Chatbot Projects at Public Authorities: the 3 Success Factors

Chatbots are also making their way into the online communication of public authorities, and users are willing to enter into dialogue with the digital helpers. To ensure that they use the digital communication option repeatedly and regularly, some guidelines should be followed when creating and optimizing the bots. 3 factors contribute a lot to the success of a chatbot project.

Chatbots do a great job when it comes to providing users with neutral and fast help by always being available and at the same time relieving employees of standard concerns. But they only do this if they are designed to meet the needs of their users and if the right use cases are selected. This is also shown by the current Chatbot Study 21. To ensure that the expectations of the bot are fulfilled, project teams would do well to pay attention to a few points during its development and further optimization.

As part of my certificate thesis entitled «A chatbot for the IV office of the Canton of Bern: Why and under what conditions the digital helper promotes dialogue with stakeholders», I specifically investigated how the stakeholders of the IV office of the Canton of Bern (IVBE) react to the new digital helper, how they assess the interaction with it, and what is important for them in order for their requirements to be met.

My study results make it clear that the findings and recommendations for action from the Chatbot Study 21, which looks at chatbots in a wide variety of companies, can be transferred to the stakeholders of administrations, in this case to the example of the IV-Stelle Kanton Bern.

The two chatbot studies on the topic of user and employee needs and satisfaction show that 3 important success factors must be taken into account when developing a successful bot, which affect the users on the one hand, but also the company's own employees on the other.

Success factor 1: electronic data analysis AND target group survey when implementing the chatbot.

A well-founded survey and analysis of your own data (website statistics, telephone and email receipts, etc.) are the be-all and end-all for the development of a chatbot that meets the requirements of the stakeholder group. In order to get to the bottom of the needs and sensitivities of your stakeholder group, you need to specifically ask users about them. The answers will reveal possible priority areas of use and characteristics for your chatbot. 

I will go into a little more detail here about some of the evaluations of the research I conducted in connection with the development of a chatbot for IVBE to illustrate what is important to chatbot users. Some of the results are surprising:

Chatbots are on the rise and users are ready for bots and have a positive attitude towards them - even in the administrative environment.

My survey clearly shows that IVBE customers are very open to using a chatbot and have no fear of contact. As many as 67.5 percent of respondents answered «yes» or «rather yes» to the question of whether they would be willing to interact with a chatbot. This is a significantly higher percentage than in the Chatbot Study 21, where 44 percent were in favor of chatbot use. 

The respondents have no concerns about data privacy. This is an important finding for administrative communications in particular.

The respondents trust IVBE with regard to data security. This is shown by the approval rate of 73 percent. This is 13 percent higher than in the chatbot study 21 (60 percent). 

The question of whether they are willing to provide personal user information during a chatbot interaction in order for the problem to be solved was answered in the affirmative by 62 percent of respondents. This is 12 percent less than in chatbot study 21. 

The fact that the bot knows the respondents and can access their user history is rated positively by 37 percent. The respondents are somewhat more reserved on this point than in Chatbot Study 21, where the figure was 52 percent.

The desire for self-service portals is increasing. The priority areas of application for which respondents can imagine a chatbot confirm the general expectation that government business should be conducted online wherever possible.

Success factor 2: Talk to exponents personally after implementation as well

After the chatbot is implemented, its development is not yet complete. Conduct so-called «sounding boards» not only before, but also during and after any chatbot implementation with exponents of the stakeholders who interact with your chatbot. These provide valuable additional information for the further optimization of your chatbot - also with regard to possible deviations in self-perception and perception by others. In the best case, through honest exchange and genuine interest in the needs of your stakeholders, you turn pure information recipients into communication partners and ambassadors for your company.

In our specific case, we initially asked ourselves whether it wasn't absurd, in line with our self-image as «people for people» working at the IV office of the Canton of Bern, to use a robot, of all things, to promote dialog with stakeholders and achieve greater customer proximity. And whether the use of a digital helper is more likely to damage the reputation, which is already not in the best of shape otherwise? The conversations with policyholders (in our case, people with disabilities) yielded illuminating insights and reveal even more plus points of a chatbot as a dialog medium:

• Chatbots are accessible for all people with various disabilities. A strong pro-argument, not only for the IV.
• No one contacts the IV voluntarily. A chatbot would lower the inhibition threshold to obtain information and make contact anonymously for the first time.
• Especially young insured persons and people with mental illnesses see a chatbot as an attractive dialogue channel. A chatbot could make it easier for IV (and others) to access these stakeholder groups.
• Society (and therefore our stakeholders) are already very «Messenger-savvy». Short questions and answers, the principle on which a chatbot also functions, fit in with today's times. This would benefit the image of IV (and probably other companies as well).

In summary, it can be said that a chatbot cannot and will not replace personal consultation - but perhaps it can prepare it.

The results of the survey underline this. With a positive share of over 82 percent, the respondents consider the preparation of a consultation by a chatbot to be rather useful (32.4 percent), useful (43.2 percent) or even extremely useful (6.8 percent). This is more than 20 percent higher than in the Chatbot Study 21. On the other hand, the respondents can hardly imagine a complete consultation by a chatbot, with an approval rate of just under 15 percent.

Success factor 3: Don't forget your own employees. Get them on board early on

Not only the users, but also the company's own employees should definitely be taken into account during bot development. A chatbot must never be a project of the IT or communications department alone. For a successful implementation, all departments on which a chatbot has an impact must be on board from the very beginning. In most cases, information and persuasion work must be carried out internally beforehand, as experience shows that the level of knowledge regarding «What is a chatbot anyway?», «What can it do?" and "What can it not do?» varies greatly among the company's own staff.

If the employees see a clear benefit for the company and for themselves - especially with regard to the increase in automation and thus efficiency - they are prepared to release personnel resources for the realization and subsequent management of a chatbot. At IVBE, it was precisely the prospect of the above-mentioned resource-consuming preparations for a consultation being handled by a chatbot and the employees being able to devote themselves again more to the exciting work that requires their real know-how that was very motivating. In most cases, promising synergies with various (internal and external) departments also arise in this context, and further fields of action emerge that require corresponding joint preparatory work.

Conclusion

Consider the input of your stakeholders when planning and developing your chatbot so that you can tailor it to their needs. and be sure to collect user feedback after implementation to further optimize the digital channel.

Accompany your chatbot project with communication throughout all phases. Ideally, you should inform and integrate employees in advance and communicate actively and regularly about the project status during implementation. It is also crucial for the success of the project to communicate the launch widely to the stakeholders and to support it. Afterwards, they must be continuously informed about further developments and actively involved in adjustments. 

This way, your own chatbot enthusiasm will quickly and sustainably spread to others and the project will be supported by everyone.