Research by consultant Accenture forecasts the economic impact of generative artificial intelligence (AI) in the enterprise. The 2024 report suggests more than$10.3 trillion in additional economic value can be unlocked by 2038 if organizations adopt generative AI responsibly and at scale. 

In addition, business leaders noted in the research that generative AI will ultimately increase their company's market share, and 17% anticipate an increase in market share by 10% or more. As much as 95% of workers see value in working with generative AI -- but their top concern is that they don't trust organizations to ensure positive outcomes for everyone from the introduction of emerging technology.

Also: AI's latest trick: Better KPIs for measuring business success

Meanwhile, MuleSoft's ninth annual Connectivity Benchmark Report, produced from interviews with 1,050 IT leaders across the globe, reveals that the AI inflection point amplifies the need for a coherent IT strategy. Eighty-seven percent of IT leaders report that the nature of digital transformation is changing. AI further complexifies the tech landscape, with 991 apps in the average enterprise. IT budgets need to increase to meet this surging demand. 

The research also found that integration and security concerns are the biggest barriers to AI adoption. The AI genie is out of the bottle, with over three-quarters of organizations reporting they use multiple AI models. As many as 90% say difficulty integrating AI with other systems is a barrier, followed by 79% reporting security concerns. But perhaps the biggest barrier to AI adoption is a mix of data silos and systems fragility holding companies back. Almost universally, 98% of IT leaders report facing challenges regarding digital transformation. 

The persistence of data silos is mentioned by 81% of respondents and the fragility of tightly coupled and highly dependent systems by 72%. Data silos are preventing automation projects from being completed on time and within budget. Automation is still a source of contention between IT and the business. Business users benefit greatly from the automation of their work (1.9 hours per employee per week) and demand more flexibility to automate. However, the majority of IT departments still need to figure out how to enable this automation in a secure and governed way. Two-thirds (66%) of automation projects have IT as the sole gatekeeper. 

To better understand the impact of generative AI adoption that aims to improve customer experiences, I interviewed two of the world's leading experts on customer relationship management (CRM), customer experience (CX), and customer service. Michael Maoz is senior vice president of innovation strategy at Salesforce. Before joining Salesforce, Maoz was a research vice president and distinguished analyst at Gartner, serving as the research leader for the customer service and support strategies area. Ed Thompson is a senior vice president of market strategy at Salesforce. 

Also: How renaissance technologists are connecting the dots between AI and business

Before joining Salesforce, Thompson was a research vice president and distinguished analyst at Gartner, covering customer experience (CX) and CRM strategy and implementation. Maoz and Thompson shared their points of view on what businesses need to consider and implement before applying generative AI solutions to their customer service applications and processes. 

Over 95% of large organizations are running a pilot or live production of some form of generative AI. These companies have moved beyond researching and evaluating possibilities and moved into taking action. Midsize organizations have often identified over 100 possible use cases and the largest have over 500 defined. Among the business leaders we work with, responsible for marketing, sales, customer support, and overall customer experience, the conversations have progressed from "What could generative AI do?" to answering the question: "Does generative AI deliver value for my business unit, for my business, and for my customers?"

A few organizations have made headlines using generative AI for drug discovery or chip design, as they use skilled internal resources to tune large language models for high-value, game-changing use cases. Some organizations have become proficient in tuning small language models for specific industries or a single type of high-value process. However, a much higher percentage of organizations have taken a "lowest risk" approach and started with internal projects that do not expose the generative AI outcomes directly to customers or suppliers.

Also: The future of generative AI: Here's what technology analysts are saying

For those who have moved beyond employee productivity or application development, the most common area of focus has been the customer service department -- more so than marketing, sales, or commerce. So now everyone is wondering, what have early adopters achieved with generative AI in customer service? The answer is a mixed bag. Some are having spectacular successes, but many are struggling to obtain value. The rude awakening many are experiencing is that switching on generative AI does not necessarily (and likely will not) deliver returns.

Generative AI is not a silver bullet

Businesses around the world hope that, beyond the hype of generative AI, there lies a near-term path to improving business efficiency and in parallel a longer-term ability to grow revenue. There is one, not insignificant, consideration to weigh before the true savings can be measured. In 2024, as in 2023, generative AI and ChatGPT both trail "Customer Service / Telephone number" as search terms on Google in most countries. Most of those searches involve a customer's quest to reach a human being. There is great frustration because most businesses are working hard to make it difficult to reach a person.

This gap between the corporate commitment to removing the human connection in customer service and the customer's desire for a human connection almost always points to a bad business process. The business must examine why the customer doesn't use the self-service channel. This discovery process is a precursor to deeper self-service powered by generative AI.

Also: 4 ways to overcome your biggest worries about generative AI

Our first recommendation is to step back and ensure the customer service process you want to supercharge with generative AI satisfies customers. Layering generative AI on a broken process won't provide a magic Band-Aid. First, fix the broken process, and then you can safely run ahead with generative AI. But surely there must be some easy wins with generative AI in customer service?

"When we add generative AI to a bad customer self-service process, we have made things worse, only at a greater scale and with more severe consequences than before. A chatbot that gives customers incorrect information without generative AI does so because the underlying knowledge base is incorrect. Once generative AI is added, the company will find that the incorrect information is now given to the customer with perfect confidence by the bot, and not because it is hallucinating. The result: Now, the company has more eloquent false information disseminated in a new form," said Michael Maoz, SVP of innovation strategy at Salesforce.

