Bot VS Applications: pros and cons. Case of the bot accountant
Working with the most complicated tax applications for accountants is not new for our team. But the market dictates its own rules, and one day our partners came up with the idea of creating a chatbot that would completely replace reporting services for entrepreneurs and allow POPs to pay taxes and submit reports without delving into accounting nuances and without filling out tax forms. In other words, the entrepreneur simply answers the questions, and the bot, on the basis of the answers, independently generates and submits to the state bodies all reports.
As a result, in May of this year we launched the entrepreneur bot accountant market. We tell how it was.
In favor of the bot, we had several arguments.
Firstly, its accessibility - the user works in a messenger familiar to him, has the ability to view the interaction history, he does not need to download and install anything.
Secondly, ease of use and error insurance. The format of the bot allows you to control the actions of the user in the product: literally "lead him by the hand" through all the tax jungle. The user simply cannot proceed to the next step without going through the previous one, which avoids many errors.
Thirdly, the communication format in the bot is only textual. In the case of taxes, this is even good. I really wanted to save entrepreneurs from direct contact with frightening tax forms and forms, while changing the paradigm of consumption.
In addition - for us it was a professional challenge - no one has done this before.
At the bot design stage, the question arose - to create a bot based on commercial platforms for developing bots, or to develop your own platform. The advantages of commercial platforms are obvious - they are easier to launch and use, dialogs can be edited without the participation of developers. But we understood that the capabilities of commercial platforms could limit us in some solutions and functionality.
As a result, we settled on our own development. Perhaps this was a more difficult path, but, in our opinion, more correct from a strategic perspective.
In practice, this meant for us that the slightest changes in the bot and changes in the dialogs (and this process is fascinating, but endless) are hardcoded by the developers. This is a very distracting resource, which could be directed to real development.
As a result, we separated the release of small features (including minor changes to dialogs) and full-fledged new functionality - with a clear understanding of the resources that can be allocated for this. But pretty soon they came to the understanding that creating a simple file, possibly a gui interface, so that analytics and marketers could edit the dialogs anyway, would be necessary in the future.
How is this solved in applications? Most changes to the product by default are made by developers. And from this point of view, the speed of creating and making changes to dialogs in the bot is an important advantage for developers, which translates into the ability to respond more quickly to user requests.
Work on the bot was a challenge for us in terms of building interaction with the user through dialogs, rather than a graphical interface. After all, the textual format of interaction limits some possibilities, and far from all ideas can be implemented in a bot. In addition, many simply more convenient and familiar to work with the graphical interface - it is more informative and enjoyable for them.
But it was precisely the limitations of the text format that led us to strive to maximize the simplification of the client path. As a result, if using standard reporting services it takes about 10 minutes to prepare and submit reports, then in the bot you can pay taxes and submit a report three times faster.
If with other products in our arsenal we can always track the sources of visits and conversions and analyze the user’s path, then when working with the chatbot it’s a little more difficult to track the effectiveness of any actions. So far, the only thing we can see is the step at which the user who has already started working with the bot stops.
For the first 3 months of work, about 15,000 people entered our bot accountant. But conversion, i.e. real payments amounted to only 5%.
The inability to see and analyze the effectiveness of certain actions, in the end, led us to the only logical step - we created a landing page for the bot, which not only details the capabilities of the bot, but also lends itself to analysis. But in order to more deeply understand how the user perceives the bot, we came to the need for phoning and personal contact with users.
What did we find out?
Despite all the advantages of such a bot, we had to deal with the most powerful of the factors - with the habit. At this stage, the users of the bot are mainly IT, retail and advertisers - those categories of consumers who are traditionally not afraid to try something new. But on the other hand, we were very inspired by the cool reviews of those users who nevertheless trusted the chatbot and now pay taxes with its help.
In general, the experience with the bot showed striking differences compared to traditional applications. But most of the shortcomings that we encounter can be offset by a competent approach. And we are still learning to use the advantages of the bot to the maximum.
In the near future, we set ourselves the following goals:
Already now we can replace the accountant for 70 - 80% of FLP of the 2nd and 3rd groups, but at this stage there are additional wishes from users and partners who always want more. What are we, frankly, happy.
As a result, in May of this year we launched the entrepreneur bot accountant market. We tell how it was.
Why exactly the bot?
In favor of the bot, we had several arguments.
Firstly, its accessibility - the user works in a messenger familiar to him, has the ability to view the interaction history, he does not need to download and install anything.
