Excel Under Threat: The New Era

Key insights

• Endex: An Excel add-in that reads messy workbooks like a human analyst to find hidden errors, extract PDF tables into formulas, and build full financial models.
  Reviewer ran four practical tests on common pain points—cleanup, error detection, PDF extraction, and model building—and found Endex speeds up work but still needs human validation for complex judgments.
• Copilot: Microsoft’s AI assistant now sits inside Excel as chat and agent-based tools that let users ask questions and get guided help.
  It aims to replace some manual formula work with conversational prompts and step-by-step task automation across Microsoft 365 apps.
• =COPILOT(): A new cell function (in preview) that accepts natural language prompts, like asking for a list or summary, and returns structured outputs directly into sheets.
  It reduces the need to hand-write complex formulas for common data tasks.
• Power Query: AI enhancements speed up data cleaning and transformation by auto-detecting entities, categorizing fields, spotting outliers, and suggesting fixes.
  These upgrades make merging and standardizing sources faster, especially for large or inconsistent datasets.
• Automation: Excel’s AI adds features such as formula auto-complete, a Clean Data button, and pivot table auto-refresh to cut repetitive work and lower error risk.
  Users get faster insights through automatic charts, trend detection, and built-in suggestions that help both beginners and experts.
• Microsoft 365: Many AI features require a Microsoft 365 subscription and show up first in preview channels.
  Microsoft is bundling more AI and management tools into its plans, which will affect pricing and availability starting mid‑2026.

Video summary

In a recent YouTube video, Mynda Treacy (MyOnlineTrainingHub) [MVP] put a new AI add-in called Endex to the test inside Excel. She introduced the tool's promise to interpret messy workbooks, find hidden errors, extract data from PDFs into formulas, and even build whole financial models from scratch. Consequently, she set up four practical tests that demonstrate how the tool works in realistic finance and modeling scenarios. The video aims to show what Endex can do in practice rather than simply relaying vendor claims.

How Endex approached common Excel headaches

First, Treacy showed how Endex tackles the task of understanding large or poorly structured workbooks, an issue many analysts face. The add-in appeared to read sheet context, detect key inputs and outputs, and generate a clearer structure that made the model easier to navigate. Moreover, she highlighted the tool’s ability to surface potential errors, which can save hours when auditing inherited files.

Second, the video demonstrated Endex extracting data from PDFs and converting it into workable spreadsheet formulas, thereby reducing repetitive manual copying and reformatting. This feature seemed to perform well for straightforward tables, yet Treacy noted that more complex or inconsistent PDFs still required human review. Therefore, while Endex speeds up initial extraction, it does not fully remove the need for careful verification by an experienced analyst.

Building models and automating templates

Next, Treacy tested Endex’s ability to generate templates and full financial models from a brief description, which is one of its headline capabilities. The add-in produced working structures with formulas and layout suggestions, and the generated models were usable starting points for further refinement. Still, she emphasized that outputs often need tailoring for specific conventions, company logic, or regulatory nuances, so professionals should treat the AI as an assistant rather than a replacement.

Finally, she walked through a cleanup scenario where Endex reorganized messy sheets and suggested improvements to naming and formula consistency. In many cases the changes improved readability and reduced the risk of reference errors, but Treacy observed that automated fixes can occasionally alter intended calculations if assumptions are ambiguous. Thus, the workflow benefits from a two-step process: allow the AI to propose changes and then validate them manually.

Comparing Endex to Microsoft’s Copilot and related AI

Treacy also contrasted Endex with built-in and emerging Excel AI features such as Copilot and formula functions that interpret natural language. She noted that Endex focuses specifically on financial modeling and workbook repair, giving it a narrower but deeper set of capabilities for analysts. Conversely, Microsoft's ecosystem aims for broader integration across Microsoft 365 with chat interfaces, function calls, and automated suggestions that serve many use cases beyond finance.

Consequently, Endex can outperform general-purpose agents on finance-specific tasks, but it may lack the seamless cross-app workflows and governance expected from enterprise-grade Copilot integrations. For organizations that already rely heavily on Microsoft 365 agent features, choosing between a specialized add-in and platform-native AI requires weighing depth against integration. In practice, many teams may use both, applying each where its strengths match the task.

Tradeoffs and practical challenges

There are clear tradeoffs when adopting AI tools like Endex: speed and productivity gains come with oversight and validation costs. While AI can automate tedious work and surface issues quickly, users must accept that false positives and incorrect assumptions will occur, especially with complex or bespoke models. Therefore, teams must allocate time to review AI-generated outputs and to document decisions the tool makes on their behalf.

Data governance and security are additional concerns, particularly when tools process sensitive financial models or proprietary data. Organizations need to evaluate how inputs and outputs are stored, how models are shared, and whether the vendor’s approach aligns with internal compliance requirements. Hence, the decision to adopt such tools should balance potential efficiency gains against risk management and control processes.

What this means for users and next steps

Treacy’s tests show that Endex can significantly speed up model comprehension, error detection, and template creation, which will appeal to analysts who spend much of their time debugging and restructuring spreadsheets. However, she stresses that the best outcomes come from combining the AI’s output with experienced human judgment, rather than relying on automation alone. Teams should trial the tool on non-critical files, establish validation checkpoints, and define clear ownership for any changes the AI recommends.

In conclusion, the video provides a pragmatic look at how purpose-built AI add-ins can transform specific workflows in Excel, while also reminding viewers of the work that remains for humans. As AI features continue to evolve inside Microsoft 365 and through third-party tools, organizations will need to weigh functionality, integration, and governance to choose the right mix of solutions for their needs. Treacy’s hands-on approach gives practitioners a useful starting point for evaluating whether and how to incorporate Endex into their modeling toolkit.

Keywords

end of Excel, is Excel dying, future of Excel, Excel alternatives, AI replacing Excel, Excel vs Google Sheets, Microsoft Excel updates, spreadsheet future