Not All AI are Inaccurate, When Data is Right: Sujith Abraham, Salesforce
In an era where consumer trust is at a record low with AI, businesses are under pressure to deliver exceptional experiences while navigating the complexities of this technology. In an exclusive interview with Sujith Abraham, Senior Vice President and General Manager of Salesforce ASEAN, he sheds light on how companies can rebuild trust in AI and drive customer success. "It would be a mistake to broad brush all AI as inaccurate,” he states, and emphasise the crucial role of data in ensuring AI accuracy and effectiveness.
Inaccuracy in AI
Sujith suggests that companies must examine the root causes of inaccuracies: “Is it because they are DIY-ing their AI? Do they have access to ALL of their customer data and are they unifying it without copying it?” By asking these critical questions, companies can begin to address the issues and ensure that their AI solutions deliver accurate and relevant results. Instead of focusing on the speed of innovation, businesses should focus on leveraging their data to create better customer and employee experiences. This includes a unified view of the customer, allowing for more impactful and personalised interactions through AI. "Right data makes good AI," he noted, further stressing the importance of a strong data foundation. He also provides an example of how Saks Fifth Avenue, one of Salesforce's clients, utilises Agentforce to provide real-time, tailored recommendations and resolve queries, improving customer satisfaction and operational efficiency
Most Singaporeans believe that companies are reckless with their data
The State of the AI Connected Customer research, which was published by Salesforce, has revealed that a significant 61% of consumers in Singapore express concerns over companies' handling of their data, indicating a perception of carelessness in data management. This lack of data trust and the perceived inaccuracy of AI, seems to indicate that the AI implementation for businesses is more of a strategic long-term luxury than an immediate solution where short-term financial gains take precedence, which is more in line with the current business landscape. However, Sujith disagrees with this assessment and reiterates that AI is not solely a long-term objective; rather, "AI agents can provide immediate value to businesses." AI agents are capable of responding to consumer enquiries without the need for human intervention. This can result in increased customer satisfaction and higher margins and increased customer satisfaction if the AI is grounded in the right data.
Transparency and trust are paramount for consumers when interacting with AI agents
Sujith also notes the importance of transparency, quoting from the research that 76% of consumers want to know if they are interacting with an AI agent. Furthermore, a majority of consumers are more likely to engage with an AI agent if its logic is explained and if there is an easy way to escalate to human support. By making these aspects clear, businesses can foster confidence in their AI systems.
Millennials and Gen Z, are more open to AI but are sensitive to their needs and perspectives
The customisation of AI is exceedingly complex due to the profound sensitivity of today's social subtleties and viewpoints. Although the younger generation is more likely to adopt new technology, the challenge lies in the formulation of AI responses that are both authentic and respectful in the presence of rapidly evolving social environments, where comprehension can vary considerably even among demographic groups that appear to be similar. To accomplish this, AI systems must exhibit exceptional sensitivity in navigating intricate cultural, emotive, and historical contexts, a feat that occasionally even humans were unable to achieve.
However, Sujith believes that Agentforce can overcome this issue through its Atlas Reasoning Engine, which uses Large Action Models (xLAM) to simulate human-like reasoning and continuously evolve based on context. Built on the Einstein Trust Layer with features like dynamic grounding and toxicity detection, the AI solution is designed to deliver highly accurate, contextually relevant outcomes while adhering to the company’s AI policy. "Our Acceptable AI Use Policy also ensures we have specific protections in place like disallowing the use of our AI solutions for explicitly predicting protected characteristics such as age, gender, health status, medical conditions and financial status to name a few" he highlighted.
AI is now?
According to Sujith, companies should not shy away from AI implementation out of fear of inaccuracy. He notes that by not moving forward with AI adoption, companies are "missing out on the transformative potential the technology offers". He argues that instead of slowing down, companies should focus on building strong data foundations and adopting solutions that can be implemented quickly and efficiently. He points to platforms like Agentforce as a way for companies to quickly implement scalable AI that is also capable of delivering excellent customer service. He concludes that "Autonomous AI agents can help businesses rebuild trust and scale to meet increasing demands, by elevating and personalising customer experiences at scale in a trusted manner”
Source: tech360tv