AI and Privacy: Walking the Tightrope to Acceptance and Success

When it comes to AI in financial services, there’s a lot to be excited about. Especially in keys areas like critical decision making where its potential for helping humans to make more accurate assessments and decisions represents a real breakthrough. And AI is also key to enhancing customer service and helping deliver deep and rewarding experiences.But, as with any new technologies, there are concerns which need to be addressed. For most of us, the prospect of allowing AI powered algorithms to make decisions about our money or personal safety sounds risky. And the memes and cliches about AI, as well as the scare stories about its effects on jobs and the spread of misinformation, threatens to slow our ability to benefit from its enormous potential.Those concerns are natural, and we need to be wary about how banks, airlines, online shopping brands, and the government store, use, and share it. You just have to look at the research that’s being conducted all over the world. The public are sceptical about AI. Work done by Forbes Advisor found that in the UK 90% of people are worried that their data will be hacked. 63% worry that the data the government holds about them isn’t safe. And levels of trust in technology platforms hover around 50%.

Transparency and honesty engender trust

At our Financial Services Connect event in London last month there was agreement that the answer to those concerns is to show consumers how AI can work to deliver personalised, generative experiences which get things done quickly and accurately, and gives the specialists they rely on, the ability to do more for them. That approach is supported by academic research into how the public begins to trust new technologies. When you involve the public and clearly explain how technologies, systems, and protections work, you build greater trust. Transparency and engagement works.That’s why Salesforce’s announcement of the Einstein 1 Platform as a key element in our Service Cloud is important. It’s aimed at reassuring the public that we’re walking that tightrope between innovation, deeper privacy and the ultimate protection of personal data. The platform addresses the issues which regularly hit the headlines, head on: namely, the ownership of data, reducing the frequency of errors and protecting the integrity of each dataset, and ensuring AI doesn’t ‘hallucinate’ or come to illogical conclusions and dead ends. And it does it while making sure that AI always delivers appropriate responses free of bias or toxic influences.So, how does the Einstein 1 Platform change the game around AI? It’s all about trust. And that’s the foundation of the platform, which has Einstein 1 Data Cloud natively integrated into it. That allows organisations to connect any data to achieve a unified profile of each customer. Most importantly, Einstein enables you to infuse AI, automation, and analytics into every customer experience.Central to the way Salesforce works with data is the Trust Layer. You can bring in data from multiple sources using different technologies, but that retrieval is always secure. Data is masked and de-masked (names, credit card numbers etc.) and zero retention means that data is never stored unnecessarily, it’s always where it is for a purpose, and when that task is done, it goes back to where it came from. Securely.

Pioneering a trust-based approach to AI

Einstein Copilot: Making it easier to unlock the potential of AI

Salesforce has worked hard on not only developing our Service Cloud to get ahead of privacy concerns, but also to ensure that ‘the AI dream’ comes true in ways that suit our customers. We’re engineering the algorithms to do both simple and complex tasks logically: for instance, if they don’t know the answer, they won’t make one up, they’ll say they don’t know. Many AI engines don’t do that. And that gives AI a bad name.What will give AI a good name is its power to support and expedite human interactions. That’s especially important in financial services. At our FINS Connect event we heard about a range of use-cases that are happening right now. One told the story of a customer who was just starting the process of applying for a mortgage. As she began to upload her documentation she realised that she’d lost her debit card. That meant she would have to stop the application process and then get in touch with a different part of her bank, something that, she was sure, would not only lose the time she’d already spent, but result in even more wasted time navigating the bank’s different departments.She knew she had to do something, so she called to report her lost card and request a new one. The agent (an actual person) quickly cancelled the lost card and ordered a new one. Then she surprised the customer by knowing all about her mortgage request. AI was working behind the scenes to match the two events and suggest both specific actions and wording to the agent so she could reassure the customer on both fronts: Her debit card was in the post, and her mortgage application was progressing smoothly. Decisive, informed, and secure action like that is what will build trust in AI.

Behind the scenes what really counts is that the technology seamlessly integrates with your business where it matters most; the point where you interact with a customer. Just as I described in that story. So, Einstein Copilot is our new generative AI-powered conversational assistant that can be matched to each of the Salesforce applications you’re using. It works seamlessly into workflows and is key to achieving significant productivity gains.Of course, every use-case is different, just as every customer is. So, the platform offers Einstein Copilot Studio which enables you to customise your Copilot for your unique needs. You can tailor it to prompt specific answers and conduct specific conversations to match your objectives and the needs of your people. Again, the Trust Layer is its foundation. Conversations generate data and that data has to remain private. Especially when generative AI is used to analyse it and then generate responses.The tool kit offers Prompt Builder, Skills Builder, and Model Builder, which simplifies the AI-user interaction, promotes versatility and adaptability in terms of the way AI works each time, and you can add in bespoke language – relevant to your products etc – to achieve personalisation even when it’s an algorithm that’s communicating with the customer.

Conclusion: Generative AI is here and now is the time to trust it

Parker Harris, the Co-founder and CTO of Salesforce, summed up this whole discussion succinctly when he said, “A company’s AI strategy is only as good as its data strategy.”Generative AI is all about data. Without data the algorithms can’t function. And that data has to be pristine, protected, and powerfully targeted. When you achieve that, and can prove it, then customers will trust you with their data because they know it’s being used for their benefit. The wave of AI enabled communication tools that are being used in everyday commercial communication is growing. People are used to being offered chatbots online and interacting with AI powered assistant on the phone or on a screen. But their acceptance of the technology can only be based on how good their experiences are. It’s obvious that better, faster, and more accurate responses will drive acceptance levels higher. And that’s important because the range of possible applications is vast: AI is moving forward among organisations.We must engender trust. That’s the last barrier. The utility of generative AI is clear, but trust matters when it comes to achieving the ‘AI dream’ – when it’s an accepted element in nearly everything we do; when we engage with our bank, search for insurance or investments, or seek advice about how to manage our financial lives, for instance. That’s why I believe that the forward-thinking developments around our Einstein 1 Platform are game changers. Trust is the cornerstone of all the AI solutions we offer, and that’s how we’re going to make the AI dream – a reality.

Aman Virk

Aman Virk, Senior Director, Customer Transformation, Financial Services, Salesforce. Aman is a Customer Transformation Senior Director at Salesforce. She supports the Financial Services sales team as a trusted industry advisor, shaping the go-to-market strategy, building partnerships across strategic accounts and driving our social agenda on financial inclusion and ESG strategies.