Searches in our files and emails are painful.
We’re used to it now. We all play the game of trying to remember some key word that will pull up just the right file except it doesn’t, or the appointment last year that will clear up some mystery if we can just find it on the calendar, or the email we kind of remember from last week but it’s not in the inbox and it’s not in junk mail and it’s not in deleted items and WTF??!! Outlook search is useless and Windows search is don’t even bother, and Google Search is struggling to show good results from the seriously enshittified internet.
I want to describe how AI will improve that process for employees in large companies. Some of the words are a little geeky but stick with me, this is Geek Lite because I don’t understand it very deeply either.
If you work in a large company, you probably have access to hundreds of thousands of company documents and too much of your time is spent looking for needles in those haystacks. You might get the wrong result, or you might get thousands of irrelevant results, or you might not realize you’re looking in the wrong file store or the wrong database.
There’s one important extra part of enterprise searches: you only get to look at files that you’re cleared to see. Maybe you don’t have permission to get into HR folders, or financial records, or company contracts, or the CEO’s calendar.
In the IT world it’s called “identity management.” When you log in, the system knows who you are and what you have permission to see. It’s a central element of security for a large company. Most business hacks involve a bad guy logging into an account disguised as someone with permission to access important company files.
Large companies have servers in the cloud or onsite that hold all the company documents, all the emails and calendars, all the team chats, all the contracts, all the finance data, everything. That data is scattered in different places: email servers, file servers and network drives, databases, collaboration platforms like Teams or Slack, third-party platforms, and more.
One of the exciting things that AI can do is sift through large amounts of data. You will be able to use AI to look through all of those sources to find answers to your questions. The AI will pull information together from multiple sources and summarize it in understandable paragraphs.
In large businesses, AI will deliver personalized answers that draw from all the data everywhere in the company at the same time, but it will do it in a way that respects user permissions.
Let me give you an example. This is hypothetical and made up and probably wrong in deep and fundamental ways but shut up.
Imagine that Megalomania Studios has had huge box office success with its SF movie Quantum Quandary.
The question is: “How much did Ryan Reynolds get paid for his role in Quantum Quandary?
We’re going to look at how AI might answer that question for three different people: a random person; a low level studio executive; and the CEO.
You ask that question as a member of the general public. The AI pulls from publicly available information and says something like this.
“While exact figures aren’t always released, industry reports suggest Ryan Reynolds commands salaries in the tens of millions of dollars for leading roles in blockbuster films like Quantum Quandary. Estimates place his pay somewhere between $20 and $30 million, though this is speculative.”
That’s decent if a little vague. Maybe there’s a link to an article in Variety about the actor’s movie deal.
Now imagine you’re working as a low-level executive at Megalomania Studios and you tap this question into a company laptop while you’re logged into the company network. You have access to some internal studio data but not the most sensitive financial details. The AI response tells you this.
“Internal budget documents show a figure in the range of $25 million for Ryan Reynolds’ performance. This includes his base salary and potential bonuses tied to box office performance. However, this figure doesn’t include any backend participation or merchandising deals.”
Finally, you’re the CEO of Megalomania Studios with complete access to all company records. The AI response is something like this.
“Ryan Reynolds’ total compensation for Quantum Quandary is $32.5 million to date. This breaks down into a $20 million upfront base salary, a $5 million box office bonus triggered by exceeding expectations, and a $7.5 million share of the merchandising revenue. There will be significant additional payments beginning in the third quarter for backend participation. There is an additional contract with Mr. Reynolds’ LLC negotiated at the same time as the Quantum Quandary contract to cover restructuring of backend participation for previous movies, bonuses for awards nominations, and long-term adjustments to residuals if sequels and merchandising revenues meet certain thresholds. A detailed breakdown of these components is available in the attached financial report. There have been two meetings recently with representatives of Mr. Reynolds about renegotiating profit participation if he can convince Hugh Jackman to join the sequel.”
Obviously an executive assistant could put that information together for the CEO in half an hour. But the AI will provide that answer in seconds and stay available for whatever tangent then crosses her mind. “How much will Mr. Reynolds make from the Funko Pops deal?” “Does the studio pay for Mr. Reynolds’ wig of 100% ethically sourced yak hair?” “What is the status of negotiations over character licensing rights for the Quantum Quandary theme park ride?” You know, the usual industry stuff.
The answers are different because the AI knows who you are and what you have permission to find out. The CEO can ask when her next meeting is scheduled but a sales team member doesn’t have access to the CEO’s calendar. Over in HR, a manager can ask for a detailed performance review for John but John’s random work colleague can’t get that information.
There are two easy predictions.
One: This is transformative! Finding information and obtaining summaries from corporate data in seconds instead of doing arduous searches – that’s a game changer.
Two: Microsoft will continue to be one of the biggest companies in the universe.
I’ll tell you more about that in the next article.