A whole new world: The exciting new roles AI is creating in customer support

As the customer service space finds new and exciting ways to work with AI, the roles, responsibilities, and career paths that make up support teams are being reshaped.

At Intercom, we firmly believe that AI will turn customer service into a more fulfilling, impactful, and opportunity-filled career. As our AI bot, Fin, begins to handle more of the repetitive queries that fill our team’s day, our CS reps have more time to focus on complex customer issues. This allows them to apply and develop their problem-solving and relationship-building skills, becoming subject matter experts that our customers can consult when they need to.

“Many future customer service roles will focus on maximizing the human-AI partnership – meaning customer service will no longer be a step towards a different career, but one of the most exciting careers out there”

Not only that, but we’re already starting to see a whole new category of roles and responsibilities spring up around AI in customer service, with a focus on ensuring that teams’ AI tools are working as effectively as possible alongside support teams, and offering every customer the best possible experience.

In short, many future customer service roles will work together to maximize the human-AI partnership – meaning customer service will no longer be a step towards a different career, but one of the most exciting careers out there. Here are just a few of the brand new roles already emerging as customer service embraces AI.

Knowledge Manager

As the old saying goes, an AI bot is only as good as the content you feed it. High-quality support content is critical to the success of an AI bot, so prioritizing the management of your knowledge base is a must in an AI-first world.

Support teams are ideally placed to spot gaps, identify inaccuracies, and improve the flow of your support content. Knowledge management within a support team is not new, but it has traditionally been dispersed across the team – with no specific owner and few concrete processes – and depended heavily on institutional knowledge passed from teammate to teammate.

Here are just some of the responsibilities that will be associated with the role:

• Content creation: Filling any identified gaps in support content, keeping existing articles up to date, and creating new content alongside product launches or feature releases.
• Help center management: Monitoring the accuracy of content across the help center, flagging articles that need to be updated, replaced, or removed, and paying attention to the number of views each piece of content is getting to assess its value.
• Conversation analysis: Assessing whether the content the AI bot is providing is truly answering the question the customer is asking, and whether updates are required to the knowledge base.
• Bot performance analytics: Linking the bot’s performance to potential content improvements to better meet customer needs.

Although a knowledge manager will be the dedicated coordinator of knowledge management efforts across the company, knowledge management should be, and often already is, within the remit of everyone on the support team. If you’re a small support team and aren’t in a position to hire a dedicated knowledge manager, there are small steps you can take to begin to build on the knowledge management function within your existing team to prepare for a world of AI-assisted support:

1. Work with your team to identify the questions you need your AI bot to be able to answer, and ensure the content to support those answers is up to date, accurate, and easy to understand. You can start small by focusing on your top 10 FAQs.
2. Encourage support reps to keep track of the questions your AI bot wasn’t able to answer, or that frequently result in customers asking to speak to a human rep. Would a simple content update solve the issue?
3. Are there members of your support team that are particularly skilled at, or who show an interest in, creating content? Enable them to carve out a couple of hours each week to address the gaps and opportunities the rest of the team has flagged.
   Learn as you go! Monitor the results of your work, and continue to update and add to your support content as needed.

After decades of the same style of customer support, where a customer asks a question and a support rep answers it, this new approach will require a slight culture shift. But we believe it will improve the experience of both your team and your customers, maximizing your support reps’ knowledge to benefit the customer more than ever before. Ideally, the first time you answer a question is the last time, as your AI bot will be able to answer the same question any time it’s asked again in the future.

“When you’re struggling with an overflowing inbox, it’s easy to view conversations as items on a checklist to be ticked … but the reality is that customer conversations are just one step on a customer’s larger journey”

We’re enabling this very culture shift on the Intercom Support team using our Snippets feature. Snippets allow our team to quickly add content specifically for our AI bot that improves its answer quality and coverage, but isn’t available within your public-facing support content. It encourages our Support team to think about the customer journey and the challenges they might encounter along the way, and enables them to get ahead of those challenges with high-quality, helpful support content.

Find more tips on optimizing your knowledge base for an AI bot

Conversation Designer

When you’re struggling with an overflowing inbox, it’s easy to view conversations as items on a checklist to be ticked – the customer asks a question, your team answers it, and that’s it. But the reality is that customer conversations are just one step on a customer’s larger journey.

That’s where a conversation designer comes in – their role is to optimize the end-to-end support experience for your customers, spanning bots, automation, and human customer service, and find the obstacles to seamless customer support. The emergence of this role indicates a growing connection between customer support and customer success. Focusing on the customer journey encourages a more holistic and proactive approach to the customer experience, as opposed to the more traditional, reactive type of customer support.

