Excelmatic

Excelmatic (the name RowSpeak is also used on the page) is an AI data analysis and reporting tool for spreadsheets and business data, positioned to let users converse with Excel and other data files through natural language to complete analysis, tabulation, and reporting. The following is organized based on the information explicitly stated in the materials.

I. Purpose Excelmatic is used to analyze messy Excel, CSV, PDF tables, and image tables, helping to identify trends, explain data changes, and generate charts, dashboards, and reviewable reports. Its core process is: start with existing files → ask questions in plain language → review and share results. The materials emphasize that results are reviewable and assumptions can be adjusted.

II. Main Capabilities

1. Data source support: can upload spreadsheets, CSV files, PDF tables, screenshots/image tables (supporting PDF, JPG, PNG, and scanned documents), as well as files exported from business systems and files on shared drives, inboxes, or desktops. The materials state that the initial process does not require a data warehouse or real-time data connections.
2. Image and PDF table extraction: can accurately extract nested tables, receipts, and financial statements from images and PDFs, with intelligent column alignment, and supports one-click export to Excel.
3. Multi-table merging: can perform multi-table joins by matching keys. In the example in the materials, Book ID is used as the matching key to look up a book number reference table in sales order details and fill in two blank fields.
4. Analysis and explanation: can compare different periods, explain the reasons for changes, discover abnormal fluctuations, and explore numbers in natural language.
5. Dashboards and charts: can generate dashboards with AI in seconds. The example dashboards listed in the materials include a monthly financial dashboard (revenue, expenses, cash flow, profit margin, budget variance), a sales performance dashboard (revenue, pipeline, conversion, best-selling products, regional performance), an inventory operations dashboard (inventory levels, fast-moving items, low-stock risk, inventory changes), and a marketing performance dashboard (spend, conversions, ROAS, CAC, channel performance).
6. Regular business reports and automation: can save questions, metrics, and validation rules once; when new spreadsheets or business export files arrive, automatically recalculate variances, perform checks, and compile a management summary ready for reporting. The materials say that updated files can be placed in the same folder or sent by email, and RowSpeak will automatically rerun all formulas, charts, and review checks.
7. Report detail example: the materials show an operating cost control dashboard example that can switch by store, compare multiple years, and recalculate instantly, including revenue, labor cost ratio, fixed salaries, and social insurance, six core metric cards including all-risk service fees, as well as cost share charts, expense rankings, and fiscal year profit margin trends. Another example gives a monthly profit and loss variance analysis conclusion: from January to August 2025, total revenue was $2,862,500, total expenses were -$2,020,600, net profit/gross profit was $841,900 (profit margin 29.4%), and actual overall was $152,900 below budget, of which revenue exceeded budget by $74,500, but expenses exceeded budget by $227,400.

III. How to Use Upload existing files (Excel, CSV, PDF tables, screenshots, or business export reports), ask questions in natural language, then review the results and share them. The materials explain that key assumptions can be revised, such as revenue fields, date ranges, grouping dimensions, refund handling methods, or chart types, without rebuilding the entire analysis. It also supports checking assumptions, adjusting visualizations, and refining reports before sharing.

IV. Applicable Scenarios and Audiences Applicable to finance and accounting (monthly analysis, variance explanations, cash flow review, management reporting), sales and operations (pipeline, revenue, fulfillment, and periodic operations reports), marketing and growth (campaign performance, conversion trends, ROAS, CAC, channel performance), inventory and inventory changes (fast-moving consumer goods, inventory changes, abnormal changes, and potential inventory risks), management reporting, as well as individual analysts, business owners, and operators. The materials state that it is aimed at individual analysts, finance professionals, operations personnel, business owners, and teams handling periodic business reports.

V. Limitations and Information Not Yet Provided The materials clearly state that direct connectors to tools such as Google Analytics, Google Sheets, and databases are not yet provided, and such connections are "coming soon"; at this stage, the same analysis and reporting process can be completed by uploading export files from these tools. The materials do not provide specific pricing, fee standards, free quotas, supported interface languages, specific model names (local deployment mentions being based on open-source large models), and other information.

VI. Data Privacy and Security The materials state that users' files and conversations are private, encrypted, and isolated; business data will not be used to train AI models; uploaded data sources and conversations remain within the workspace and are not used to train or improve its general AI models; files are encrypted during upload, processing, and storage, and access is limited to authorized users and services; files, conversations, and analysis results are isolated from other users and organizations, and workspace access is controlled by the organization to which they belong. For organizations that require complete data privacy, RowSpeak can be deployed locally within their own private network and run based on open-source large models, so that data and AI workflows are under the organization's control, and they need to be contacted to discuss deployment requirements.

VII. Summary of User Feedback User feedback cited in the materials mentions: monthly variance reports that used to take half a day to prepare can now produce a reviewable first draft in a few minutes; data exported weekly from three advertising platforms can be aggregated into one view and changes can be seen, and assumptions can be checked before submitting to management; during forecast reviews, CRM export files can be used directly, and the sources of pipeline changes and specific deals can be shown; when handling monthly variances for multiple small clients, the same set of review rules can be reused, using time to explain the numbers rather than rebuild spreadsheets; users believe that what is valuable is not only getting answers, but also being able to see assumptions, revise assumptions, and share results with the team.