Art by Clark Miller
Tactical use cases. Need for change management. Searching for business outcomes. Multiple fundamental challenges. These are the results of the survey of 160 executives conducted by The Information about how enterprises are implementing artificial intelligence, and they paint a picture of typical early days for the adoption of new technology.
A majority of survey respondents (53%) say that to succeed companies must excel at AI.
Today, the gains from AI are mostly in productivity and efficiency, with implementation in areas such as research and development or sales and marketing. The AI landscape offers individual use cases rather than a concerted approach. This article will show where corporations are today in terms of AI and address these questions:
- What is next for AI in an enterprise?
- What’s the right speed of change?
- How do we prepare organizations for this change?
State of Play: Inching Toward Maturity
“AI has the greatest potential to put society at an inflection point that we’ve seen in a long time,” writes one survey respondent. However, we are in the very early stages of AI implementation, 76% of survey respondents believe. Even keeping abreast of developments is a tall order, with less than a third (32%) of corporations considering themselves mature in this area.
Nick Smith, founder and CEO of Sailes,the company that builds unique AI to automate all sales prospecting , puts the current level of maturity in perspective: “We are in infancy in terms of implementation of AI because the length of the runway—the vision and the possibilities—are so beyond what we have ever imagined.”
Just 17% of corporations are mature (mostly or fully) at creating business value from AI and measuring it. “It’s too early to push for business value of AI, when we are still in the very early stages of implementing it into enterprise systems and workflows,” says one survey respondent, noting that AI comes on the heels of other technologies, with companies still trying to figure out business value from cloud.
“With AI, companies have a chicken-and-egg problem,” says Chun Jiang, co-founder and CEO of Monterey AI, which uses large language models to analyze unstructured qualitative data, such as support tickers, transcripts and surveys. The chicken, in her metaphor, is the implementation of an AI-driven tool before it’s clear if this is the right solution (the egg) for the efficiency problem the company is addressing. “Do they really need to look at an AI solution for the sake of it? But if they don’t try it out, they might not find the opportunities to improve,” is how she frames this challenge.
The ramp-up in AI over the next 12 months will not be a case of fools rushing in (see chart, “Inching Toward Maturity”). “Start small, show early success in niche functions, and grow from there,” advises one survey respondent. Forty-two percent of survey respondents say corporations are taking a wait-and-see approach to AI, and just a third say corporations are directing significant investment to enterprise AI.
Current Gains: Efficiency, Productivity and Speed
Are corporations thinking big enough when it comes to AI? Some survey respondents decry the lack of an enterprisewide strategic approach. “There is a gap between the value of tactical point solutions versus enterprisewide strategic solutions,” says one survey respondent. Adds another: “Basically, although there is a lot of ‘bubbling up’ of AI use cases, there is little evidence of top-down efforts to put serious resources into place.”
Nick Smith, founder and CEO of Sailes, agrees that current AI initiatives mostly originate from the bottom up, with “early adopters going in on their own without even getting approval from the company to use AI.” However, he is increasingly seeing large companies with AI strategies that encompass all business units, as well as the emergence of new AI-related departments, such as AI compliance.
The benefits corporations are realizing from AI come mostly in increased efficiency and productivity gains (67%), faster speed of business (52%) and more informed, data-driven decision-making. Less than a third see higher-level gains such as improved data quality management or increased profitability. When looking at AI by function, executives are most satisfied with its implementation in research and development as well as in sales and marketing, but much less so with its applications in human resources or finance (see chart, “Levels of Satisfaction”).
Sailes offers AI-driven tools that pair salespeople with AI bots unique to their personality and style. The company’s platform can accelerate the sales process 16 times. “We have taken the idea of enabling sales and turned it into multiplying sales,” says Smith.
AI tools can drive immediate results. Monterey AI’s AI-driven integration of support tickets, transcripts and surveys, for example, , , can show unbiased recommendations for product improvement instantaneously. Similarly, with an AI tool that parses qualitative responses to an employee survey, Jiang’s Monterey AI can cut the effort needed to analyze these entries, which can number a million, from months to 15 minutes.
But what really matters, Jiang points out, is for the end user, whether a customer or an employee, to see improvement. That means the insights generated by AI tools need to improve operations or services, which requires further cooperation and leadership.
Several survey respondents called out the impact of AI on the workforce and the potential AI has for talent management. The survey shows that only 12% of executives are currently satisfied with the implementation of AI in human resources, and 12% are realizing the benefits of AI for human resources.
“Enterprise AI is going to significantly impact the workforce. The corporations where the HR departments see it as an opportunity for cost optimization through job cuts will lose out to those that see it as an imperative need to upskill their employees for the jobs of tomorrow,” notes one survey respondent. “Finding the fundamental value will require much bigger thinking than we are seeing from enterprises today, at least an order of magnitude beyond the current asks.”
