Your D365 Data Management Framework Isn’t Ready for AI. Here’s How to Fix That.

April 16, 2026
Budgeting and Planning
Tim Turner

Ask any SME business leader what their biggest priority is right now, and AI will come up within the first two minutes. Ask them how their data is tracking and the conversation gets a lot quieter.

That gap between the pressure to use AI and the reality of the data underneath it is where most organisations are stuck right now. And it's not a small gap. Research shows that somewhere between 75 and 80% of businesses that have moved on AI haven't seen the results they were expecting.

The technology isn't the problem. The data is.

Fragmented systems. Inconsistent definitions. Sources that have never been properly connected. If the foundation isn't solid, and if the underlying data isn’t ready, it doesn't matter how sophisticated the AI tool sitting on top of it is. You'll get fast answers with inaccurate and inconsistent results.

The Sticky Position Most SME Leaders Are In

Here's what makes this lack of data readiness genuinely difficult: most mid-market and SME leaders already know their data isn't where it needs to be. That's not the revelation. The paralysis comes from what to do about it.

On one side, there's real pressure to act. AI is moving fast. Competitors are making noise about it. The board expects progress. Sitting still feels like a risk.

On the other side, making the wrong move feels just as dangerous. The questions that keep leaders up at night aren't abstract, they're specific and reasonable:

What if we choose the wrong platform and we're stuck with it for years?

What if the data turns out to be worse than we think and we make decisions we can't stand behind?

We don't have the team for a full data overhaul. Is this too big for where we are?

These fears are legitimate. Platform lock-in is real. The cost of unwinding a poorly chosen architecture is real. And for a business without a dedicated data engineering team, the complexity of standing up a modern data environment from scratch is very real.

The problem is that waiting for certainty before acting isn't a strategy. It's just a slower version of being left behind.

AI-Ready Data, On Demand

Unifying and future-proofing data for AI

How TimeXTender Solves the AI-Ready Problem

TimeXtender is a metadata-driven data integration platform, which sounds technical, but the practical implication is straightforward: it removes the biggest risks that make data modernisation feel so daunting.

The first risk it removes is platform lock-in. Most data infrastructure gets expensive to change because of the logic that drives it. The transformation rules, the business definitions, the data model are all written directly in the language of whatever platform it runs on. SQL code for SQL Server. Spark notebooks for Microsoft Fabric. When you want to move platforms, you're not just migrating data. You're rewriting your entire pipeline, usually with different specialists for each environment.

TimeXtender works differently. All of that logic is defined at the metadata level, once. The platform then generates the correct code for whatever target technology you're pointing it at. Want to run on Microsoft Fabric? It writes the Spark notebooks. Moving to Snowflake? Done. The model doesn't change. Just the destination.

This means you're not making a permanent bet when you choose a starting point. You can begin on-premise, move to the cloud when you're ready, add platforms as the business grows, and swap endpoints without dismantling the work you've already done. The investment builds over time rather than becoming a sunk cost every time the technology landscape shifts.

The second risk it removes is complexity. Instead of stitching together separate tools for ingestion, transformation, modelling, and delivery (each with their own interfaces, specialists, and failure points), TimeXtender handles the entire pipeline in one place. Less to manage. Less to break. Less expertise required to keep it running.

The architecture is built around three straightforward stages.

Ingest is an always-on service that pulls data from all your sources: ERP, CRM, flat files, APIs, incrementally and on a schedule. Data lands here raw. No manual staging runs, no business-unit abstractions to maintain.

Prepare is where transformation and modelling happens. Raw data comes in, clean analytics-ready data goes out. This is your data warehouse stage, sitting on top of whichever cloud storage platform you're using — Microsoft Fabric Lakehouse, Snowflake, Azure Data Lake, AWS, or SQL Server if you're staying on-premise for now.

Deliver is where you define outputs. One model, multiple endpoints simultaneously. Power BI for your finance team. Tableau for operations. A REST API for system integrations. And increasingly, an AI-facing endpoint called an MCP server (more on that in a moment). You build the model once and point it wherever it needs to go.

Why The AI Lands Differently Here

Most conversations about AI and data stop at "connect a language model to your database." TimeXtender goes a step further, and the difference matters.

An AI tool that can access your data isn't the same as an AI tool that understands your data. A field called CUST_PRCGRP tells an AI nothing. But if your platform lets you attach a plain-language description such as "the pricing group assigned to this customer, used to determine discount eligibility,” the AI now has the context it needs to use that field correctly. It won't confuse a table called "Client" that actually stores server names with a table of actual customers. It won't misread a shipment date field as a billing date.

This is what AI-ready data actually means: not just accessible, but interpretable.

TimeXtender supports this through built-in metadata enrichment and an MCP server endpoint - a Model Context Protocol server, which is rapidly becoming the industry standard for connecting AI tools to business data securely and consistently. Connect ChatGPT, Claude, or a locally-hosted model to your MCP endpoint, and it's working from your governed, described, trustworthy data model, not guessing.

The practical result is that someone can ask "which customers are overdue and don't have a follow-up scheduled?" and get a real answer, grounded in real data, without a developer writing a query.

What Keeps Data Trustworthy Over Time

Getting data into a warehouse is the starting point, not the finish line. The part that actually determines whether your reporting and AI outputs stay reliable is what happens after, and this is where a lot of data programmes quietly fall apart.

Data Quality lets you define rules that catch problems at ingestion rather than six weeks later when someone spots something wrong in a board report. A customer record with no price list? The right person gets an alert immediately. A transaction that fails a validation rule gets flagged before it corrupts downstream reporting. It moves data governance from reactive to proactive.

