The rapid adoption of AI has quickly become yesterday’s news. Yet, we remain at the very edge of its capabilities. In recent months, the internet has been abuzz with debate around the ethical implications of new models entering the market – DeepSeek probably rings a bell. 

As AI agents and conversational AI continue to evolve, the debate is only set to intensify. However, as Twilio solutions engineering lead Christopher Connolly suggests, marketers have a huge opportunity to leverage data together with conversational AI to build customer trust like never before.

Alibaba Group entered the artificial intelligence (AI) race last month with the release of its open AI tool, Qwen 3. Arriving just months after DeepSeek’s controversial launch, Qwen 3 looks set to further stir the pot on the ethical AI debate in Australia. While we work to address concerns of AI misuse from these emerging technologies, we cannot ignore the fact that AI is already commonplace across local businesses.

A recent study conducted by Google and Ipsos found 74 percent of AI users in Australia are using it for work. AI stacks have also become a must-have on marketers’ wishlists, quickly evolving from essential question-and-response platforms to dynamic, outcome-driven systems capable of driving impact throughout the customer journey. 

However, as concerns around DeepSeek and new entrants demonstrate, there is still work to be done to prove the trusted use of AI. One core area of opportunity that no marketer should ignore is conversational AI, which can be used ethically and effectively to build consumer trust and drastically improve business outcomes. 

How conversational AI bridges the chatbot trust gap

Relationships thrive on meaningful conversations – something most chatbots, often serving as temporary stopgaps, struggle to deliver. These bots rely on scripted responses, which limit their ability to solve issues or foster genuine human interaction.

Conversational AI bridges this gap by using advanced multimodal large language models (LLM) to really understand what people want and respond in a natural way. Unlike basic chatbots, it can handle extended interactions such as rescheduling tickets or facilitating product returns. 

Conversational AI can also serve as a sales assistant that can guide customers to the right products, replacing traditional e-commerce search categories and reducing unnecessary steps in the purchasing process. It can reason with customers, recall past interactions and deliver tailored responses while being available twenty-four-seven to resolve issues promptly and efficiently without human intervention. As a result, contact centres can scale basic tasks typically handled solely by human agents, significantly optimising resources.

Global financial services company ING took this leap, introducing custom APIs with Twilio providing AI-powered voice, chat and video capabilities to enable it to resolve many queries autonomously. Moving to dynamically generated conversations provides superior customer experiences by providing real-time updates and guiding customers through each step with a tone that conveys reassurance and clarity. 

Using conversational language that mimics natural speech patterns can make customers feel heard and understood, which is key to building trust. ING is also reducing risks related to accuracy, bias, security and compliance, alongside real-time feedback loops to ensure continuous improvement.

Conversational AI brings the human factor to customer experience

Of course, for people to feel comfortable enough not dealing with a human, the AI interaction needs to feel natural. Recent advancements are improving the way AI acts and reacts. Leveraging AI, brands can create interactive voice response (IVR) flows with synthetic agents capable of handling routine enquiries such as product returns or fee waivers using natural, contextually aware responses.

Cedar, a North American healthcare financial services company, has streamlined patient communications using AI-driven automation and empathetic support by integrating an AI-powered Conversation Relay service, SMS and voice APIs to enhance patient accessibility, improve financial interactions and reduce wasted expense. AI-powered voice agents provide smart, personalised support, offering instant responses, and can also trigger human intervention when needed to create a meaningful experience.                   

Marketers need accurate data foundations

So, how do you build conversational AI into a content and engagement strategy? The short answer is simple – decent data foundations. When agents use incomplete or inconsistent information with customers, it often leads to confusion, missed opportunities for meaningful connections and impersonal or disjointed experiences. High-quality, clean and accurate data leads to better AI models that can understand and respond accurately to user enquiries. 

Consolidating data from various sources like Martech platforms, customer relationship management (CRMs) and IT service management (ITSMs) into a centralised repository that captures all touchpoints and interactions with a brand’s products or services provides a more comprehensive view of AI models to deliver personalised and contextually relevant responses. 

For example, Twilio Segment can create ‘golden customer profiles’ that integrate individual customer data, including preferences, purchase history and demographic details, into a unified view. As they can better understand the context, these agents can offer faster resolutions, assisting with completing a purchase, suggesting alternative products, or addressing concerns that may have led to cart abandonment.

Of course, ongoing data maintenance and refinement are essential to keeping the data ecosystem safe, adaptable and continuously improving. 

Build conversational AI on trusted infrastructure

Instilling trust in conversational AI also hinges heavily on the nature of the infrastructure it is built on. DeepSeek has placed concerns about data travelling across borders in the spotlight, and we know that trust is quickly lost if customers believe their data is transiting to destinations they disagree with. The reality is that data on DeepSeek is sent to China only on the mobile app, however DeepSeek (the model) is an open-source model that can be run locally or on owned infrastructure, without data heading across geographies. 

This is why it’s important for marketers to work with a partner to help differentiate between applications and the underlying AI model when considering the use of conversational AI. Once the distinction has been made, it’s possible to then navigate the pros and cons to develop a fit-for-purpose model.
           
Marketers have a huge opportunity with conversational AI to prove that AI can be used for good through faster responses and smarter customer connections. Most importantly, conversational AI presents an opportunity to alleviate consumer fear and establish trust by demonstrating real value from data.

Christopher Connolly is solutions engineering lead for communications business, Twilio. Based in Sydney, Connolly leads the development and implementation of innovative, customer-centric strategies. A recognised thought leader in the field, Connolly holds eight patents for customer experience innovations. His extensive international experience across banking, telecommunications and government sectors has equipped him with the expertise to craft bespoke solutions tailored to unique needs across diverse markets.