Do qualitative thematic analysis & reporting with this ChatGPT -based tool (AILYZE)

A new AI tool built specifically for qualitative thematic analysis is gaining attention for turning uploaded interview transcripts into structured themes, code-book style outputs, and—crucially—evidence in the form of quotes. The platform, called AI lies (a wordplay on “analyze” and “AI”), aims to reduce the workarounds researchers previously had to invent when using general-purpose chat systems not designed for qualitative coding.

The workflow starts with uploading a small set of documents. On the free “light” plan, each query accepts up to three files, which may require chunking for larger studies. The interface also includes guidance on what to ask and what to avoid; for example, it discourages prompting for numerical answers, reflecting that the tool is optimized for qualitative interpretation rather than quantitative tasks. Early use comes with minor glitches, but ongoing improvements are described by the person responsible for the platform.

To test the tool, the creator uses a hypothetical leadership study and generates three interview transcripts via ChatGPT to ensure the analysis begins without prior knowledge of the content. After upload, the tool offers multiple modes: file summarization, answering targeted questions, and running thematic analysis. Summaries can be returned as bullet points or essay-style text. Targeted questions can be asked, though some prompts may produce unhelpful or “silly” results when they don’t map cleanly onto the study’s context—such as a question about whether participants are “happy,” which yields difficulty determining that attribute while still returning content-based summaries.

The core feature is thematic analysis. When requested in essay format, the tool generates a set of themes (nine in the example) and places supporting quotes under each theme. The output is designed to be more than a high-level narrative: it provides evidence tied to the underlying interviews, which the creator emphasizes as essential for qualitative rigor. Additional instructions can be layered on top of the initial analysis, allowing users to refine how themes are framed or expanded.

The tool also supports cross-document comparisons. In the example, the analysis is extended to compare challenges reported by male versus female leaders. To make that comparison, the tool needs participant attributes (e.g., marking which participants are male or female). The resulting comparison avoids overclaiming: it reports no clear indication that female leaders face more challenges and notes that both groups use similar strategies. When the user requests quotes for the comparison, the tool supplies document extracts, reinforcing the claim with direct textual evidence.

Despite the strong utility, the workflow is framed as assistive rather than fully automated. The creator stresses that AI output should not replace researcher judgment or the full analytic process; instead, it can accelerate coding, theme drafting, and evidence gathering—especially the quote-and-extract functionality that many general chat tools struggle to deliver reliably.