The Death of VBA? Three New Tools Write Python for Excel Automatically

For decades, automating Microsoft Excel meant relying on legacy macros and Visual Basic for Applications, commonly known as VBA. While powerful, mastering VBA required a dedicated programming background, leaving countless spreadsheet users trapped in repetitive manual cycles of formatting, data cleaning, and reporting. Today, the landscape of spreadsheet automation is shifting dramatically as a new generation of tools emerges to write the underlying code automatically.

Rather than forcing finance professionals and data analysts to learn complex programming languages from scratch, modern software is bridging the gap by translating spreadsheet actions or natural language prompts directly into Python. Among the frontrunners in this space are Mito, Microsoft’s deeply integrated Copilot, and Julius AI. Each platform offers a fundamentally different approach to automating spreadsheet workflows, varying significantly in cost, architecture, and target audience.

Mito

Taking the familiar concept of a macro recorder and entirely modernizing it, Mito swaps out aging VBA for robust Python code. Operating as a specialized spreadsheet interface that runs seamlessly inside Jupyter notebooks, as well as Streamlit and Dash applications, Mito translates every single edit made within its grid into equivalent Python code displayed directly in the cell below.

Whether a user writes a lookup function, filters a complex column, constructs a pivot table, or generates a dynamic chart, the corresponding code appears instantaneously. When the time comes to update the report the following month with a fresh data export, the entire workflow rebuilds itself automatically, eliminating the need to repeat repetitive manual clicks.

This dynamic makes Mito particularly valuable for data analysts who spend significant portions of their work weeks rebuilding recurring monthly reports. Because the generated code remains completely visible, individuals who never intended to learn Python can gradually absorb the syntax simply by observing it populate in real time. Furthermore, Mito integrates an artificial intelligence chat assistant capable of generating or repairing code from a plain-English description, easing the friction when a complex data transformation is easier to describe verbally than to execute manually through interface clicks.

Despite its utility, setup represents the primary hurdle for new users. Mito functions as a collection of Jupyter extensions, requiring a straightforward command-line installation using package managers for its core packages before launching through JupyterLab. For professionals who have never interacted with a notebook environment, this introduces a brief learning curve before the familiar spreadsheet interface appears, setting it apart from tools that live natively inside traditional spreadsheet applications.

Mito’s business model is split between open-source offerings and tiered enterprise services. The core platform is open source, with a complimentary free plan that supplies the basic spreadsheet interface alongside a restricted allocation of monthly AI completions. For professional users requiring heavier workloads, a paid tier unlocks unlimited completions, advanced formatting features, and enterprise-grade telemetry controls. Meanwhile, enterprise customers retain the flexibility to connect proprietary language models, including locally deployed options, ensuring sensitive corporate data never leaves internal systems. Ultimately, Mito functions as an effective bridge, preserving the intuitive nature of spreadsheet workflows while producing reusable, shareable, and reviewable scripts.

Copilot, of all things

While some industry observers might question the inclusion of Microsoft’s flagship enterprise assistant in a pure code-generation discussion, the evolution of Microsoft’s native ecosystem makes it a central player. Through a strategic partnership with Anaconda, Microsoft integrated Python directly into Excel cells, allowing code to execute natively within the workbook architecture. Copilot now bridges the gap by writing that code entirely from plain-language user requests.

Users can instruct Copilot to eliminate duplicate records, reshape a complicated table, or visualize a emerging trend, prompting the assistant to insert the necessary Python code, execute it securely, and write the resulting data directly back into the spreadsheet. Crucially, the generated code remains visible and fully editable, ensuring that users with coding proficiency retain the ability to inspect and modify the underlying logic.

This native integration has matured at a rapid pace across multiple platforms. Advanced analysis capabilities rolled out widely to desktop environments, followed by expanded multi-platform support across web and mobile ecosystems. Subsequent feature updates introduced specialized planning modes that outline proposed procedural steps for user approval prior to executing modifications, alongside dedicated semantic skills capable of generating advanced data science formulas on demand. Because deployment timelines can vary considerably across different enterprise cloud tenants and subscription models, organizational IT departments typically verify specific feature availability before rolling out team-wide initiatives.

Beyond functionality, this approach introduces distinct security advantages. Python execution within Excel occurs inside isolated cloud containers managed by Azure infrastructure rather than executing directly on local hardware. Furthermore, native security protocols such as Protected View actively intercept and block Python formulas originating from internet downloads or untrusted sources, returning a designated error state until the file is explicitly verified by the user.

The primary limitation surrounding native integration remains financial. Utilizing the advanced in-Excel editing and code-generation experience requires a specialized enterprise productivity license, as standard entry-level chat tiers do not include the underlying computational backend. For organizations already heavily invested in Microsoft’s corporate software ecosystem, this automation layer arrives without incremental infrastructure costs. However, for independent users or smaller firms, the mandatory subscription overhead makes it a significantly pricier alternative compared to standalone web platforms. Its core strength remains data governance, keeping sensitive enterprise workflows strictly confined within established corporate security boundaries.

Julius AI

Positioned as the most hands-off option among modern automation platforms, Julius AI approaches data analysis through a conversational paradigm. Rather than managing local development environments or navigating proprietary spreadsheet plugins, users simply upload or connect their existing data files—ranging from standard spreadsheets and comma-separated values to cloud-based sheets and database connectors—and ask questions using natural language. Behind the scenes, the platform interprets the request, writes and executes the necessary Python or SQL code, and returns polished charts, structured tables, and synthesized written insights without requiring any local software installation.

For recurring and repeatable analytical tasks, Julius utilizes structured notebooks that combine text narratives, data inputs, and sequential analysis steps into reusable templates. An analytical notebook configured around a specific period’s financial records can be rapidly pointed toward subsequent datasets with minimal friction. For advanced users seeking finer control, the platform accommodates manual code cells, allowing analysts to inject custom Python scripts for specialized operations while the system manages underlying technical dependencies and library imports.

The platform has also expanded its accessibility through third-party integrations, including conversational applications that allow users to update formulas, consolidate disparate worksheets, clean messy data, and revise budgetary forecasts directly within saved cloud workbooks. These integrations can subsequently generate visual dashboards, graphical charts, and editable presentation files. While tiered subscription models provide flexible entry points for individual professionals and growing teams, pricing structures frequently evolve alongside platform updates.

As with any automated system relying heavily on machine learning models, certain tradeoffs accompany the convenience. When an automated language model attempts to parse and clean unstructured columns with ambiguous headers, occasional misinterpretations can occur, making manual spot-checking against primary source files a prudent practice for critical financial reporting. Additionally, transmitting proprietary company data to external cloud-based platforms necessitates careful review of corporate data privacy policies and compliance requirements, a consideration that varies depending on whether an organization utilizes local deployments, managed cloud environments, or dedicated enterprise security tenants.

Selecting the appropriate automation pathway ultimately depends on an organization’s existing technical ecosystem and daily operational workflows. Platforms built around interactive notebook interfaces suit analysts seeking to generate transparent, reusable scripts, while native enterprise integrations serve organizations deeply embedded within standardized productivity suites. Meanwhile, conversational analysis platforms appeal broadly to users looking for rapid, natural-language insights from raw data files. While each of these tools successfully lowers the historical barriers associated with programming, none entirely eliminates the professional responsibility of verifying analytical outcomes.