8 ways to build a better business bot

A wise person should have money in their head, but not in their heart. — Jonathan Swift

As we enter the second year of the bot industry, entrepreneurs and investors are wondering about how bots will generate revenue and become sustainable businesses. In this article I will review the direct and indirect ways that bots can drive revenue to a business. It is important to note that this is far from a comprehensive list, as I am sure entrepreneurs will come up with many other ways to make money in this industry.

1. Subscription

The subscription business model is currently the most common way that bots for work drive revenue. In this model, the bot provides an ongoing service the user subscribes to and pays for.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":2130680,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,","session":"B"}']

An example of a subscription-based bot is Growbot — a bot that helps teams celebrate professional wins in Slack, promoting great team spirit and cohesion. The bot looks for keywords like kudos and collects and amplifies these congratulations.

Most of the features provided by Growbot are free — the company does not charge the user to install the bot for using its core features. The Growbot team deliberately did not want to charge for aspects of the bot that promote usage, engagement, and virality. It did, however, carve out a set of features that are important to big companies and HR departments. Growbot charges for a premium service that includes company data dashboards and values integration.

Growbot asks users to pick a plan during setup.

Here is how it highlights its premium functionality on the web:

Growbot also lets users upgrade their bot at a later stage, enabling a use case where the team works with the bot for a while and then decides to upgrade.

Another example of a subscription-based bot would be Statsbot. Statsbot is an analytics bot that exposes stats from Google analytics and Mixpanel. Statsbot decided to go with a trial model in which they provide you with a fully functional premium bot for a trial period. An interesting strategy Statsbot applied is to let the bot itself prompt the user for payment.

[aditude-amp id="medium1" targeting='{"env":"staging","page_type":"article","post_id":2130680,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,","session":"B"}']

As you can see, both bots drive the user to set their subscription on the web. This is because none of the chat platforms currently support in-platform subscription payments. Driving the user out of the conversation and into a web interface might affect the conversation but currently, that is the only option available for subscription bots.

Once the payment has been set up, it is important to acknowledge it in the chat interface. Notifying the user of upgrade, downgrade, and recurring payments is a good way to promote transparency and user satisfaction.

The subscription model, when done right, is very lucrative because the lifetime value (amount of money the user will pay in the lifetime of using your product) of the user is usually higher than other models.

[aditude-amp id="medium2" targeting='{"env":"staging","page_type":"article","post_id":2130680,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,","session":"B"}']

2. Ad-serving

Ad serving has been the bread and butter of many web and mobile businesses. Bots are in a unique position when it comes to ads, as they can build a personal relationship with the user, as well as collect a lot of personal information that can contribute to more personal and fine-tuned ads that lead to better clickthrough rate.

Here’s how it looks on the Kik platform:

As you can see in this example, the bot represents a teen influencer bot on the Kik platform, and the user actually asks for the ad as part of the conversation. Most advertisers will tell you that a user asking for an ad is as good as it gets when it comes to conversion to paying users. The bot captures the intent at exactly the right moment and serves a relevant ad to the user.

[aditude-amp id="medium3" targeting='{"env":"staging","page_type":"article","post_id":2130680,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,","session":"B"}']

I talked to Andy Mauro, cofounder and CEO at Automat, who built the bot. He stressed that a huge motivator was to both provide an engaging fan experience and to deliver value to the brand — and to do so in an authentic way that is more conversational and personal than traditional advertising. Using chatbots, you can add conversion opportunities (in their case a coupon to start) that make it easier to demonstrate real value and sales. Andy added a sentiment question in the bot which, as of today, has averaged out to 88 percent loved it or liked the bot.

Note that while some platforms are happy with bots serving ads, other platforms do not allow it. Slack, for example, does not permit the serving of ads on its platform because it’s a business communication platform. It would not make sense to serve ads in a business tool context, especially one that that promotes productivity and focus.

3. Data for analytics and market research

Bots can collect a great deal of data from users. Through engaging in a conversation or playing a game, bots can learn about user preferences and interests.

An example of such data collection is the Swelly bot. Swelly lets users play a game of choosing between two options. Once the users chooses an option, they are rewarded by being shown their cohort results (creating social hook to re-engage with the bot) and get prompted for another choice.

[aditude-amp id="medium4" targeting='{"env":"staging","page_type":"article","post_id":2130680,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,","session":"B"}']

Here is how it looks on Facebook Messenger:

Now imagine you were a fast food brand, for example, and wanted to know which type of sandwich customers would love or what type of fast food imagery they would find more appealing. This type of bot would be a gold mine for your research. By collecting this type of audience preferences, the Swelly bot can generate a detailed report in a very rapid and precise way. While this bot does not charge the end user, I am sure it will one day make a lot of money helping brands understand their audience.

4. Selling goods and services

Bots can also become the channel in which you sell good and services. These can be tangible goods the bot sells directly in chat or a paid service that the bot exposes through chat.

