Researchers use AI to turn open text feedback into numbers in 2026

In short: In 2026, more research teams are using generative AI to convert open-ended text, like survey comments and interviews, into countable metrics.

What's going on

Qualitative data is the kind that comes in words, not numbers, like written survey answers, interview transcripts, and customer notes. For years, teams had to read it manually and label it by hand, which was slow and hard to scale.

Now large language models, or LLMs (AI systems that work with text), are being used to do much of that labeling automatically. The AI can tag a sentence as a topic like “pricing” or “delivery,” group similar ideas into themes, and estimate sentiment, meaning the emotion in the text, like frustration or happiness. Think of it like sorting thousands of sticky notes into neat piles and then counting each pile.

These tools also turn results into charts and summaries, such as word clouds, tables that compare groups, and slide decks. Audio interviews can be converted to text first using speech-to-text (software that writes down spoken words), which makes it easier to scan and analyze at scale. Some teams report this cuts manual work by about 20 to 30 percent or more, so people can spend more time interpreting what the results mean.

Adoption is rising quickly, with reports suggesting many teams have moved from general AI use to specialized research tools by 2025, and a large share expect AI helpers to handle more than half of projects soon.

What to watch

Human review still matters. AI can be consistent, but it can miss context or nuance, so teams will need clear checks, sample audits, and rules for when a person must step in.

Source: TechCrunch AI