Timextender AI early access¶

Thank you for joining our early access program!

As part of this Early Access program, you can try our new AI capabilities that make your data platform more intelligent, more governed, and much easier to work with.

We will reach out soon to schedule an onboarding call with you and your team.

This short guide explains what you have access to, why it matters, and suggested first tests for partners and customers.

1. Timextender MCP Server – Natural language over governed models

The Timextender MCP Database Server lets AI agents (like Claude or local tools via MCP) query your data through Timextender’s semantic layer, using business terms instead of SQL. It connects to your Deliver instances and exposes governed semantic models over stdio or secure HTTP, with automatic detection and reload when you deploy updated instances.

This means AI can answer questions like “Show Q4 revenue by region” using your approved logic, relationships, and metrics, while respecting domain isolation, read‑only safeguards, and audit logging.

Please visit Timextender MCP Server Overview for detailed descriptions.

How to start:

• This feature requires our newest version of Timextender Data Integration, released in January 2026.
• Within TDI, you will need to create a new Endpoint of type: Timextender MCP Server
• You can download the newest version of Timextender MCP Server here, and follow the documentation.

What to test:

• Connect MCP Server to a Deliver instance and verify that business terms from the semantic layer appear in the AI tool.
• Ask the AI agent a few typical business questions (e.g. revenue by region, customer segments, product performance) and confirm the results match your existing reports.
• If you have multiple domains (e.g. Sales, Finance), spin up separate services and confirm each only exposes its own semantic layer.

Value to highlight with customers/partners:

• Semantic‑first: AI uses governed business definitions, not ad‑hoc SQL.
• Production‑ready: zero‑downtime semantic updates, audit logs, and read‑only enforcement.
• Future‑proof: built for SQL Server today, designed to support more backends like Snowflake and Fabric.

2. Xpilot for Data Quality – Generate rules with AI

Xpilot for Data Quality helps you define and improve data quality by suggesting rules based on your datasets and instructions, and then refining them interactively. It works directly in Timextender O&DQ, with support for “simple” and “enhanced analysis” modes.​

Internal testing shows Xpilot can propose useful rules, understand dataset descriptions (especially with enhanced analysis), and even work with instructions written in local languages, though you may still see some off‑target or duplicate suggestions in certain scenarios. ​

How to start:

• Your service will need to be upgraded to Timextender O&DQ 26.1. Please contact support@timextender.com and specify what service to upgrade.
• Follow the Generate Data Quality Rules with Xpilot.

What to test:

• Create a dataset with a clear description and click Generate Rules.
• Try both simple and enhanced analysis and compare how relevant the suggestions are.
• Modify the instructions (e.g. focus on completeness, uniqueness, valid codes, etc.) and regenerate rules to see how Xpilot adapts.
• Preview rules and validate that the generated SQL and logic match your expectations. If you see issues, use “Ask Xpilot to fix the error” and iterate.

Value to highlight with customers/partners:

• Speed: drastically reduces time to a first set of useful data quality checks.
• Guidance: gives less technical users a way to create and refine rules without starting from a blank page.
• Real‑world fit: supports large tables and complex scenarios, with active improvements based on internal and partner feedback.

3. Orchestration Error Insights – AI help for failed runs

Orchestration Error Insights aims to make it much easier to understand and fix package failures in Orchestrator. Instead of long, noisy logs, you get a cleaned‑up error view, plus AI suggestions powered by Xpilot.

The vision for v0 is: when an orchestration package fails, you see separated error and full logs, the package configuration, and AI guidance on likely causes and next steps for script‑based packages like DataTransfer, PowerShell, and CommandLine.

How to start:

• Your service will need to be upgraded to Timextender O&DQ 26.1. Please contact support@timextender.com and specify what service to upgrade.
• Follow the article Orchestration Xpilot Error Insights.

What to test:

• Trigger a few controlled failures in test environments and open the error in Orchestrator.
• Compare the Error log vs Full log to see if the filtered view makes the problem easier to understand.
• Review Xpilot’s suggestions: do they help you diagnose and fix the problem faster than reading the raw logs alone?

Value to highlight with customers/partners:

• Faster troubleshooting: less time digging through logs, more time fixing issues.
• Self‑service: helps engineers solve common issues without always escalating to support.
• Consistent experience: log‑based help in the UI today, with potential for aligned email notifications and deep links later.

4. How we’ll work with you As an Early Access participant, you can start testing these features right away, on your own schedule. ​

Our team will also reach out directly to:​

• Help with onboarding and setup where needed.
• Collect feedback on what works, what’s confusing, and what is missing for production use.
• Prioritize improvements and roadmap items together with you.

If you want to go deeper on any of the features, reach out to your usual Timextender contact and we’ll set up a focused session.