AI Trends from 700+ Business Leaders Reveal Global Keys to Success

Nine out of ten leaders know a strong data strategy is critical to AI success – so why are only a third of them integrating a unified data strategy across their company? In January 2024, Salesforce commissioned Forrester Consulting to find answers to these gaps and more in the rapidly developing AI landscape. 

The research, conducted with 773 business leaders in 14 countries, sheds light on global business decision-makers’ mindsets around AI-powered CRM.  These trends prove that regardless of fast shifts in AI, there are certain foundational principles that companies must prioritize for growth.

Before we explore these key success factors and recommendations, let’s understand the current state of AI-powered CRM.

Gaps in basic AI understanding reveal room for growth

The research reveals organizations worldwide are embracing AI across various CRM use cases. But, it’s not all smooth sailing. The research uncovers critical gaps that could impact the success of AI adoption in CRM.

Here’s a striking revelation: Despite all respondents making plans to adopt AI, only half were able to choose the correct definition for both predictive and generative AI when presented with both side by side.

Considering that an AI strategy likely encompasses both types of AI models –  which serve specific purposes and goals – this signals an opportunity for more comprehensive education. People need to know the specific use cases each type of AI enables, the anticipated business outcomes, and how to design a strategic plan for AI grounded in those goals. 

Another key finding is the significant gap in data maturity and readiness. While 92% of leaders emphasize the importance of a robust data strategy, only 34% claim to have one implemented across their business. Bridging this gap is crucial to using AI effectively.

Trust also emerges as a primary concern, with respondents citing security issues and skepticism about the output quality of generative AI. Fear of unintentionally exposing private customer data and potential damage to brand reputation looms large as barriers to purchasing generative AI.

Three foundations for strong AI-powered CRM

1. Ready your data

Data quality and availability is the crux of successful AI implementation. Over 50% of survey respondents agree that improved data quality is essential. Despite its importance, they indicated that their top challenges with their organization’s CRM include data quality issues and a lack of data skills. 

On the flip side, companies with a higher degree of data maturity are not only more likely to have adopted AI already but also are more likely to use a unified CRM across their business, thus realizing greater front-office productivity and impact to customer satisfaction.

Our recommendations: Focus on cleaning your data, eliminating silos, and ensuring a holistic view of customer data. Take a balanced approach to data maturity aligned with your strategic goals. Knowing it’s not realistic to improve everything at once, work to understand the specific data requirements needed to deliver your AI use cases in a phased approach. Most importantly, keep data quality and availability front and center as you build a strategy in tandem with AI. 

2. Build trust in AI

Building trust is a must in AI. Almost all respondents (96%) said that trust is important – or even critical –  when partnering with an AI vendor. Genuine concerns about unintentionally exposing private customer data, infringing copyright, or violating data regulatory compliance requirements raise a lot of questions. Specifically, organizations are seeking vendors who already have security protections – such as data masking (the practice of anonymizing sensitive data)  – baked into the tool. Another way to protect from potential risk is to choose a vendor that offers AI as part of their core CRM offering, so there are no extra hoops or complexities with trying to integrate an outside source.  

Our recommendations: Work with a trusted AI vendor that provides meticulous management of both AI inputs and outputs. Your data should be masked when shared with any large language models. When considering a vendor-hosted or external model, ensure that the context of inputs and prompts will not be stored and you never lose control over the use of your data. Beyond privacy and security concerns, prompts and outputs should be automatically scanned for harmful outputs. Finally, depending on the AI use case, consider keeping a human in the loop to ensure quality, accuracy, and trust.

3. Make space for education and upskilling

As AI becomes embedded in organizations, nearly half of survey respondents agreed that continuous upskilling is necessary. This, paired with the general lack of understanding around AI concepts, highlights the opportunity for a thoughtful approach to AI education. Additionally, they noted that a lack of data skills is a primary challenge with the current use of their CRM systems. 

Our recommendations: Beyond ensuring employees have a general understanding of your AI strategy and goals, start tactically with employee training to create and refine prompts. A prompt is detailed instruction provided to a large language model to help it generate an output. A better prompt yields a better, more relevant output from the AI model. Every employee can be a prompt engineer – this training will maximize the potential benefit of AI while helping people work more efficiently. In addition, establishing corporate policies that educate employees to evaluate AI outputs for accuracy, bias, toxicity, and potential harm is essential. 

See how top leaders make the most of their AI investments by creating strong data practices, a culture of continuous learning, and unwavering trust. These insights and more can be found in the full study, so your team can create a foundation of excellence with your own AI-powered CRM.