Secondly, ease of use and error insurance. The format of the bot allows you to control the actions of the user in the product: literally "lead him by the hand" through all the tax jungle. The user simply cannot proceed to the next step without going through the previous one, which avoids many errors.
Thirdly, the communication format in the bot is only textual. In the case of taxes, this is even good. I really wanted to save entrepreneurs from direct contact with frightening tax forms and forms, while changing the paradigm of consumption.
In addition - for us it was a professional challenge - no one has done this before.
The invention of the bicycle - the pros and cons
At the bot design stage, the question arose - to create a bot based on commercial platforms for developing bots, or to develop your own platform. The advantages of commercial platforms are obvious - they are easier to launch and use, dialogs can be edited without the participation of developers. But we understood that the capabilities of commercial platforms could limit us in some solutions and functionality.
As a result, we settled on our own development. Perhaps this was a more difficult path, but, in our opinion, more correct from a strategic perspective.
In practice, this meant for us that the slightest changes in the bot and changes in the dialogs (and this process is fascinating, but endless) are hardcoded by the developers. This is a very distracting resource, which could be directed to real development.
As a result, we separated the release of small features (including minor changes to dialogs) and full-fledged new functionality - with a clear understanding of the resources that can be allocated for this. But pretty soon they came to the understanding that creating a simple file, possibly a gui interface, so that analytics and marketers could edit the dialogs anyway, would be necessary in the future.
How is this solved in applications? Most changes to the product by default are made by developers. And from this point of view, the speed of creating and making changes to dialogs in the bot is an important advantage for developers, which translates into the ability to respond more quickly to user requests.
Graphics or texts?
Work on the bot was a challenge for us in terms of building interaction with the user through dialogs, rather than a graphical interface. After all, the textual format of interaction limits some possibilities, and far from all ideas can be implemented in a bot. In addition, many simply more convenient and familiar to work with the graphical interface - it is more informative and enjoyable for them.
But it was precisely the limitations of the text format that led us to strive to maximize the simplification of the client path. As a result, if using standard reporting services it takes about 10 minutes to prepare and submit reports, then in the bot you can pay taxes and submit a report three times faster.
Difficulty analysis
If with other products in our arsenal we can always track the sources of visits and conversions and analyze the user’s path, then when working with the chatbot it’s a little more difficult to track the effectiveness of any actions. So far, the only thing we can see is the step at which the user who has already started working with the bot stops.
For the first 3 months of work, about 15,000 people entered our bot accountant. But conversion, i.e. real payments amounted to only 5%.
The inability to see and analyze the effectiveness of certain actions, in the end, led us to the only logical step - we created a landing page for the bot, which not only details the capabilities of the bot, but also lends itself to analysis. But in order to more deeply understand how the user perceives the bot, we came to the need for phoning and personal contact with users.
What did we find out?
- Many users entered the bot out of curiosity;
- Some could not understand the principles of the Telegram. For example, where we suggested choosing one of the two buttons, they answered - typing. We happened to read outraged comments about why the bot does not understand us.
- Since there was nothing similar on the market before, a certain distrust of its work remains.
Despite all the advantages of such a bot, we had to deal with the most powerful of the factors - with the habit. At this stage, the users of the bot are mainly IT, retail and advertisers - those categories of consumers who are traditionally not afraid to try something new. But on the other hand, we were very inspired by the cool reviews of those users who nevertheless trusted the chatbot and now pay taxes with its help.
Where are we going next?
In general, the experience with the bot showed striking differences compared to traditional applications. But most of the shortcomings that we encounter can be offset by a competent approach. And we are still learning to use the advantages of the bot to the maximum.
In the near future, we set ourselves the following goals:
- continuous improvement of dialogs in order to strike a balance between the simplicity, information content and emotional appeal of the text and flow;
- closing all the needs of FLP of the 2nd and 3rd groups in terms of accounting services;
- the development of such communication mechanisms with the user in the bot itself that would not cause a spam sensation, but at the same time be effective;
- creation of the maximum convenient mechanism for consulting (providing advice in the interface of the bot itself in the future) as a necessary element for building user confidence in the product;
- creation of interfaces for making quick changes to dialogs without the participation of developers.
- integration with other services.
Already now we can replace the accountant for 70 - 80% of FLP of the 2nd and 3rd groups, but at this stage there are additional wishes from users and partners who always want more. What are we, frankly, happy.
All Articles