{"id":9569,"date":"2026-09-07T04:45:03","date_gmt":"2026-09-07T04:45:03","guid":{"rendered":"https:\/\/yodaplus.com\/blog\/?p=9569"},"modified":"2026-09-07T05:18:15","modified_gmt":"2026-09-07T05:18:15","slug":"what-tools-are-analysts-using-to-automate-modelling-workflows","status":"publish","type":"post","link":"https:\/\/yodaplus.com\/blog\/what-tools-are-analysts-using-to-automate-modelling-workflows\/","title":{"rendered":"What Tools Are Analysts Using to Automate Modelling Workflows?"},"content":{"rendered":"\n

Analysts automating modelling<\/a> workflows today combine four layers: Excel with AI features built directly into the spreadsheet, general-purpose AI assistants like Claude and ChatGPT for drafting and research, purpose-built research platforms for data retrieval, and dedicated AI FP&A platforms once a model outgrows what one analyst can manage manually. Gartner’s latest CFO survey found AI adoption across finance functions climbed from 37% in 2023 to 59% today, with nearly 60% of CFOs planning to increase AI spend by 10% or more in 2026.<\/p>\n\n\n\n

That growth is real, but it is worth being precise about what it actually changed. Here is the tool landscape analysts are working with right now and where each layer fits.<\/p>\n\n\n\n

Excel Remains the Base Layer, Now With AI Built In<\/h3>\n\n\n\n

Spreadsheets are still the dominant surface for financial modelling, and that is not expected to change soon. The reason is not inertia. Excel remains the most flexible analytical surface available, and the AI features added between 2024 and 2026 have made it meaningfully more productive rather than replacing it.<\/p>\n\n\n\n

Microsoft 365 Copilot brings AI assistance directly into live Excel workbooks, handling formula generation, pattern recognition, and natural language data analysis. Enterprise pilots report financial modelling running 30 to 40% faster with Copilot active, one of the most consistently cited productivity gains across AI tool categories. Copilot operates within an organisation’s existing Microsoft tenancy, keeping data inside the firm’s own environment rather than sending it to an external model provider, which has made it the lowest-friction starting point for many teams.<\/p>\n\n\n\n

General-Purpose LLMs as the Unofficial Junior Analyst<\/h3>\n\n\n\n

Alongside Excel, general-purpose AI assistants like Claude and ChatGPT function as an informal extra set of hands, used for drafting commentary, structuring a model’s logic, and stress-testing assumptions before they go into a live spreadsheet. Wall Street Prep’s June 2026 comparison tested Claude, ChatGPT, and Microsoft Copilot Agent Mode against real investment banking standards, building a fully integrated three-statement model, giving analysts a concrete sense of where each tool actually helps versus where it still falls short on production-grade work.<\/p>\n\n\n\n

These tools work best positioned as a reasoning partner rather than a model-building engine on their own. Analysts use them to sanity-check an assumption, summarise a filing before building it into a model, or draft the first version of variance commentary that a human then edits and finalises.<\/p>\n\n\n\n

Purpose-Built Research and Data Platforms<\/h3>\n\n\n\n

A separate category of tools focuses specifically on data retrieval and research rather than modelling itself. Equity analysts commonly rely on platforms like AlphaSense, Kensho, and Kavout to search earnings transcripts, filings, and broker research quickly, feeding cleaner, pre-processed information into whatever modeling tool comes next.<\/p>\n\n\n\n

These platforms solve a different problem than Excel or a general-purpose LLM. Instead of building the model, they compress the time spent finding and validating the inputs a model needs, which is often the more time-consuming part of the workflow for research-heavy roles.<\/p>\n\n\n\n

Where Python and Code-Based Tools Fit In<\/h3>\n\n\n\n

For quant-leaning analysts and work that needs to scale or run automatically, Python and R inside dedicated platforms increasingly complement Excel rather than replace it. Custom backtests, regression analysis, derivative pricing scripts, and large-dataset extraction jobs are handled through code-based environments, with AI-assisted coding tools helping generate and debug that code faster.<\/p>\n\n\n\n

Most finance teams end up running all three surfaces at once: Excel for bespoke, one-off modelling; a recurring dashboard tool like Power BI or Tableau for standard reporting; and a programming environment for anything that needs to scale beyond manual spreadsheet work.<\/p>\n\n\n\n

AI FP&A Platforms for Scaling Beyond One Analyst’s Model<\/h3>\n\n\n\n

Once a model outgrows what a single analyst can reasonably hold in their head, purpose-built AI FP&A platforms with audit trails and explainability become the relevant evaluation category, rather than continuing to stretch a spreadsheet or a chat-based assistant further than it was designed to go. These platforms typically manage the full model lifecycle, covering building, refining, maintaining, scaling, governing, and communicating results, with governance features that a general-purpose tool does not offer out of the box.<\/p>\n\n\n\n

What Analysts Are Actually Doing Day to Day<\/h3>\n\n\n\n

Despite the tool landscape expanding quickly, most analysts have not fundamentally changed their daily workflow. Analysts still open Excel, reuse existing templates, and build models largely the way they always have, with AI tools layered in for specific, bounded tasks rather than replacing the core process. Banks in particular are piloting a growing number of AI tools, but security concerns and uneven reliability have kept many deployments in an experimental phase rather than full production use.<\/p>\n\n\n\n

Common Challenges in Adopting These Tools<\/h3>\n\n\n\n

Integration friction with Excel-centric workflows:<\/strong> Analysts depend heavily on Excel, and tools that require moving data in and out of a separate browser interface too often create enough friction to slow real adoption.<\/p>\n\n\n\n

Over-reliance without questioning outputs<\/strong> Some analysts accept AI-generated figures without adequately verifying assumptions, particularly when a tool’s output looks polished and confident regardless of underlying accuracy<\/a>.<\/p>\n\n\n\n

Governance and transparency requirements<\/strong> Regulatory expectations require clear explainability for any AI-assisted output feeding into a decision, which not every general-purpose tool is built to provide.<\/p>\n\n\n\n

Skill gaps in supervising AI output<\/strong> Analysts need new competencies in data literacy, scenario validation, and model oversight to use these tools effectively, and many firms are still building out that training.<\/p>\n\n\n\n

Best Practices for Choosing and Using These Tools<\/h3>\n\n\n\n