Roles & skills for successful bot development
If your company is embarking on their Conversational AI journey, one of the first questions you are likely to ask yourself is how to set up a team that can successfully build, test, integrate and deploy bots.
Since Conversational AI is slowly starting to consolidate itself as a professional area (if you are still sceptic about this, take a look at the growing number of job offerings here), there is a lot of information lying around that can sometimes be overwhelming and a tad confusing.
Fear not, this article gives a brief overview of the essential roles and skillset required for successfully building and launching a Conversational AI project.
To paint the picture a bit clearer, the different roles will be explained aligned to the different phases of a project. (Spoiler alert: there’s no need to have big teams right from the start). For each role, we’ll briefly describe their main functions within the team and the skillset required to cover that role.
Governance & Planning Phase
Even though management is not a project phase per se, it’s still essential for successful project delivery on time. This brings us to the first essential role to cover these tasks: The Project Manager/Scrum Master.
Project Manager/Scrum Master
This is the role in charge of overseeing project completion, task and sprint planning and coordinating the team’s efforts to deliver the project. This role is also in charge of reporting the status of the project to ensure all stakeholders have visibility. Although at early stages of project development this role is covered by one person, these responsibilities could be also split between two different people: one in charge of the project governance as a whole and the other responsible for managing sprints. In terms of profile and required skills, the person to take on this role should have experience with Agile/Scrum methodologies, project management, have some degree of knowledge of Conversational AI as well as of the Business processes and technological setups.
Design Phase
This phase is where the blueprint of the bot is created. A mixture of technical and business skills is required. The main roles involved in this phase are Conversational Designers, Business Analysts, and Technical Architects.
Conversational Designer
With a strong focus on user experience, this role should be devoted to defining the bot’s persona and personality, the flows and conversation structure and maybe even the look & feel of the bot to make sure it stays on brand and interacts smoothly with users.
People working in this positions usually come from Linguistics, UX Design, Communication and Marketing. In small projects, the person on this role usually acts as well as a Conversational Copywriter (See next section for more on this role).
Business Analyst
This role works very closely to the Conversational Designers. They are responsible for defining the business requirements and making sure the use cases the bot is handling are correctly addressing the business needs, and therefore, this role is generally covered by someone who has a profound knowledge of the business.
It is often assumed by the member of the company’s department that originated the use case. This could be Marketing Analysts , Customer Service Agents, Customer Experience Managers, etc.
Technical Architect
This role is in charge of making sure that the architecture and bot hosting infrastructure is in place and that all the security and network requirements are met. The person covering this role needs to have a very good understanding of the technical set up as the final project is likely to be integrated with existent systems. More technical in nature, this role is generally covered by a member of the IT department (could be a backend developer or DevOps engineer).
Development Phase
This phase is where the team “gets down to business” and starts building the flows , training the Machine Learning model and building the integrations with the company’s backend systems. In early-stage projects when teams are small, the roles associated with the phase are usually delegated into one single person.
Conversational AI Developer
The person occupying this role will spend their time picking up the stories defined by Conversational Designers and use Teneo Studio to build them into the flows. This role is also in charge of assisting with the configuration of entities, listeners, and conditions amongst other tasks. This role is usually covered by Computational Linguists, AI Software Engineers and Developers.
Conversational Copywriter
The person behind this role typically comes from a Marketing background and is the creative mind that chooses the words and expressions that the bot will use. It is their responsibility to make sure the conversation script stays on brand and helps deliver a good user experience.
AI Trainer
This is the role in charge of developing, training, and optimizing the Machine Learning model behind the bot. People covering this role usually come from Tech-related and/or linguistics-related backgrounds (i.e. Computational Linguists, Machine Learning Engineer, Data Scientists).
Technical Developer
This role must have experience in the company’s technologies and the APIs to which an integration is needed since they oversee coding and building whatever integration is required. Technical consultants using Teneo must be versed in Java and Groovy.
Quality & Optimization Phase
In this phase the bot’s performance is analyzed with the purpose of enhancing it. The main role involved here is Quality Assurance Engineer/Tester
Quality Assurance Engineer/ Tester
This role is in charge of developing and implementing tests during development and before going-live to ensure the bot meets the defined quality standards.
People occupying these roles must have some experience with testing and IT Quality assurance processes. There are some companies which integrate this role’s responsibilities into those of Conversational AI Developers explained in the section before, since this is a function deriving directly from the developers’ work.
Bot Analyst
This role mostly analyzes and measures the bot’s performance, user interactions and the overall quality of the project. It is also responsible for the confection and distribution of reports and dashboards to the different stakeholders. The desired profile is a person with experience in data analysis and reporting tools.
In small teams, the role of Quality Assurance Engineer and Bot Analyst can be delegated into the Conversational AI Developer role described above .
Complementary Roles
Since Conversational AI projects have their own unique nature, they may require additional specific roles to make sure they are successful.
Those projects that involve expansion to new markets and territories may require the figure of a Localizer, i.e. someone who can translate and adapt the bot’s content to each language. There are some projects that may call for Frontend Developers to make sure the bot is correctly integrated into a Mobile App or Website as well. Finally, projects that involve channels using voice call for specialists in Automatic Speech Recognition that can make sure the proper Text to Speech and Speech to Text technologies are in place.
Working in Teams with Teneo
Making all these roles collaborate and work coordinately is very simple in Teneo. Apart from being able to define user permits, Teneo’s version control features allow users to easily keep track of changes made to projects. Using Version Flags, teams can work collaborate on the same project while separating responsibilities.
Final Considerations
Even though there are several roles to cover…don’t panic! Your company doesn’t need to hire one person per role to successfully deploy your bot.
During early stages of development for a project, it is common to have more than one role centralized within the same person: Usually the person acting as Conversational AI Developer tends to perform QA Testing and the same person analyzing use case designs conversations.
As projects become more complex and companies increase their confidence and experience in Conversational AI, so does the need for more specialized roles. This often translates into an increase in the amount of team members.
If hiring the service of a company for development (For example, an Artificial Solutions Partner), roles tend to mix between different parties: Developing and QA Testing might lay on the contractor side while hiring companies tend to keep the Conversational Design aspect.
The team creating and deploying a bot is not the sole responsible for a successful delivery of the project; other departments like Legal and Compliance, IT and Marketing are always involved and consulted during the different development stages.
Conclusions
As previously explained, there is no single formula to create a team. Each company should adapt their team size and skillset to the requirements and demands of their projects making sure essential roles like project management, conversational design er (absorbing other functions and roles as needed), and technical development are always covered.
Now we’d love to hear about your team. Are you covering these roles within your team? How big is your team? Let us know in your comments below!