Editorial policy and corrections

Who writes the Depra AI site?

Zia Wasi, co-founder of Depra AI, leads the research and the guides. He is a GEO researcher. GEO, or generative engine optimization, is the work of getting a brand named and cited in AI answers. He studies how ChatGPT, Gemini, Google AI Overviews and Perplexity choose and cite brands, with a focus on India.

Zia Wasi leads the Hinglish vs English AI shopping study and its October 2026 re-run, and the open IndicGEO repository that holds its data. He also collected Depra AI's research on AI models and the search indexes behind them, from public sources. Guides and comparisons carry his byline. The study page is published by Depra AI and names him as the lead of the research here and on his author page. Product, help and company pages, like this one, carry no byline and speak for the Depra AI team.

How are facts about other tools checked?

Every fact about another tool is read on that vendor's own public pages, usually its pricing page and homepage. We do not take vendor facts from review sites, other comparison pages or AI answers.

• Dated stamps: each vendor fact carries the date we read it, in the form "checked 13 Oct 2026", with a link to the page it came from. Before a vendor fact changes on our site, we read the vendor's page again.
• "Not published": when a vendor's own pages do not say something, such as which engines a plan includes, our tables say "Not published". We do not guess and we do not fill the gap from elsewhere.
• When a vendor's pages disagree: we say that they disagree and give both readings, each with its source.
• A disclosure on comparisons: our comparison pages say that Depra AI publishes them. Our comparison table and the Depra AI alternatives page also list what Depra AI does not offer today.
• Facts only: we describe other tools by what their own pages say or do not say. We do not rate them, and we do not comment on their design or their funding.

How is Depra AI research run and published?

Each Depra AI study is designed before any data is collected, and its raw data is published so anyone can check our numbers.

• Design set first: the prompts are frozen and the statistical tests are chosen before collection starts. Any change after that goes into a public deviations log, with the reason.
• Raw data public: the August 2026 raw answers, the prompt set, the brand lexicon and the analysis code are on IndicGEO, Depra AI's open research repository. The October 2026 re-run files are being added there as rerun-2026-10.
• Independent recomputation: before a number is published, separate automated checks (AI agents working from the raw data, without our analysis code) recompute it with their own code. For the October 2026 re-run, three checks recomputed the headline numbers from the 479 raw answers and reproduced them. They flagged two definitions, which the study page now states precisely.
• Interest stated: Depra AI funds its research and has a commercial interest in the results. Each study says so, and reports its weak findings as weak.

How does Depra AI use AI in writing?

Facts about Depra AI come from our own product records and price list. Facts about anyone else come from a source we opened, linked in the sentence with the date we read it. An AI tool's draft is never the source for a fact. Study numbers are counts of AI answers that we collected and published as data.

How do I request a correction?

Email team@depra.ai. Name the page, quote the sentence, and link to the source that shows the right fact. A vendor can write about its own product the same way, with a link to its own page.

The team replies within an hour during working hours in India (IST). If the source holds, we correct the page, move its updated date, and add a line to the log below. If we disagree, we reply with our source.

Corrections log

Factual corrections to published Depra AI pages are listed here, oldest first, from October 2026.