Enterprise CIOs are under immense pressure to adopt AI quickly, but data concerns are preventing widespread implementation
CIOs have always been critical to the organization. However, the explosion of AI has thrust CIOs into the spotlight, with the expectation that they need to be AI experts.
Enterprise CIOs are feeling the pressure to adopt AI quickly. However, according to a new Salesforce survey, most are still just experimenting with the technology – not fully implementing it. The Salesforce survey, conducted with 150 verified CIOs from companies with over 1,000 employees, revealed that while 84% believe that AI will be as significant as the internet, only 11% have fully implemented it.
CIOs are hesitant to go all-in on AI, pointing fingers at data and security concerns as the main barriers to adoption. As we move into a world of autonomous AI, how can CIOs move past experimentation into implementation? DigiconAsia checked in with Gavin Barfield, Vice President & Chief Technology Officer, Solutions, Salesforce ASEAN, to find out more…
While most CIOs agree that AI is a game-changer, many are cautious. What are some key obstacles to AI adoption?
Barfield: CIOs are under significant pressure to adopt AI quickly. Yet, many businesses remain in the experimentation stage with the technology. A recent Salesforce survey found that despite 84% of CIOs believing that AI will be as significant to their businesses as the rise of the internet, only 11% have fully implemented the technology.
Despite the overwhelming hype around AI, many companies find that early iterations of generative AI such as copilots fail to deliver real value. Copilots could only provide suggestions and assistance but not take action autonomously, as they often operate outside the flow of work and are disconnected from organizational processes. This makes it difficult for businesses to deploy AI within the enterprise to solve real business problems.
In contrast to copilots, autonomous AI agents understand context, process inputs, and take action on behalf of humans to execute tasks seamlessly within workflows, delivering personalized and efficient outcomes. Another challenge with AI adoption is fragmented data sitting in silos across different systems – from structured data such as transaction records, inventory and customer profiles; to unstructured data such as documents, PDFs, and emails. When data is siloed, AI can’t get a unified view to deliver accurate, contextualized, and actionable outputs.
Unsurprisingly, the same Salesforce survey found data to be one of the main hurdles to AI adoption, with over half (52%) of the CIOs citing the lack of trusted data as one of their top AI implementation fears. Many do not trust that their data is accurate and up to date, while others struggle with incomplete and inaccessible data.
Security is also another major hurdle to AI adoption in the enterprise. While CIOs are keen to implement AI across their organization, many are concerned that AI use can introduce security and privacy risks. Without the necessary guardrails in place, AI runs the risk of hallucinating, producing toxic and biased content, and even exposing sensitive customer data – which can erode trust in the organization.
Which parts of business can an organization get the most value from adopting AI?
Barfield: We are entering the third wave of AI with autonomous AI agents. These AI agents are changing the game, as they can make decisions and take action without human intervention. AI agents allow businesses to automate repetitive tasks and free employees to focus on higher-value work, which ultimately improves operational efficiency, productivity and growth. However, businesses will need to identify use cases that address critical pain points in their operations for initial AI deployment, before gradually expanding based on measurable outcomes.
Customer service is one of those functions. According to a McKinsey study, 75% of AI’s value comes from areas like customer operations, marketing, and sales—functions where AI agents can directly engage customers and drive measurable business outcomes.
With AI agents, organizations can now reinvent the service experience and deliver personalization at an unprecedented scale. Unlike traditional chatbots that follow predefined, rules-based dialogues, AI agents can reason and ground answers in relevant knowledge and context. They can converse in natural language, switch topics and have a conversation similar to how we would with a human.
For instance, they can offer personalized recommendations based on customers’ preferences, or suggest products based on customers’ past purchases. They can also anticipate customer needs and provide proactive support, such as sending reminders for upcoming appointments or notifying customers about potential issues before they arise.
