By Nina Trentmann

Nov. 19, 2019 2:56 pm ET

After a year of deal-making and deal-avoiding—including seven acquisitions and the rejection of takeover bids—executives at Smurfit Kappa Group PLC thought it would be good to gauge how well they had communicated the company’s strategy and its story along the way.

The Dublin-based company commissioned an investor-perception study from consulting firm Rivel Research Group, which polled dozens of shareholders and analysts. The results were an eye-opener for executives, who had assumed Smurfit Kappa—Europe’s largest producer of paper-based packaging—would be viewed as focused on growth.

“People thought of us as a leveraged vehicle and felt that our approach was much more about debt reduction than about investing,” said Ken Bowles, Smurfit Kappa’s chief financial officer. “It didn’t fit with where we were and what we were doing.”

The company pored over its corporate communications—annual reports, presentations, news releases, earnings-call transcripts—and realized that executives’ talk of efforts to reduce the company’s €3.1 billion ($3.4 billion) debt pile at the end of 2018 was what actually resonated with the outside world. Not the acquisitions.

“We misfired with our communications,” Mr. Bowles said. The company has since revised its corporate literature to address the mismatch and now chooses its words more carefully, highlighting its focus on sustainable growth instead, he said.

Smurfit Kappa is part of a growing group of companies turning to external assessments to fine-tune their communications. The target audience isn’t just human analysts and investors: Machine-learning tools, perception studies and speech recognition are also used by finance and investor-relations executives seeking to anticipate the reaction of robots that increasingly sway the market by buying and selling shares on the basis of data analysis.

Meanwhile, campaigning from activist investors, a stronger focus on environmental, social and corporate governance and changes to financial regulation also add to the need to get the message across.

“We really have to change the way we are writing and scripting,” said Hala Elsherbini, senior vice president at Halliburton Investor Relations & Communications, a Richardson, Texas-based investor relations and communications agency unrelated to oil-field services company Halliburton Co.

These days, a lot of investors are using language- and sentiment-analysis tools, which means companies have to adjust how they talk about their finances, said Ms. Elsherbini, who said she is working with about 20 U.S.-based companies to help them communicate more effectively.

Algorithmic traders using sentiment analysis—an automated process that identifies positive, negative and neutral opinions in a body of text—can react negatively to wording such as “contrary to,” Ms. Elsherbini said, adding that she prefers the more positive-sounding “different to” when drafting a company release or an earnings call script. Similarly, she avoids frequent use of terms like “shortfall” or “decline” and focuses on just numbers instead.

“Don’t overstate the negative,” she said.

Other companies are pitching data-crunching software tools that they say can find signals in companies’ public statements.

Prattle, a St. Louis-based text analytics company, uses machine learning and natural language processing to create a lexicon of a company’s commonly used words. This can help institutional investors get a better understanding of a company’s tone and the potential implications. 

Carlos Brito, chief executive of Anheuser-Busch InBev, poses with a beer after a news conference. The company now focuses on more face-to-face interactions between executives and investors. PHOTO: FRANCOIS LENOIR/REUTERS

Most of Prattle’s revenue comes from quantitative hedge funds that subscribe to its services, said Chief Executive Evan Schnidman. The company—which was founded in 2014 and acquired this year by investment network Liquidnet Holdings Inc.—also has a tool that corporate investor-relations teams can use to test the potential market impact of their company’s language, he said.

Rival software firm Amenity Analytics, based in New York, has come up with a “deception score” tool that warns investors when companies may be using evasive language, euphemisms or stalling tactics, according to CEO Nathaniel Storch. “As an investor, you are always trying to understand something that other people are missing,” said Mr. Storch, whose company has more than doubled its customer count over the past year, according to a spokeswoman.

Big companies aren’t all using these technologies to improve their communications. Brewing giant Anheuser-Busch InBev SA, for instance, recently gathered information on language-recognition technology to get a better understanding of what is possible, but so far isn’t using it, said Lauren Abbott, global vice president for investor relations at AB InBev.

Still, AB InBev over the past year has adjusted its communication strategy following an investor-perception study similar to the one commissioned by Smurfit Kappa. The company now focuses on more face-to-face interactions between executives and investors and uses its earnings release to provide a more strategic update instead of regurgitating numbers, Ms. Abbott said.

Brian Rivel, chairman and CEO of Rivel Research, which conducts nearly 170 investor perception surveys a year, said the firm’s revenue from that part of its business has roughly doubled to $8 million this year from $4 million in 2009. Behind the increased interest, he said, is his clients’ belief that calibrating their communications can boost their shares.

“There are two things that motivate people: fear of activism, and opportunism,” Mr. Rivel said. “CEOs want their companies to have a higher stock price.”

Write to Nina Trentmann at Nina.Trentmann@wsj.com