CEO Copilot
Personalized copilot for managing your company
Note: all the technologies mentioned here except for LLMs are widely available for quite some time. Seeing that no solutions, similar to the described, are widely used now, gives me a hope we will not be there soon, and still rely more on human-to-human interactions in the company management. However, thinking about how the future can turn out for us is always interesting and sometimes even useful, especially if the picture we paint is not very bright.
Imagine that you are managing a company. Not very small, but not too big either (I’ve never worked in big ones, so cannot perform this “imagine” thing about them, sorry). The big proportion of your time will be spent on strategic things, but you’ll probably spend tons of time on various status checks. You need to know how people are doing, are they happy or not, and what’s on their minds. Probably there are some particularly important projects that you’d like to track as well. Of course, it’s important to see some key metrics and trends. And maybe it’s useful to be aware of new big or unexpected problems happening too.
So many things to think about!
How do you perform all these checks now? I suppose a part of them is done using various tools for technical, customer support and sales metrics monitoring (Grafana, Monday). Lots of useful info can be seen using project management software (Jira). All these dashboards, roadmaps etc. As for people-related topics, another category of tools exists that allows to perform various surveys and analyze the data (CultureAmp).
However versatile and rich all these tools are, they still lack a universal and convenient UI for accessing random information and making insights. Their solution is providing possibilities to configure all kinds of dashboards to look at the data from different perspectives. This means that there is always someone who has to perform work to configure these dashboards before they are ready to use. As a consequence, you as a CEO will often come up with questions that cannot be answered by these tools immediately. The existing solutions for you are:
Obtain a certain level of expertise to be able to use SQL or JQL to query data yourself.
Ask your employees to answer the questions.
Option 2 is the typical way to go now, especially considering that lots of things happening in a company are not covered by the existing data analytics tools. You talk to your managers, and they answer the questions, give status updates, and raise any outstanding points to you.
Q1 is answered by the dashboard based on the data from Sales Team; Q2, Q3, and Q4 are asked on a 1-2-1 with your VP; they answer Q4 themselves, and go with Q2 and Q3 to the team and Team Lead respectively. Not very fast, but quite reliable, and you talk with nice people during the process.
So, despite all the new tools and technology progress, a CEO’s primary interface to their company is conversations (written or verbal) with people.
Now, let’s look at a typical software development remote company. All the project management is done in Jira, various information is stored in Notion. People communicate via Slack, Teams or Zoom. There are also software tools for Sales, Marketing, HR, Recruiting and all other teams. Anyway, the specific tools can be different, but their functionality is very similar.
Every remote company already sits on top of the data that could be analyzed and used for providing useful insights at any time. And with our desire to automate everything (software is eating the world, right?), I would ask the question: why do you even need people to access the company information? The two things already mentioned are that not all information is analyzed with the tools currently available, and there is no convenient user-friendly interface to access the data in place.
And if my understanding of the current AI developments is right, we’ve come to the point when we might have solutions for the two problems. And if this is the case, it could potentially disrupt the way modern companies are managed and work in general!
Lets see what I mean here.
People
The biggest part of a typical company information not readily available in some BigQuery-like storage is people. That’s why products like CultureAmp exist. They provide tools to understand how employees see their work, are they happy or not, do they share the company vision, if they plan to look for another job soon etc.
And actually, I think our technical capabilities are enough to see such information already without even asking people explicitly. If all the communications are done using software tools, a company could analyze the data from chats and provide a lot of useful hints, statistics etc.
You will always know.
The tool providers could even consider privacy aspects, and let’s say never share the actual messages from particular people, but still give hints like “the average sentiment towards company values for Sales Team has changed from neutral to negative, take a look!”. The data is here, the tools are here, so what’s stopping us?
A small caveat
While we’re used for our text messages and emails to be stored and then analyzed, there is still verbal communication happening (people’s favorite Zoom meetings). If a company decides to analyze the recordings, it probably might have some issues with privacy protection, as people want to have their small talks without being recorded. The solution? Add a switch “Working conversation” - “Personal conversation”. “Working conversation” is enabled by default, and everyone knows the data will be analyzed. If you want to have some personal talk, no problem, just switch to “Personal conversation”!
And to help people not to forget to enable “Working conversation” mode, our evil video conferencing app could use some small nudges like “Your personal conversation is happening for 10 minutes already. Maybe you started talking about work, but forgot to switch the mode (wink)?”. They could also show some stats like “Average person in our company has 30 minutes of personal talks, and 2 hours of work-related discussions. This week your stats are 60 minutes of personal talk and 1.5 hours of work discussions on average. Please ask your manager to take a look and help you to be more effective (smile).”
Helping you and your colleague to have efficient and result-oriented call.
Interface
Having all the information stored and ready to be requested, we need an interface. ChatGPT and similar products provide you such possibility! It can understand and answer your questions regarding a set of text data, and even provide ad hoc analysis! See for example ChatPdf. Going forward, you won’t need any predefined dashboards that limit your imagination! With the tools like GPT Code Interpreter1, you will see any data insights about your company you could ever dreamt of!
Now Q1, Q2 and Q3 are answered instantly based on the data from Sales Team. Only more complex Q4 requires a person to answer it.
Yes, the tools are not yet as good at answering the questions as humans, but they are readily available at any moment, you don’t need to schedule a call or wait for a person to answer after the weekend. And if these new tools have all the information about the company available, they often can be faster. They will be almost as good as people at many tasks, and it’s interesting to see if this will be enough to completely change the way you see your company as a CEO.
Future
I do not know where we will end up. But as the technology is quickly becoming better, we will see very interesting changes soon regardless of whether we want it or not.
See great examples by Ethan Mollick here https://twitter.com/emollick/status/1652170706312896512