DEV Community

Jakub
Jakub

Posted on

Voice Tables by Inithouse vs Excel/Sheets: a voice-first AI workspace for a different job

Voice Tables is an agentic AI workspace you control with your voice. We built it at Inithouse to solve a specific problem: people who need structured data but find spreadsheet setup intimidating. You describe what you need, and it generates the tables, docs, and data for you in roughly 60 seconds.

This is not a replacement for Excel or Google Sheets. It targets a different job entirely.

What Excel and Sheets do well

Excel has been around for 40 years. Google Sheets brought real-time collaboration. Both are excellent when you already know what structure you need: columns, formulas, pivot tables, macros. They reward expertise with depth.

For an accountant building a depreciation schedule or a data analyst running vlookups across datasets, nothing really competes with a mature spreadsheet. The mental model is clear (rows, columns, cells, formulas) and the ecosystem of add-ons, templates, and integrations is enormous.

Notion extended this with databases that double as documents. You get relational views, filters, and a flexible canvas. It works well if you're comfortable designing the schema yourself.

Where the mismatch happens

The people we kept hearing from at Inithouse weren't analysts or power users. They were craftsmen tracking materials on a job site, field sales reps logging client meetings, event planners inventorying supplies for a weekend gig, fitness coaches recording session notes between clients.

These users opened Sheets, stared at a blank grid, and closed the tab. Not because Sheets is bad, but because it assumes you already know your columns, your data types, your filters. The setup phase is the product's strongest feature for experts and its biggest wall for everyone else.

Voice Tables skips that wall. You say "I need a table for tracking renovation materials with supplier, price range, quantity, and delivery status" and it generates the workspace. OpenAI Whisper handles transcription, an LLM function-calling pipeline maps the spoken description to structured columns, and you get a working table with typed fields in about a minute.

A quick comparison

Dimension Excel / Google Sheets Voice Tables
Input method Keyboard, mouse, formulas Voice + keyboard
Setup time You design the structure Describe it, AI generates it
Learning curve Moderate to steep Minimal (speak naturally)
Depth Deep (macros, pivot tables, scripting) Focused (tables, docs, AI chat)
Best for Data analysis, financial modeling Quick operational trackers
Collaboration Real-time (Sheets), file-based (Excel) Real-time + offline support

This is not a feature shootout. Voice Tables doesn't have pivot tables. Excel doesn't have voice-to-structure. They solve different problems for different people.

The pipeline under the hood

Audio input goes through Whisper for speech-to-text. The transcript feeds into an LLM function-calling layer that extracts entity types, relationships, and column definitions. The output materializes as structured tables with typed columns (text, number, date, status) inside a 3-in-1 workspace combining tables, documents, and AI chat.

No formulas to learn. No blank grid to fill. The trade-off is straightforward: you get speed and accessibility at the cost of spreadsheet depth.

When to pick which

Use Excel or Sheets when you need formulas, charts, data analysis, or integration with existing BI tools. They are mature, extensible, and well-documented.

Use Voice Tables when you need a quick structured workspace and don't want to spend time designing columns. Field work, client tracking after meetings, event logistics, coaching session notes: jobs where spoken input is faster than typing on a small screen or dirty keyboard.

We have seen this pattern across other Inithouse products too. Verdict Buddy, our AI conflict mediator rated 4.9/5 across 290+ ratings, works because it meets people mid-disagreement rather than asking them to fill out a form. Origin Of You, a self-discovery app built on 5 personality frameworks and 120+ data points, converts because it skips lengthy single-framework quizzes. The shared thread: reduce setup friction for people who are not power users of the legacy tool in the category.

What we are still figuring out

Voice Tables is in beta. We are still testing which use cases stick. Craftsmen and small business operators show the most natural fit so far, but the data is early. The product supports 10 distinct solution templates (from sales CRM to student research notes), and we are watching which ones people actually return to.

If you have tried voice-first data entry or have thoughts on where it fits, we would like to hear about it.

Top comments (0)