Early success in customer service: Start simple

We have already seen dozens of customer service departments that are successfully using generative AI across four powerful use cases:

1. Post-call wrap-up and case summarization: Customer advisors and support agents find this part of the job tedious and difficult. As a result, they enter the smallest amount of information necessary. When using generative AI, especially when there are prompts to guide them, agents can save a minute or two after each call or case, while providing more and better summary information. That summary information can be analyzed to find patterns in the calls, which show the health of the process and suggestions on how to fix the issues. Generative AI is also a valuable source of training material for onboarding new agents and advisors.
2. Personalized content for the customer during a chat or phone session: When generative AI is built into the chat or phone conversation, the support agent can use the customer's questions as prompts as to the best advice or answers to provide. When connected to a strong knowledge base and customer profile, the generative AI can assemble a highly personalized answer for the agent.
3. Customer onboarding: The most critical post-sales step in the customer lifecycle is probably onboarding. Welcoming the customer, reminding them of products and services they have purchased, and leading them through personalized training is a great way to show the customer that you value their business. This approach can be a completely automated self-service process.
4. Sentiment analysis/Service Intelligence: Generative AI is used to analyze the words and responses chosen by the customer, and the responses given by the system or agent. The CRM system can notify the support agent or the service manager of any emerging issues and quickly intervene to fix the concerns.

As an aside, an exciting additional technology is voice-to-text, where humans can speak naturally and generative AI understands the context and delivers a text answer back to the customer. This approach can be a bit more complex than helping with call wrap-ups as the voice quality, language, accent, and complexity of technical terms can interfere with the process. But, in short, the second recommendation for IT leaders is to start simply with generative AI.

Complex customer interactions are more difficult 

After more than 18 months of experience, we have ample proof of the ability of generative AI-powered bots to deflect calls to human agents, and also as a way to deflect human interactions when the customer has a simple query. The query can be requests such as bank balance, order status, billing information, update profile, meter reading, deadlines, or answers to a wide range of frequently asked questions. All of these answers involve facts. Businesses require a solid knowledge base, a properly defined customer service process, and connections to the right systems. Not much workflow or data cleansing is required.

"At the moment, companies fall into two schools of thought or two approaches, as regards intelligent bots: They either start by thinking about enhancing the productivity of their agents, or they have as an aim the use of generative AI to reduce the number of customer service agents. Managers of customer service departments are often skeptical about the promises of generative AI-powered bots as a tool to help with agent productivity," said Ed Thompson, SVP market strategy at Salesforce. "Over the years, many promises were made by software suppliers that failed to materialize. By comparison, the heads of digital, COOs, and finance heads are not as focused on agent productivity. They tend to be excited about the possibility of reducing costs using the new generation of generative AI-powered bots."

Also: Generative AI advancements will force companies to think big and move fast

Thompson also said: "Generative AI is bringing new concerns such as bias, hallucinations, and toxic answers. But the old concerns remain and may be more serious. Perhaps generative AI will provide the impetus to solve unintegrated and siloed data sources, out-of-date data, poor quality data, incorrect data, and bad service processes and workflow design. Recommendation: Don't repeat the mistakes of previous bot implementations."

When the customer request is more nuanced, such as when the answer to a question is something like, "that depends on your priorities," or, "that requires that we weigh many factors," generative AI is not the best answer until the process is changed to one with a single right answer.

An example of a simple transaction turning into a complex transaction is when "where is my order?" becomes: "I would like to split my order and send one part, the part ready now, to one location and charged to one credit card, and the rest of the order that is on backorder sent to a second location in another country and charged to a second credit card not yet on file. Could you help me with that?" The technology for these more intricate processes requires coordination from multiple teams across the business and IT.

The complexity derives from the fact that many more systems, workflows, and rules must be invoked. That process means integration and design logic. Many complex interactions are not easily handled by automation or generative AI unless significant customer experience process design work is undertaken, followed by significant data access, workflow, and prompt building. We would not recommend these areas as the best place for a beginner to start on the generative AI journey.

Summary: Success requires a systematic approach

The point that we want to drive home is that processes must be checked and double-checked before applying generative AI. Mistakes can come back to embarrass the company that built the generative AI system. Recent examples include a Chevrolet chatbot telling the customer false information that resulted in financial loss and Air Canada attempting to make the legal argument that its chatbot was a separate entity whose answers were not the company's responsibility. The machine cannot discern between good and bad data or good and bad processes. The machine only sees data and follows whichever process and rules were created for it.

That reality brings us to our third recommendation: You had better ensure that the data your generative AI consumes is accurate and that the customer service process you have designed is within the ability of the generative AI to manage success. Start small. Start where there is measurable value and benefit. Start simple. Start with clean data and invite a small number of loyal customers into your design process to ensure you are on a path to success.

This article is the first of a two-part series on generative AI and its impact on customer experience. Join us for part two, where we will discuss ways to operationalize your generative AI strategy for speed, scale, and profit.

This article was co-authored byMichael Maoz, senior vice president of innovation strategy, Salesforce, andEd Thompson, senior vice president of market strategy, Salesforce. 

Artificial Intelligence

I asked Gemini and GPT-4 to explain deep learning AI, and Gemini won hands down

How to use ChatGPT's file analysis capability (and what it can do for you)

I tried Copilot Notebook: Microsoft's new AI tool offers two handy prompt features

What to know about Mistral AI: The company behind the latest GPT-4 rival

• I asked Gemini and GPT-4 to explain deep learning AI, and Gemini won hands down
• How to use ChatGPT's file analysis capability (and what it can do for you)
• I tried Copilot Notebook: Microsoft's new AI tool offers two handy prompt features
• What to know about Mistral AI: The company behind the latest GPT-4 rival