Kayak co-founder Paul English’s consumer-facing mobile app Lola has proved a hit with investors since launching in 2015, raising $43 million, but his dream to reinvent the travel agent business has smashed into the reality of today’s consumer landscape.

Speaking Wednesday at the Phocuswright Europe event in Amsterdam, English said the company would be pivoting to focus on those that travel frequently such as corporate users.

Version one of Lola appeared in the Apple app stores in early 2016 with the plan to combine the popularity of messaging, knowledge of human travel agents and power of artificial intelligence (AI).

Judging by English’s decision to change the business model, it seems it did not achieve what he wanted.

Instead, a second version, described by English as a “pivot” or “refocusing,” will be made available at the end of the summer and will be “quite different” to what exists today.

One element that will be added to Lola is the ability for the user to have the freedom to book their own trips, or do self-service, and not just rely on the agent to curate their trip.

English said the second lesson he had learnt was that “the people who really love the Lola vision are the people who travel a lot because if you travel twice a year—most Kayak customers do—Lola’s not really that exciting for you but if you’re on the road every week, or every couple of weeks, then Lola gets to know your preferences, and it works unbelievably well.”

Rather than chase after the average leisure traveler, English wants to target small- and medium-size companies that might be better served by Lola’s approach.

English said that Lola currently had 11 business clients based in and around Boston and these were among the “the most rabid” users of the app. It made him realize that there might by avenues to explore in this sector.

“Well TMCs [travel management companies] first of all are not known for their innovation, and they’re not known for building consumer-grade software,” he said.

One of the ways English said Lola could potentially monetize its offering is through a subscription service.

At launch, English stated he would hire 100 travel agents within a year but Lola ended up hiring about a dozen. He said on Twitter that the hiring spree didn’t take place because Lola’s artificial intelligence was kicking in faster than expected.

Photo Credit: The first version of the Lola app. Paul English says the new version out later this year will target more frequent travelers. Skift