[aditude-amp id="medium5" targeting='{"env":"staging","page_type":"article","post_id":2130680,"post_type":"guest","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"bots,business,","session":"B"}']

An interesting example of a paid service exposed via chat are the ride services, Uber and Lyft, that developed interfaces for both Slack and Facebook Messenger:

I find these example very interesting because these are services that were available on mobile and consumed mainly in a form of a mobile app. The conversational interface lets you consume these services and order a ride without having the app installed on your phone. As the bot market grows, we might see more and more paid services move to the chat platforms.

You cannot talk about goods sold via bot without mentioning Amazon Echo. Since we introduced Alexa to our house and started using it in our kitchen, more and more of our household goods are being ordered through it. “Alexa, add napkins to the shopping list” is easy, frictionless, and catches the intent, at the right time, with a bot that can address that intent.

5. Referral fees

This is another major business model on mobile and web that is moving to bots — a bot that can help you decide what to buy or which service to consume and then refer you to the right service, rather than actually completing the transaction itself.

A good example of a referral business model in bots is the Kip bot. The Kip bot, available on Slack, Messenger, Kik, and Telegram is a shopping assistant for teams. Kip helps you find what to buy and then refers you to the merchant that can fulfill the order.

The advantage of Kip is that it is not bound to a single vendor; it can potentially serve as a cross merchant and provides the user with results sourced from multiple merchants. When the user completes the transaction on the merchant’s site, Kip gets a referral fee for sending the user to that vendor.

There are many other use cases like this in our lives that follow the same pattern, from travel agent to a car salesperson. We will be seeing a lot of agents like Kip that will help us choose between vendors.

6. Brand promotion

Another indirect way to drive your business through a bot is by promoting brand recognition. Thinking about the bot as a front-end representative to your product and service can make a lot of sense, and having a delightful bot that provides a useful service can generate strong attachment to your brand.

This pattern is explored by many marketing managers. I think the key is to explore the right value proposition the bot can provide as well as the persona it has, both of which will reflect heavily on the brand perception customers will develop while engaging with the bot.

Here is an example of a branded bot on Kik:

As you can see, the bot clearly represents the H&M brand. It uses a young and friendly persona (note the use of emojis and its casual tone) but also has emphasis on providing value to the user, in this case searching for clothes very early in the conversation. Providing value is key, because a bot that does not do that might actually generate a negative brand association, very similar to a company representative that keeps wasting your time.

7. Extending a paid-for product

Similar to mobile apps, extending a paid product or service through a bot is a very common pattern for companies who already have an established offering and want to extend this offering to another user experience.

Trello is a well-established paid service that provides task and project management. Noticing that a lot of its clients are using Slack as a way to communicate, Trello decided to extend its service to that new platform and provided parts of its functionality inline in the chat interface.

As you can see, the interaction is contextual and actionable. It makes a lot of sense for users to interact with their service in this intuitive and collaborative way. Choosing which part of the functionality to expose in chat is key to the success of use cases like this. If you expose too much you might provide a cumbersome experience, and if you provide too little it might be not valuable enough.

Why is extending to new platforms useful for your business model? The first reason is that it improves user engagement and the usefulness of your service — users have yet another touchpoint to get value from the product. The second reason is that it improves your product defensibility by mitigating the risk of a competing product gaining traction on this new platform and stealing your clients.

8. In-bot virtual goods

In-bot virtual goods is another model that I think is going to be big in the coming years. An example would be a game bot that sells you chips or powerups, or a bot that sells you music and photos. While I could not find good examples of in-bot payment in real life today, this model has been one of the most lucrative models in the mobile app industry.

When should you start charging users?

I interviewed several bot builders and they all came with the same answer: You should only start considering charging for your bot when you feel that you have reached product market fit. Users need to get hooked on your product, use it on a regular basis, and derive a lot of value from it before you ask them to pay. Whether you are providing a trial, freemium, or in-bot virtual goods, make sure your users love your product before you try to charge them.

Another point many bot providers have stressed is that it is important to decide what to charge for. Charging for features that promote engagement and virality of your product might be counterproductive.

When I joined Google eight years ago, we all thought that paid apps were going to be the way developers would make money out of mobile. What happened in reality is that in-app purchases became the way users preferred to spend their money. The point is that this industry is very young, and we are still exploring the right monetization strategies. You should explore these and other strategies for your particular use case and experiment until you find what works.

Let’s get one thing clear: A bot is a type of user experience and a way to expose products, services, or a brand. The only way to make money out of bots, without having a service or a product, is to be a bot builder and have someone pay you to build that bot. End users do not pay for bots — they pay for the services they expose. They pay for the product the bot promotes, and they are influenced to connect to the brand which it represents. But still, in the same way we have companies that called themselves mobile businesses, making most of their revenue through mobile, we are going to see more and more companies calling themselves bot companies.

Bots will not only become a way to generate revenue, they will be taking big chunks of legacy web and mobile business and become big business of their own. What I usually tell investors who ask me about this market is, “If I do my job right, the next Slack-sized startup is going to be built on the chat platform.”