The artificial intelligence of large language models is revolutionizing business

 When an expert gets an extra brain from artificial intelligence for ideation and texts, it may mean a revolution in the business environment. Technology application agency Evolver and consulting agency Ellun Kanat tested how to take the artificial intelligence leap at an explosive speed.

The importance of language model-based artificial intelligence has grown rapidly in recent years. Consulting firm Ellun Kanat likes to look boldly into the future, so it was clear that even with artificial intelligence taking rapid development leaps recently, they needed to explore its possibilities for business proactively. The use of artificial intelligence in consulting, and especially in communication, was tried out in a collaborative project with Evolver. 

Ville Tuominen, Business Director of Ellun Kanat responsible for product development, says that the company has long thought that artificial intelligence and automation will shake up production. The human role on the creative side was thought to be protected. However, rapidly developing artificial intelligence in text and image creation proved the assumption to be incorrect.

Together with Evolver, we started thinking about whether artificial intelligence-based text tools can help, especially in the production of texts that require creativity. The experts were put to the test.

“We have recognized that artificial intelligence will change work and how work is done. In this experiment, we wanted to find out its capabilities of in the context of challenging work,” explains Tuominen.

The Jasper.AI text tool was put to the test. The experts used artificial intelligence to support their own work, for example in articles, blog posts, clean writing of memos, and ideation of event concepts and names.

“In the fall of 2022, Jasper.AI was the best-known tool for writing copy on the market and a team of three people started testing it. We held three two-week sprints, each with its own topic. At the end of the sprints, we gathered to reflect on the results,” explains Evolver CEO Arttu Kataja, who was responsible for facilitating the project.

Finally, the results were drawn together in a workshop where the experts could give direct feedback on the experiment.

“There was frustration with many things that made it difficult and even prevented the proper use of artificial intelligence in expert work. The biggest challenges were related to the accuracy of the content and the lack of source information, which are essential in expert work,” Kataja describes and continues: “There were also flaws in data protection that shouldn’t exist when processing confidential information. So the text had to be worked on quite a lot before it was given to the artificial intelligence to process.”

Kataja emphasizes that despite the shortcomings, there were also many opportunities apparent: “Artificial intelligence does not replace human creativity as a whole, but it is a very useful tool for ideation and the production of basic texts.”

From trial to product – in less than two months

Artificial intelligence proved to be effective for producing general copy, as long as a human is involved in checking the content and editing the text. Artificial intelligence is not as directly suitable for expert work as it is for supporting work and duplicating a finished text.

“The better the text is reproducible, the better the artificial intelligence works. If the customer often needs the same type of text with a certain tone of voice and an unchanged core message, then artificial intelligence is very useful. It’s a different matter if artificial intelligence always has to guide you to a new need,” explains Tuominen.

Production started immediately based on the findings. Based on the experiment, a communication strategy facilitation service compatible with artificial intelligence was launched.

“We first produced an artificial intelligence-compatible communication strategy. The key was to understand at which points the artificial intelligence changes the process and how the choices made are clearly communicated to the artificial intelligence,” Tuominen says and continues:

“In practice, we have to think about instructions that can enable the production of simple texts pretty close to the right one, or how to make different versions of an existing good text.”

The timing couldn’t have been better. At the same time as the hugely popular ChatGPT language program was released, Ellun Kanat already had product development in full swing.

“It’s crazy that ChatGPT was released at the end of November and Ellun Kanat has a product out at the turn of the year. It shows how fast things are happening.”, says Kataja.

“In product development, we often start with the basic product and fixing the related problems. Here, we started thinking about the solution ahead of time, wondering what question it answers. After the experiments, the intended use was found, and I myself am satisfied with how quickly we were able to launch the product,” Tuominen reflects.

Kataja also admires how open-minded employees were to experiment with the possibilities brought by the new tool: “At Ellun Kanat, there is a genuine curiosity and understanding that communication professionals have to integrate new technology into their own work. They don’t see artificial intelligence as a threat to their jobs, but as an opportunity for development.”

The sharing of ideas is also encouraged through a dedicated discussion channel. Tuominen believes that the interest in utilizing artificial intelligence is already a win: “We don’t know the business benefits yet, but the product has attracted a lot of interest. Above all, it has opened up a discussion among the staff about which direction we are going.”

“I see that our customer organizations benefit the most from this and our task is to think about how we can help our customers join this fast-moving train,” Tuominen adds.

Now is the time to wake up to the possibilities of artificial intelligence

ChatGPT put the possibilities of big language models on everyone’s lips at the end of 2022. However, Kataja estimates that many are still not prepared for the accelerating pace of development with which artificial intelligence will change business models:

“Few business developers are aware of the speed at which artificial intelligence is developing. Something new happens every week! If you’re not inside the community, it can be difficult to notice those market signals, even though ChatGPT made it visible.”

The use of artificial intelligence will become more common and the field of language model-based artificial intelligence will see a huge leap in the next 12 months, Tuominen assures: “There probably aren’t many fields where you can be lulled into thinking that this doesn’t apply to us. Everyone should at least think about the effects.”

Tuominen estimates that in the near future artificial intelligence will have a significant impact on business if it can be utilized in a sensible way. He is particularly interested in seeing if artificial intelligence will change the usual ways of doing things:

“I would hope that as the use of artificial intelligence becomes more common, we would get rid of old ways of doing things and start working more rationally with new tools.”

In digitization, people often simply transfer old way of doing things to a digital platform, which doesn’t allow for the exploitation of the hidden possibilities offered by the new tools.

“It’s worth considering in advance how the game will change when a new player comes. From a business point of view, it is also interesting how we think about working methods within the constraints of new tools. If this can be done, then the revolution is close”, Tuominen smiles.