Ranking The Features of AI-Enabled Excel 2019
Last week Microsoft offered the world a sneak peek of Excel 2019, which is due to start rolling out early next year. The headline item from the closed-doors unveiling was the news that Excel, for the first time, will include capabilities that let teams work with machine learning algorithms inside the legendary spreadsheet software.
We don’t know a lot about the new capabilities beyond what Microsoft has sketched out in a blog post and demoed for TechCrunch’s Frederic Lardinois, but from what the team has shown so far, Excel 2019 will be able to understand more about your data and contextualize it with external information. In the demo, Microsoft touted three new features of Excel that promise to bring AI to your spreadsheets in the near future. Whether the features will be useful is a different story.
Here are three new things Excel 2019 will be able to do, ranked according to the likelihood of you actually using them:
To learn more about how AI is changing your finance team, check out The CFO’s Guide to AI Strategy.
Import machine learning models into Excel
Machine learning—the brains behind artificial intelligence—is a powerful method of data processing that lets algorithms find patterns in large datasets. Starting with this new release, Excel will be able to import these algorithms directly, and presumably, run them inside the software.
This isn’t something everyone’s going to use immediately, but it marks an important moment for tech-forward Excel jockeys and machine learning beginners alike. If this feature allows teams to start using machine learning as routinely as they build pivot tables, Excel 2019 might ultimately mark the moment when AI fully democratized for the business world.
Usefulness: High. Just like VisiCalc got PCs on corporate desks en masse, spreadsheet software could also bring machine learning to a desktop near you.
Automatically generate data visualizations
Visualization is an essential analytic technique in the era of big data. It makes information contextual and accessible; almost tangible. Abacus has a feature called Insights that helps you understand your spending through data visualization. As of 2019, so will Excel.
Microsoft touts their “Insights” as a feature that identifies interesting information and presents it graphically—and automatically—to users. “It is meant to take any list of data and then start to generate insights,” says Jared Spataro, the Excel team member who demoed TechCrunch. “It will look at combinations, charts, pivot tables and it will recognize those that are most interesting by looking at outliers, looking at trends in the data, looking at things that represent changes.”
While visualization is indeed an important, almost necessary tool for understanding data, my hesitation with this feature comes from past trauma inflicted by Microsoft Office’s automated helpers. This feature sounds like an artificially intelligent Clippy, which is a terrifying thought. Office 97’s annoying paperclip was notoriously bad at predicting what you were trying to do and what you needed help with. Will Excel’s Insights suggestions be much much better?
Usefulness: Medium. The spreadsheets of tomorrow will contain more data than we’ve ever had to wrangle before. Therefore, data visualization will be essential. If Excel can take a light and unobtrusive touch with these automated graph suggestions, the feature could be helpful. We’ll have to see how this one goes.
Contextualize your data by pulling from the web
When you input data into Excel, the software tries to detect whether you’ve entered text, a date, currency, etc. With the help of Microsoft’s Bing API, Excel 2019 will attempt to understand even more.
The software will be looking for a data type for which it can search the web and find additional information. Microsoft offered an example: if you tag a column “Company Names,” Excel will recognize the dataset to be a series of companies for which it can search Bing to find additional data. You could instruct Excel to automatically pull in live data about the companies’ market caps, headquarters addresses, and so on.
Giving Excel greater access to the outside world via a search engine API catches it up with other spreadsheet software like Google Sheets, but their use case seems half-baked. Porting live data across dynamic spreadsheets is already nerve-wracking; entrusting your models to the discretion of Excel’s natural language processing sounds like malpractice.
Usefulness: Low. The level of utility will depend on your use case and your level of technical skill, but it’s hard to imagine many people putting their numbers in the hands of unproven natural language processing. It will be interesting, however, to see what emerges from an Excel that is fully tapped into search engines.
Excel looks poised to compete in an AI world
In recent years, Excel users who have wanted to augment their analysis with machine learning have had to export their data to smaller, more experimental analytics software. With this latest release, Microsoft is demonstrating they’ve gotten the picture.
We don’t yet know what other capabilities will be baked into Excel 2019, but from this preview, we can tell they’re heading in the right direction. Excel is ground zero for the type of task-oriented manual work that, moving forward, will increasingly automate off the desks of busy finance people. It’s encouraging and exciting that Microsoft is planning for Excel itself to lead the charge into that future.
See the Abacus report “The CFO’s Guide To AI Strategy” for more on how artificial intelligence is changing the finance team.