Edward Adjei, co-founder and CEO of Aragorn AI, says his company’s platforms automate the integration of employee data, which can occupy 40 different systems within one company. Aragorn’s technology reduces data integration time from several months to a couple of days, and human resources leaders can communicate with the platform using natural language.
Adjei sees Aragorn as a conduit for HR departments to achieve higher-end goals, like better talent management and recruitment, because it allows HR professionals to spend more time engaging with employees versus handling the databases. Additionally, HR departments can offer optimal benefits since they no longer need to choose vendors based on technology compatibility with existing systems.
Challenges: Undervaluing the Need for Change Management?
Succeeding in AI does not boil down to a couple of obvious or dominant challenges. It will require a well-thought-out, complex and multipronged approach to address multiple, almost equally difficult challenges (see chart, “Challenges to AI Success”).
The biggest group (48%) points to defining the right use cases for AI as the top challenge. At this stage, notes one survey respondent, “enterprise AI permeates efficiency efforts (note taking etc.) and go-to-market strategies (sales, productivity etc.), which are mostly about the present-day problems. We have not seen clear use cases to guide companies on future strategy, financial management and talent areas.”
The use of AI for efficiency may be the table stakes. But these obvious tactical solutions are a safer bet than forcing AI on strategic tasks or processes. “Reverse-engineering use cases for the technology is a recipe for disaster,” warns one survey respondent.
Data quality and management comes in as the most important challenge to get right (44%). Survey respondents believe “data maturity has and will continue to hold enterprises back when it comes to AI.”
The challenge of managing and producing good data is an obstacle for companies that aim to become data driven, but the age of AI is compounding shortcomings in data quality. “AI requires a level of data management and structure that is simply not there in most corporations. A big area of activity will be cleaning up existing data so that they can be used for creating AI models,” one survey respondent expressed.
While change management does not rank as high on the list of challenges, several survey respondents were concerned about the institutional inertia hampering AI rollouts, as well as the old guard’s hesitancy to embrace and understand AI. One survey respondent wrote that to realize the full potential of AI, “you must be willing to invest the time, energy and concentrated resources to launch an effective change management initiative—and greater still to maintain. Few leaders are effectual [enough] change management practitioners to make this last in the meaningful way we all hope.”
For Aragorn AI’s Adjai, the AI solution that people will accept is the one that takes into consideration their workflows. “Ideally, there should be a platform that adapts to [people’s] process, not the other way round,” he says. “Not every organization needs to fit into one specific box.”
Conclusion
We are still in the infancy of AI. The survey of and comments from 160 executives conducted by The Information (see “Methodology”), and interviews with AI technology companies’ founders who work with large corporations, reveal some guidelines about implementing AI initiatives:
Start with the business case, not the technology. Focus on an efficiency, quality or speed issue that AI may solve, and start building the AI solution from there. In this way, treat AI like any other investment in your business growth.
Look for ROI that goes beyond tactical gains. AI can show immediate results in terms of speed and efficiency. In the long term, that may not be enough. AI tools need to play a role in improving the end-user experience, be it that of the customer, employee or vendor.
Encourage new initiatives but provide guardrails. With many AI projects originating from the bottom up, it’s important not to stifle them and wait until a formal top-down AI strategy is put in place. Still, there needs to be awareness of the risks of AI and guardrails around its development.
Methodology: Based on a survey of 160 respondents conducted by The Information in March 2024.
- Size. A majority of survey respondents (60%) came from companies with revenues of up to $100 million; 14% had revenues between $100 million and $1 billion, and 26% had revenues of $1 billion or more.
- Industries. The top industries represented in the survey were technology, media and communications (46%), professional services (19%), financial services (9%), and healthcare and life sciences (6%).
- Function. Top functional areas were general management (32%), technology (19%), marketing and communications (11%), and sales (10%).
- Rank. Forty-five percent of executives were C-level, including 19% who were CEOs, 25% who were directors, 12% who were managers, 12% who were at the vice president and senior vice president level, and 6% who were owners.
- Gender. A majority of respondents were men (77%); 17% were women, and the remaining 6% did not disclose their gender or were transgender women.
- Race or ethnicity. Sixty percent were white Caucasian, 17% Asian, 5% Hispanic or Latino, 4% African American.
- Age. The biggest number of respondents were between 55 and 64 years old (31%), followed by those aged 45 to 54 (25%) and 35 to 44 (17%).
• AI Startup Harvey Targets $2 Billion Valuation and Mulls Buying a Legal Research Company