Bots for everything? Or is it?

The best part about business bots is that they are built around collaboration. Wired recently published its take on . We believe in this vision of the future and have bet our proverbial house on it. We started back in 2016 to build a suite of apps that leverages Slack to enable companies to dramatically improve their productivity. Like the Wired article mentioned, there are lots of obvious early use cases for chat including expense filing, and customer support. We agree with these examples but does it mean every business activity lends itself equally to the chat interface? Well, obviously no, but the use cases fall on a spectrum from their applicability to the chat interface. Here is how to figure out when it makes sense to move a function to chat

1. When what you are doing is actually chat — customer support is a prime example of this. When helping someone through chat, be it within an organization or speaking with prospective customers on the website through a live chat plugin. Here the native unit of interaction is purely chat, so it makes sense to use bring these chats to whatever platform you use internally for chat. This is what we do when speaking to our customers to provide them support.

When a customer reaches out to us for a chat, anyone on our support team can claim the chat and start talking This way we are able to bring both in product and on website support to a page where everyone on the entire team is always logged in. In case we see a spike in the support volume, anyone on the team can pitch in since everyone is seeing all incoming requests. Before we built this solution, only the support team was logged into a customer support portal which meant long wait times when ticket volumes spiked. Not only does this benefit the customers via reduced wait times, it also means that everyone on the team is providing support and talking to the customer at least a couple of times a week. Our friends at seem to agree with us and they have recently released their Slack bot that brings conversations with the customers through their web/app widget right to Slack for support personnel to always be in Slack.

2. When pulling in data into context —  and are the perfect examples here. They both allow you to pull in data from other sources or simply by asking others in your company, that is relevant to the discussion at that time. Data can be used to settle all arguments easily and the radical transparency it brings in helps to align the various parts of the organization towards number driven goals.

Statsbot in action. Image courtesy Statsbot.com

3. Sequential workflows — These are the classic workflows where the action of one person triggers another person to perform a subsequent action. Think of someone applying for a leave in Slack using AttendanceBot. The employee says they are on leave and then gets it approved by the manager.

Leave application by an employee triggering an approval process by the manager In this case while the actions inside Slack are quite simple, there a lot of magic that takes places behind the scene. When applying for a leave, AttendanceBot checks if the employee has the required leave balance of the given leave type, figures out if the given leave type requires an explicit approval by the manager and checks if anyone else on the team is on leave during those exact days so that the manager can take an informed decision. The moment the manager approves the leave, AttendanceBot instantly notifies the employee, automatically updates the leave balance, syncs the leave to the team’s shared calendar and sets an out of message for the leave duration. Think how much easier, this is inside Slack compared to any other way. is another prime example here. It allows employees to submit their expenses to their manager for approval and after they are approved, ExpenseTron syncs those to the accounting software like Xero or Quickbooks. Once synced, the finance team can reimburse the expenses using the accounting software and ExpenseTron closes the loop with the employee by notifying them in Slack as soon as the expense is marked as paid in Xero or QuickBooks.

Another sequential workflow where an Employee submits an expense that is sent to manager for approval

4. Todos /checklists — Lots of great examples here ranging from the general purpose bot which lets anyone assign to-dos to others on the team to the more niche that is used to manage tasks around the office relating to facilities, IT, HR, pantry and others.

To-Do bot in action.

Workflows that don’t make sense on chat — Well this one a more ephemeral category. Things that don’t make sense on the limited interaction elements of chat or need a larger canvas. It could also be that nobody has yet done this well enough. If you’re working on a bot for the following categories, don’t let this discourage you.

• Shopping — be it buying regular goods, booking flights or hotels or any other scenario where discovery is an important part of the interactions, the conversational interface does become a little bit of a hassle. Discovery needs a large canvas for us to quickly go through lots of options and narrow down our selections, which often means using filters or faceted search which comes more naturally to an app. Microsoft Teams recently introduced the , which might make it easier to build such apps within a messaging platform but then a tab really isn’t a chat interface.
• Booking an Uber — given the technical limitations of not having access to the GPS data when building on top of Slack, this use case is essentially dead on arrival. Given how much we love Uber and Slack, , it does leave much to be desired. I do think this one is a low hanging fruit and the moment Slack allows developers to get access to GPS and other mobile sensors data we will start seeing more and more novel bots on Slack.
• Banking and other high security transactions — while does have a bot on Slack which allows transferring money, most banks have not embraced Slack yet. Things like accessing finger print sensor right from inside Slack, might make this easier. Access to other sensors might also similarly enable health care bots in Slack which face similar privacy and security considerations.
• Configuration and data management — While receiving alerts is great, acting on them in Slack is still hard. Unless all you need to do is receive an alert and tap on a button. Managing or configuring multiple elements is still a problem left to an independent dashboard on the browser. This is why most bots on Slack would still take you out of Slack to set up your team, configure options, design workflows etc.

All things considered, Slack (and other instant messaging platforms) are clearly the new browsers for the business world and the way of the future. Did we miss out any use case? Let us know in the comments. Also, do check out .