A market question never arrives with one clean dataset. The numbers sit in filings, trade association PDFs, pricing pages, news and last quarter's export, and every source defines the market a little differently. Rixx researches before it answers, cites every fact from the web inline, and keeps your uploaded files referenced to the page. The question every stakeholder asks first, "where does this figure come from?", already has an answer.
Ask anything: definitions, figures and quick checks
Ask a factual question and get a researched, cited answer in seconds. Rixx is especially useful for the small checks that interrupt modelling: the latest reported revenue for a company, the official definition of a segment, the date a regulation takes effect. Each answer links to its sources so you can confirm the definition matches your model. The AI research assistant page explains how answers are built.
Deep research: a market brief from many sources
For a new category or a quarterly refresh, give Rixx a deep research job. It reads across filings, association reports, company pages and news, and writes a structured brief: definition, size and growth estimates side by side, segments, drivers, key players, risks and open questions. Use the AI market research generator for market briefs and the AI report generator for any other report shape. For a framework view, AI SWOT analysis builds a cited SWOT per company.
Wide research: the whole competitive set, one table
Landscapes are where breadth matters. With wide research, Rixx splits the list across parallel agents, one per company, and merges the results into a single table with consistent columns. Compare 12 vendors on pricing tiers and target segments, profile 30 regional players, or track the last year of announcements across a category. Every row carries its own sources. See AI competitor analysis for the matrix format.
Analyse your data and documents: Python on your exports
Upload a CSV or Excel export, such as sales by region, survey responses or panel data, and describe the question. Rixx runs Python and other tools in its own private workspace: it cleans the data, computes the cuts, tests the patterns you ask about and charts the result, with the working visible. See AI data analysis.
Long reports work the same way. Upload industry studies and filings (PDFs up to 1,000 pages, plus Word, PowerPoint and images) and ask for every figure, definition and caveat you need, each with a page reference. For a stack of reports, summarize multiple PDFs pulls them into one comparison, and AI document analysis covers the rest.
Make the file: models, briefs and charts that travel
Rixx builds the deliverable:
- Excel with working formulas, recalculated and checked, assumptions sourced on their own tab. See the AI spreadsheet generator and the AI financial model page.
- Word and PDF briefs with sources kept inside.
- Editable PowerPoint with native charts when the analysis goes to a meeting. See the AI presentation maker.
Worked example: a defensible range, not one number
The job: "Size the market for AI meeting-notes software for small and mid-size businesses in North America. I need a defensible range. Give me an Excel model, a two-page brief and a chart for the deck."
Rixx may ask one quick multiple-choice question, such as how to define a small business. Then it runs as a job in the background and returns:
| Output | What is in it |
|---|---|
| Model (.xlsx) | Top-down from published figures, bottom-up from business counts, adoption and pricing, live formulas |
| Assumptions tab | Every input with its source and a note on how it was used |
| Brief (.docx) | The range, the main swing factors, and where sources disagree on definitions |
| Chart | Low, base and high cases as an editable chart |
You challenge the assumptions you know better than any public source, then present it.
Reconciling sources: a method that works
Published market figures often disagree because they measure different things. Ask Rixx to put every estimate in one table with these columns:
| Column | Why it matters |
|---|---|
| Value and year | Base years vary between publishers |
| Geography | Regional and global figures get mixed up |
| Scope | Hardware, software and services are counted differently |
| Currency basis | Nominal and constant figures diverge over time |
| Source | So you can open the original and check |
Once the definitions sit side by side, choosing a base case becomes a deliberate decision you can explain.
Which plan fits analysts
Every plan, including Free, gets every feature. Analysts who model and research most days usually choose Pro at $20/month ($200/year; ₹1,799 in India), with 200 credits a day and 20,000 a month. Lighter users start on Plus at $8/month (Starter ₹399 in India). A quick cited answer uses a few credits; a deep report or a job that builds files uses more. See pricing.
Frequently asked questions
Can Rixx build a market sizing model?
Yes. Rixx builds Excel models with working formulas, top-down and bottom-up estimates, and a sourced assumptions tab. It recalculates and checks the workbook before handing it back.
Can Rixx analyse my own data?
Yes. Upload CSV or Excel files and Rixx uses its own private workspace, with a full terminal and Python, to clean, compute and chart, returning the tables, charts and working.
How does Rixx handle sources that disagree?
It lists each estimate with its year, geography, scope and source, and explains where definitions differ, so you can choose a base case on purpose.
Can Rixx compare many companies in one go?
Yes. Wide research runs one agent per company in parallel and merges the results into one table with consistent columns and sources per row.
Can I upload long industry reports?
Yes. PDFs up to 1,000 pages work, along with Word, Excel, PowerPoint, CSV and images. Answers reference the document and page.
Where do citations appear in the files?
Briefs keep sources inside the document, and models keep a source next to each assumption, so anyone reviewing the file can trace a figure.