Here are just some of the responsibilities that will be associated with the role:

• UX mapping: For your customer journey to be truly seamless, the user experience must be smooth and intuitive.
• Collecting and interpreting customer feedback: Conversation designers will rely on any and all information they can gather about customer behavior and preferences – whether it comes from customer surveys, conversations, usage metrics, or any other source – to improve the customer experience.
• Problem-solving: Once you’ve identified the obstacles facing customers on their journey, it takes problem-solving skills to figure out a way around them.
• Workflow creation: Figuring out the best way to bring a customer along the path that will resolve their query. This requires in-depth knowledge of the user experience, company processes, and your support platform’s capabilities.

How you can begin to foster this function within your existing team:

1. Nominate support teammates to spot areas where your AI bot is missing opportunities to proactively help your customers to use your product more effectively. For example, is your bot telling your customer that a certain feature is insufficient for their needs without suggesting an alternative?
2. Identify opportunities to implement workflows that further reduce the team’s workload. If a customer asks for a refund, instead of providing an article with instructions, set up an automated workflow to allow them to submit a request right there in the chat.
3. Encourage your team to take note of any patterns in customer activity, and suggest ways that automation and AI could improve the experience.
4. If there is someone on your team who’s particularly interested in this area, nominate them to become an expert on your AI tool and its capabilities. AI bots are still incredibly new, and there are new features and capabilities being added all the time. Stay up to date with what your bot can do so your customers can see the benefits as soon as possible.

At Intercom, we’ve just hired a conversation designer, Fred Walton, to manage our customers’ end-to-end experience. Here are his thoughts on the development of customer service roles:

Conversation Analyst

The wonder of AI bots is their ability to converse in a natural, human way. Our user research shows that AI bots are already surpassing customer expectations, particularly when compared to their stilted, robotic predecessors.

And that’s not the only way AI can help you to drive business improvements. By using AI to analyze customer conversations, you can get in-depth insights into the phrasing, tones, and nuanced product terminology that arise in everyday customer conversations. But when it comes to interpreting these insights, identifying potential improvements, and driving change across the support team and the wider business, you’ll need a conversation analyst.

“Using AI-powered analysis, conversation analysts can surface key customer feedback that will impact every team in your company”

In contrast to the conversation designer, who takes a holistic view of the entire customer journey, the conversation analyst focuses on how your AI tool interprets what your customers are saying and how its responses could be improved. Using AI-powered analysis, conversation analysts can surface key customer feedback that will impact every team in your company.

Here are just some of the responsibilities that will be associated with the role:

• Data analysis: Conversation analysts need to go beyond the numbers to interpret what they mean and derive valuable insights into the way customers convey their problems, and the answers they need to resolve their issues.
• An understanding of natural language processing (NLP): NLP is at the heart of large language models (LLMs). In order to understand the way an AI bot will answer a question, conversation analysts need to develop an in-depth understanding of the way they put those answers together.
• Reporting: The insights gathered by a conversation analyst are invaluable, not only to the support team, but to teams across the business, informing decisions across product, marketing, sales, and more. Reporting these findings in a clear, actionable way is a key skill for a conversation analyst.
• Cross-team collaboration: A conversation analyst must be able to work regularly and effectively with teams across the business, to ensure communication remains open and key improvements are actioned.

How you can begin to foster this function within your existing team:

1. Dedicate some time each week for your team to share interesting issues or patterns they’ve noticed in their customer conversations, and to discuss insights and action points from the AI bot reports.
2. Some people are more geared towards data analysis and interpretation than others. If there are members of your team who have an interest in this side of customer support, analyzing the AI bot’s conversations might be an exciting opportunity to expand their role. Enable these teammates to carve out the time needed to start analyzing a small sample of AI bot conversations, and contribute their thoughts on improvements that could be made.

Prompt Engineer/Problem Formulation Engineer

We’ve all been wowed by ChatGPT’s ability to understand what we’re asking, no matter how awkwardly we might phrase our question. AI bots can provide that “wow” experience right out of the box, but when it comes to company-specific customer queries, it’s important to ensure your chatbot is performing to the highest standards. That’s where a prompt engineer, or problem formulation engineer, comes in.

Prompt Engineer

Labeled an “amazingly high-leverage skill” by OpenAI founder Sam Altman, prompt engineering involves gaining a deep understanding of the way an AI bot answers questions, creating optimized prompts and refining the bot’s responses to achieve the best results. Essentially, they ask strategic questions to produce optimal outcomes – then use those templates to inform future responses.

In the world of customer service, this means training the bot to provide the right answer every time, taking into consideration your company’s specific terminology, and the way your customers phrase their queries to provide a useful answer, and maybe even some follow-up information.

