Zoho DataPrep 3.0: Agentic AI, Code Studio and 10+ new connectors
Building a data pipeline has always meant spelling out every step. Import this file. Remove those duplicates. Join these two tables on that column. Push the result there. The tool did what you told it, and only what you told it.
The new Zoho DataPrep launch today changes all that. With Agentic Ask Zia, you describe the outcome you want, and Zia plans the steps, builds the pipeline and runs it. Every step stays visible and editable, and nothing reaches your destination until you approve it.
The release also brings a rebuilt processing engine, Python transforms in Code Studio, 10+ new connectors, and a way to run Zoho DataPrep from your own AI agents.
Watch this short video to see everything we’ve built into Zoho DataPrep 3.0:
Agentic Ask Zia: From instructions to goals
Ask Zia is now an AI agent for data preparation. Instead of giving it instructions one at a time, you give it a goal. Here’s what that looks like in practice.
1. Describe the outcome - Say you want to import event leads from Zoho Creator and company data from PostgreSQL, enrich the leads, and push them to Zoho CRM. You type that in plain words.
2. Review the plan - Ask Zia breaks the goal into tasks: import, profile, fix, join and push. You can edit the plan before Zia starts, or let it go ahead.
3. See the problems and the fixes - Ask Zia spots duplicates, messy addresses and junk records. It then lists all corrections in a preview. You decide which to keep and which to discard, then Zia applies your choices.
4. Enrich the data - Ask Zia joins related datasets and adds the missing details, so each record is more complete.
5. Confirm the destination - Ask Zia suggests sending the enriched leads to the Leads module in Zoho CRM. Nothing is added until you confirm.
The result is one prompt and one finished pipeline, with every step reviewed by you.
A team of subagents
Behind Ask Zia is a set of specialized agents that work like a small data engineering team.
Import, Transform and Export agents
The data preparation agent splits the work across three subagents. The Import agent brings data in from your sources. The Transform agent cleans and reshapes it. The Export agent sends it to its destination. Each one builds its stage of the pipeline end to end and saves its work as editable rules.
Data Quality agent
This agent looks at the quality of your data and suggests the transforms that would fix it.
Built so you stay in control
An agent that builds pipelines on its own is only useful if you can trust what it builds. Four features make that possible.
Editable rules
Every step Ask Zia adds becomes a rule you can review and change. You can open any transform, see exactly what it does and adjust it. There are no black-box steps.
Agent memory
Ask Zia remembers your preferences and the terms you use, and carries them across conversations. The more you use it, the better it fits how you work.
Smart suggestions
Ask Zia spots data quality issues, like missing values, invalid entries and inconsistent formats, and suggests the fixes to apply. You decide which ones to accept.
Anomaly detection
Outliers and unexpected values are flagged while your data is still in DataPrep, before they reach your reports and throw off the numbers.
The dForge engine: faster and lighter
Zoho DataPrep’s processing engine has been rebuilt from the ground up. The new dForge engine applies every transform faster, and works on a sample large enough to trust when you preview changes.
100x faster previews, so you see the effect of a transform almost instantly
70% faster transforms when pipelines run
50% less storage used
For teams moving large volumes of data on a schedule, those gains add up to shorter runs and lower overhead.
Custom transforms with Code Studio
Zoho DataPrep has 400+ built-in transforms. When you need something specific to your business, Code Studio lets you write your own transformations in Python. You test the script on sample data first, then deploy it to run across the full dataset. Data Engineers who know Python get full flexibility without leaving DataPrep, and the transform runs inside the same pipeline as everything else.
10+ new connectors
Zoho DataPrep adds more than 10 new connectors for importing and exporting data. You can bring data in from more of the sources your business relies on and send it wherever your teams need it, whether that’s a data warehouse, an analytics tool or a business app like Zoho CRM.
New Import Connectors: Books, Inventory, Zendesk, Woocommerce
New Export Connectors: SDP Cloud, SDP MSP Cloud, SDP Asset Explorer Cloud
Explore import sources | Explore export destinations
Headless DataPrep
Not every pipeline needs to start in the Zoho DataPrep interface. With the Zoho DataPrep MCP server, you can create, author and run data pipelines from your preferred AI agent or assistant. MCP endpoints can build end-to-end flows, or plug external agents into your pipelines. These agent capabilities are now available in Zoho Analytics data pipelines too. For developers and data teams, that means data preparation can become one step in a larger automated workflow, without anyone opening the app.
And 20+ more updates
Zoho DataPrep includes more than 20 additional updates across connectors, transforms and performance. Learn more
See it live
Join our product team for a walkthrough of what’s new in Zoho DataPrep at the launch event on October 29, 2026. Join the launch event
Set the goal and let Ask Zia move your data. Try Zoho DataPrep free for 15 days. No credit card required.
Comments