This January, we’re introducing updates that make it easier for to collaborate with customers and get more reliable results from workbook chats.

These improvements focus on three things teams told us matter most: controlled collaboration, speed, and accuracy at scale.

Collaborate directly with customers in shared workbooks

You can now share Vern workbooks with customers and allow them to edit values in specific columns only.

This makes it easier to validate assumptions, collect missing inputs, and resolve edge cases directly with customers - without exporting files or losing control of the data model.

How it works

• Go to the Company page

• Click the Share icon

• Choose View or Edit

• Select which columns customers are allowed to edit

Everything else remains read-only, so your structure and logic stay protected while customers fill in exactly what’s needed.

Much faster, more reliable workbook chats

We’ve significantly improved how tool calling works inside workbook chats.

Vern now dynamically decides whether to run a deterministic tool or an AI-powered one, based on your prompt and the task at hand.

For common onboarding transformations like:

• column stitching

• date normalization

• calculations

• standard formatting

this reduces execution time from minutes to seconds, while producing more predictable results.

The outcome: faster iterations, fewer retries, and more confidence when working with large datasets.

Reference columns directly with @ mentions

You can now reference columns by typing @ in workbook chats.

This improves both performance and accuracy, and unlocks more advanced workflows across sheets.

With @mentions, Vern can:

• reason more precisely about your data

• perform cross-sheet calculations

• infer relationships between columns

Example use cases

• calculating amounts based on payroll frequency

• applying exchange rates across datasets

• lookups similar to Excel’s VLOOKUP, without manual formulas

This makes complex logic easier to express in plain language - especially for non-technical teams.

What’s next

We’re continuing to invest in making Vern faster, more predictable, and easier to use in real onboarding workflows.

Next up, we’re focused on:

• automated transformations for specific customer source platforms

• clearer explainability when transformations run across large datasets

• expanding deterministic tooling for common onboarding patterns

As always, we’ll keep iterating based on how implementation teams use Vern in production.