Some say the role of prompt engineering won’t last long – as future AI models are trained against optimized prompts, these roles may become obsolete, or, as The Guardian put it: “In the wider jobs market, prompt engineering will probably go the way of spreadsheet management or search engine optimisation – a skill demanded in a variety of roles and prized by hiring managers as another feather in the cap of your CV.”

Problem Formulation Engineer

While prompt engineering focuses on the workings of a particular AI tool and how you can manipulate it to produce the best results, problem formulation engineering tackles the wider problem areas that exist for your customers.

This role involves identifying and comprehending problem areas, analyzing them, and defining their focus, scope, and boundaries. Developing an in-depth understanding of the problem domain makes the bot-tuning process more effective, and ultimately, delivers a better customer experience. Bots that are trained by a problem formulation engineer to deeply understand the issue a customer is experiencing will be extremely valuable assets to your business – they can suggest not just short-term solutions to that immediate problem, but related improvements that could uplevel their experience within your product.

Here are just some of the responsibilities that will be associated with these roles:

• Understanding your AI tool: AI tools will respond to prompts in different ways, depending on the LLM powering them and the source content they’re drawing from. Establishing the most suitable prompts requires an intimate knowledge of the AI tool you’re using, and the way it answers your customers’ queries.
• Understanding your customers’ most pressing issues and how they communicate them: Your support team will have a thorough understanding of the most common problem areas and the way customers phrase their queries. As a prompt/problem formulation engineer, you need to find a way to transfer this experience to the AI bot.
• Testing and optimizing: Experimentation will be a major part of this role as you test an approach, monitor customer feedback, and tweak to optimize the prompt or problem formulation.
• User research: Quantitative and qualitative user research will inform the experimentation mentioned above, giving direction to the tests you choose to run.

How you can begin to foster this function within your existing team:

1. Encourage your team to not just use your AI tool, but gain a comprehensive understanding of how it works, how it processes your support content, and how it interprets the questions it’s asked.
2. Ask your team to note any instances where customers aren’t getting the answers they need, and flag any patterns that might be solved by some bot-tuning.
3. There are lots of online prompt engineering learning and development opportunities available from companies like Coursera and Udemy. If any teammates are interested in the area, and your company has a learning and development budget, look further into these opportunities to upskill.

Support Design Strategist

This role involves taking a bird’s-eye view of the entire support experience and deciding where AI and humans fit best at each stage of the customer journey.

If you’re a support leader, you’re probably thinking, “I’m already doing that,” and you’re totally right. This work will likely fall under the support leader remit for some time, but once AI is industry-standard, companies will differentiate themselves by the seamlessness of their service delivery, coordinated by an optimal human-AI partnership. That’s when the need for a dedicated support design strategist arises.

Here are just some of the responsibilities that will be associated with the role:

1. Process analysis and improvement: As AI bots become fully integrated with support teams, support design strategists must constantly monitor, update, and overhaul legacy processes to meet rapidly increasing customer expectations.
2. Strategy and planning: Planning ahead may seem like an overwhelming task as AI transforms the world of work, but as support teams get used to working alongside AI, the support design strategist will figure out what’s working, what’s not, and what goals the team should be striving to reach.
3. Resource management: Creating the optimal human-AI support strategy requires the right balance of resources; namely time, team members, tools, and budget.
4. Collaboration with other key roles, like the conversation designer: The support design strategist will require insights from the entire support team to create a truly holistic support strategy.

The best first step you can take towards developing a holistic support design strategy is to learn as much as possible from your team as you guide them through this enormous transition.

How you can begin to foster this function within your existing team:

1. Include the team in decisions being made about their roles, and make time to discuss updates, changes, and experiments in full.
2. Reassure team members worrying about the security of their roles by acknowledging their concerns while getting them excited about the new career and upskilling opportunities AI will bring.
3. Keep the lines of communication wide open and encourage your team to provide honest feedback about how they’re feeling and how their everyday work is being affected by the introduction of AI to the team. If something isn’t working, don’t be afraid to change it – there will be lots of experimentation as AI is incorporated into more and more processes.
4. Welcome suggestions on all aspects of your team’s work, from improving workflow efficiency to automation of team processes.
5. Think about how your team’s performance will be measured in the human-AI support world. With an AI bot handling the repetitive queries that used to take up so much of their time, they can turn their attention to more impactful work and spend more time outside the inbox. How will this be reflected in their goals, career objectives, and performance reviews?

The future of careers in customer service is brighter than ever, and the roles we’ve discussed here are just the beginning. We’re excited to see customer service roles become more and more desirable as AI changes the nature of work in the customer service space and beyond.

Looking for more tips, tricks, and news for support leaders working to stay ahead of the AI curve? Sign up to the Intercom newsletter for fortnightly insights