WhatsApp and SMS updates, acknowledging the issue before complaints spike.
• Conversational AI offers quick checks: “Reply 1 to see your last transaction status.”
• Customers showing strong frustration (identified via keywords) are routed directly to trained human agents.

Within 6–12 months, the bank can reasonably expect:

• 20–35% reduction in low-complexity contact-center volume
• Higher CSAT / NPS thanks to faster, more consistent responses
• Better onboarding completion rates due to guided, multi-channel flows

From Shakedown Test to Race Weekend: A Practical Rollout Plan

Race teams never debut an all-new bike on Sunday without testing; they run shakedowns and practice sessions first. Enterprises can follow a similar four-step approach:

1. Discovery and Use-Case Prioritization

• Collect 500–1,000 samples of real customer interactions (chat, email, call transcripts).
• Cluster them into categories: information, transactions, complaints, education.
• Pick 2–3 high-volume, low-risk categories for phase one.

2. Rapid Prototyping on a Single Channel

• Build a minimum viable chatbot on official WhatsApp Business API with a provider such as SMSMasking.id.
• Keep flows simple and transparent, with clear handover to agents.
• Run during business hours before rolling out 24/7 automation.

3. Measurement and Fine-Tuning

• Track key metrics: response time, auto-resolution rate, CSAT, escalation share.
• Collect examples where the bot gets confused; feed them back into NLP training.
• Refine tone, question order, and language choices based on feedback.

4. Expansion to More Channels and Journeys

• Add SMS Masking as a notification and fallback channel.
• Extend conversational flows to billing, collections, and targeted cross-sell.
• Move orchestration onto an omnichannel layer so all channels are managed coherently.

Managing Risk: Security, Compliance, and AI Behavior

Just as an over-aggressive bike setup can cause a crash, careless conversational AI rollouts can create real business risks:

1. Data Security and Privacy

• Limit which sensitive data is ever displayed via chat (e.g. never full PINs or passwords).
• Use end-to-end encryption where possible (e.g. official WhatsApp), and secure internal APIs for SMS and backend integration.
• Comply with local regulations on data protection and electronic communications.

2. Messaging Compliance (WhatsApp and SMS)

• Use official WhatsApp Business API to ensure compliant templates, opt-in, and content.
• Control SMS frequency and content quality to avoid being flagged as spam by mobile operators.
• Provide clear opt-out options for marketing messages on every channel.

3. AI Bias and Wrong Answers

• Avoid letting unconstrained generative models answer financial, legal, or medical questions.
• Build guardrails and use structured flows for critical operations (payments, account changes).
• Audit a sample of conversations regularly to catch faulty or biased responses early.

Partnering vs Building: Strategic Choices for CX Leaders

Not every enterprise needs to build its own conversational AI stack from scratch. The build-vs-buy decision is similar to choosing between a full factory MotoGP effort and a satellite team package:

• Build if you have strong in-house tech, need deep differentiation, and are ready to invest for years.
• Partner if you want faster time-to-market and reduced technical risk.

Messaging specialists like SMSMasking.id provide:

• Production-ready WhatsApp Business API connectivity and direct-operator SMS with masking
• An omnichannel messaging platform that unifies channels and teams
• Flexible chatbot / AI integration that can be tailored to your business logic

A common pattern in Southeast Asia is to start with a trusted partner to establish the core messaging and conversational layer, then gradually build proprietary models or journeys on top of that foundation.

Racing Towards a More Mature Digital CX

Fabio Di Giannantonio shows that you don’t need to be the biggest star to deliver impressive results; what matters is consistency, reading the situation, and teamwork. The same applies to brands investing in digital customer experience.

Deployed well, conversational AI is not about replacing humans, but about:

• Giving customers fast, simple access to help on the channels they already use
• Ensuring a consistent tone and standard of service across WhatsApp, SMS, voice, and web
• Using conversation data to improve products, journeys, and operations end-to-end

In an increasingly competitive region, the brands that combine sharp data, empathetic conversations, and disciplined omnichannel messaging orchestration will be best positioned to “win the long race” for customer loyalty.

Like Diggia finding his rhythm in a grid full of giants, your organisation can also build a more mature digital customer experience – with conversational AI as a core engine, and a reliable messaging backbone from partners such as SMSMasking.id.