Saks, a premier fashion retailer, is transforming the luxury shopping experience by using Agentforce, Salesforce’s suite of fully customizable autonomous AI agents, to streamline order management and deliver a high-touch tailored customer experience for millions of customers.
Beyond customer service, AI agents can help teams across various departments handle routine tasks, freeing human employees to focus on high-value activities that drive customer success.
In sales, teams can focus on high-potential opportunities with agents pre-qualifying leads. In commerce, AI agents can respond to customers directly on your commerce site or on messaging apps like WhatsApp, and help customers make purchases faster by guiding search queries and tailoring product recommendations to the shopper.
In marketing, agents can generate a campaign brief and target audience segment, then create relevant content speaking to those audiences.
According to Salesforce’s recent survey, there is a mismatch between AI business value, enthusiasm and readiness, which can hinder the technology’s effective implementation across the enterprise. How can CIOs address this gap?
Barfield: A common challenge for CIOs is defining where and how AI should be used in their organization. While AI use cases may be more apparent in certain business functions, those departments may not be the most prepared or open to adopting the technology. For example, the Salesforce survey revealed that CIOs view customer service as having the most use cases for AI, but believe service agents are the least enthusiastic about AI adoption.
As a start, CIOs should lean on AI solutions grounded in trusted data and a complete, integrated platform that combines AI, data, automation and security. In addition, AI should be set up directly where work is done, so it can handle tasks, make decisions, and provide insights right where they are needed, boosting efficiency and productivity.
To gain employee buy-in in the technology, focus on securing quick wins by identifying high-value cases that address critical pain points.
In customer service, identify use cases that directly enhance customer interactions, such as improving response times. In sales, AI agents can pre-qualify leads so that humans can focus on high-potential opportunities. Successful pilot projects that showcase AI’s potential can help convince skeptical employees of the benefits of AI and make the case for broader AI implementation within the organization.
CIOs should also focus on enabling their workforce with the necessary skills, tools and guidelines to harness AI effectively. A positive culture where employees are encouraged to innovate and experiment with AI can help them build confidence in using AI at work. Through experimentation with AI, employees may even create new use cases that help drive positive business outcomes.
How can businesses prepare their people, data and processes to adopt AI agents effectively?
Barfield: Here are some ways…
People: Businesses must equip their people with the necessary skills to harness AI agents effectively. Fundamentally, they should understand what an AI agent is, how it reasons and takes action, which use cases it can be applied to, and how they can build their own AI agent. Understanding how they can work alongside AI agents will be critical for humans to drive customer success together with agents. To ensure that workforces have the skills to succeed with AI, Salesforce is expanding access to skilling opportunities with free AI training for anyone through its free online learning platform, Trailhead, till the end of 2025.
As AI agents become central to business operations, there is also a greater need for professionals with specialized skills to guide these systems effectively. These skills include defining agent instructions, crafting prompts, and setting guardrails.
Data: Success with AI agents begins with a trusted data foundation. Businesses must unify structured data, such as customer transaction records, and unstructured data, such as customer emails, product information and corporate policies, for AI to get a comprehensive view of the customer and deliver accurate, contextualized and actionable outputs. They should also leverage trusted and secure technology with built-in safeguards against hallucinations, toxicity, privacy risks and bias. Companies like Saks lean on Salesforce Data Cloud to unify their customer data, so that its AI agents can make informed decisions and take action to transform the customer experience.
Processes: Speed of AI implementation is critical. Now is not the time for businesses to “DIY” their AI strategy – those who have done so have faced hidden costs and a slow realization of AI capabilities. Instead, they should lean on out-of-the-box, easily customizable AI solutions such as Agentforce, that can be set up quickly, are easily scalable, and can work around the clock across any channel. Wiley, a global publishing company, saw a 40% increase in case resolution after using Agentforce to streamline customer service operations. With AI agents handling routine queries like customer access and payment issues, its customer service teams could focus on resolving more complex tasks, allowing Wiley to efficiently handle customer service requests during peak periods like the back-to-school season.