Data Enrichment handles consistency across systems, harmonising the same entity as it appears in different tools (the customer in your CRM that doesn't quite match the customer in your ERP), and adding the contextual mapping data that currently lives in Excel spreadsheets maintained by one person who might leave.

Orchestration gives you a single view of your entire pipeline. When something fails — and it will — you can see exactly where, understand why, and restart from that specific step. No hunting through logs, no re-running everything from scratch, no guessing which package broke.

A Built-In Assistant for the People Running It

For the team managing the data environment day to day, TimeXtender includes Xpilot, an AI assistant embedded directly in the platform and trained on TimeXtender's full knowledge base. Hit an error you don't recognise, paste it in and get a specific answer. Can't remember how to configure a particular step, ask and get told. It's a small thing that saves a meaningful amount of time for anyone working in the platform regularly.

Xpilot Analytics is coming soon and extends this capability to business users, translating plain-language questions into queries, interpreting results, and surfacing insights without requiring anyone to know SQL. It runs through the MCP server, so the answers are grounded in your data model rather than generated from thin air.

Who is TimeXtender For?

TimeXtender tends to be the right fit for mid-market and SME organisations that:

  • Have a need to extract data from various systems, including their core ERP such as Dynamics 365 Business Central or Finance & Operations to create an Enterprise data estate.
  • Have outgrown spreadsheet reporting but aren't ready to staff a full data engineering team.
  • Want to move toward cloud infrastructure without committing permanently to a single vendor's ecosystem
  • Need a credible, practical answer when leadership asks what's being done about AI.
  • Have tried combining multiple point solutions and found the maintenance unsustainable.

What This Means for Jet Analytics Customers

If you're currently on Jet Analytics or using the Jet Data Manager, this is the direct evolution of the technology you already have. The existing on-premise version, being renamed Jet Analytics Classic, stays supported. But through the recently extended partnership between InsightSoftware and TimeXtender, everything described in this article is now accessible to Jet Analytics customers.

A migration tool already handles the transfer of your data sources, data warehouse models, and semantic models, including historical data, so you're not starting from zero. If you've been waiting to see which direction Jet Analytics was heading, that direction is now clear.

Where Business Insights Fits In

The technology is only part of the answer. Knowing where your data actually stands today — what's usable, what isn't, what needs to happen before any of this is worth investing in — is the harder conversation, and it's the one most vendors skip past.

That's what our analytics and FP&A experts at Business Insights do. We work with Dynamics 365 customers to get an honest picture of their data environment and build a path forward that's scoped for the team and budget they actually have. Not a three-year enterprise programme. A realistic plan that starts delivering value early.

If that sounds like a useful conversation, get in touch. No obligation, no pitch deck, just a straight look at where you are and what would actually move the needle.

Business Insights helps Dynamics 365 customers turn their data into reliable reporting, analytics, and AI-ready foundations they can trust.

Contact Us To Get AI-Ready Today

Frequently Asked Questions

A metadata-driven platform stores your data logic — transformation rules, business definitions, relationships — separately from the technology running underneath it. This means when you change platforms (say, from SQL Server to Microsoft Fabric), you don't rewrite your data model. The platform translates it automatically. For mid-market businesses, this is the difference between a data investment that compounds over time and one that has to be rebuilt every few years.

No. TimeXtender supports on-premise SQL Server as a starting point, so you can modernise your data architecture without immediately committing to a cloud platform. When you're ready to move. whether that's to Microsoft Fabric, Snowflake, AWS, or Azure Data Lake, the transition is handled at the platform level without starting over.

MCP stands for Model Context Protocol. It’s an emerging industry standard that creates a secure, structured connection between your data and AI tools like ChatGPT or Claude. Without it, an AI tool accessing your database is essentially guessing at what your data means. With it, the AI works from your governed, described data model and returns answers you can actually trust. If AI-assisted reporting or natural language queries are on your roadmap, an MCP endpoint is worth setting up now.

No. A migration tool exists specifically to transfer your existing Jet Analytics project, including data sources, data warehouse models, semantic models, and historical data, across to the new platform. You're not rebuilding from zero. The main consideration is multidimensional cubes, which need to be reconfigured in tabular format either before or after migration. Business Insights has done this for a number of clients and it's a manageable process.

It depends on the complexity of your current environment including the number of data sources, the maturity of your existing data model, and whether you're moving to cloud infrastructure at the same time. The best starting point is an honest assessment of where you are today. That's something Business Insights can help with before any commitment is made.

TimeXtender supports a wide range of connectors beyond Microsoft Dynamics 365 Business Central and Finance & Operations including CRMs, flat files, APIs, and many third-party systems. The platform has been expanding its native connector library significantly, particularly following the end of its CData partnership. If you have a specific source you're unsure about, it's worth a conversation to check current availability.

Fabric is a powerful platform, but it's a collection of components that still need to be configured, connected, and managed. TimeXtender doesn't replace Fabric, it sits on top of it and automates the coordination between ingestion, transformation, modelling, and delivery. The practical difference is speed and maintainability: instead of writing Spark notebooks and managing pipelines manually, you define your model once and TimeXtender handles the rest. For teams without dedicated data engineers, that distinction matters considerably.

Tim Turner

Founder and Principal

Tim is the founder and Director of Business Insights. Tim previously led the Jet Global team and partner channel for the AUNZ region. During this time, the customer base increased tenfold, and we worked with almost all Dynamics NAV/BC partners in the region. Tim has been involved in over 100 Jet Analytics implementations and has a deep knowledge of Dynamics NAV/Business Central. Prior to this, Tim was CFO for an exporting firm in New Zealand and has worked for other organizations such as BBC, UBS, and EY.

Fact Checked & Editorial Guidelines
Reviewed by: Subject Matter Experts
©2026 Business Insights. All rights reserved.