If you want to build a new structure, the smart move is to find the best existing examples of that structure, study them, understand why they work well, and use those insights to inform your own building plans. If you do it right, you end up with something new, effective, and inspired by the past. If you get it wrong? Well, I guess that’s how Las Vegas ended up with a pyramid covered in a giant logo for a bad beer.
If you are building a support department, basing that on the structure of the best performing support teams is also a smart move. Or it was — until now. The arrival of generative AI and machine learning tools is set to change the assumptions all those incredible support teams were built under.
No doubt most of those teams will adapt and excel in the era of AI, but you don’t need to wait for them to show you the way. Let’s explore how support departments will be shaped and directed by the capabilities, costs, and competition represented by artificial intelligence.
Why AI will reshape support team structures
Technology never rests. Every new year brings new tools, new networks, and new channels that customer service teams are expected to support, integrate, or — at the very least — understand.
Most of the time, the impact of those changes can be handled by the existing organizational model, perhaps with some training and shuffling of resources. Even after the explosive growth of the mobile internet, social media, traditional chatbots, and messaging, a best-in-class software support team from 2002 looks pretty similar in structure and scale to 2023’s best-in-class example. Support professionals have always absorbed, adapted, and delivered.
The uptake of machine learning and generative AI looks very different from those earlier changes. We expect it to reshape the ways in which best-practice customer support teams are built, how they work, and how they scale.
AI will generate pressure to change in the following ways:
AI will create new financial pressures (and opportunities)
Hiring people is expensive, and most support teams could automate at least a portion of their workload through AI much more cheaply than growing their team. Even for companies that truly value human-powered service, there will be pressure to make savings wherever possible and to reserve people for the most valuable work.
AI will also create new forms of financial leverage, opportunities to greatly amplify human skills to create more business value with their time. Support agents assisted by AI-generated suggestions, processes, and data will be more productive and more effective in helping customers achieve their goals..
As the AI marketplace matures, it will become clearer where applied AI saves money, where it creates a demand for new and different services, and where it is too expensive or unhelpful to use. The true cost of those tools may not be known until the end of the initial rush to build market share. However it shakes out, money is always a driver of change, and AI is going to change financial models.
AI will add new customer service capabilities
AI tech may allow teams to offer services that were not possible (or feasible) before.
If more of your customers are able to self-serve via AI tools, does that create space for you to offer one-on-one customer success calls? Will AI translation be good enough to let you help your customers in their own first language? Could AI uncover customer insights in your support inbox that individual people would not spot?
The jigsaw puzzle of jobs to be done and resources to work with might be entirely re-scrambled by AI tools.
AI will be table stakes for support teams
Given the financial incentives, it’s likely most of your competitors will enhance their customer service offerings with some form of AI. Once AI becomes a standard tool, it will quickly become necessary to make it part of your mix. Yes, there are still plenty of craftspeople using only hand tools to produce beautiful work, but every home building contractor uses power tools or they’d never make any money.
If you wish to differentiate your service, that must come (in part) from the specific ways those tools are put to use for your customers.
AI will change onboarding and hiring constraints
Early research suggests that AI-suggested drafts in customer service have much more impact on new support agents than they do on more experienced staff. New agents can more easily ramp up their competence with the assistance of AI suggestions, whereas those who have already developed the skills and knowledge see smaller improvements to their speed and competency.
If it becomes significantly quicker and cheaper to scale up a team, companies may find they can move people in and out of the queue more often than before. Outsourcing to cover spikes in support volume may be more appealing if the quality and costs of bringing on new team members are more manageable.
Internal staff may also be more easily cross-trained on different products or roles, allowing for potentially smaller, more nimble teams.
All of these factors mean that the way we choose to build, grow, and deploy customer service teams is likely to change significantly. Predicting any specific changes takes a little educated guesswork; it’s a good thing I’m a little educated.
Predicting the impact of AI on customer support departments
It’s time to crack out your crystal ball, consult Dr. Strange, or search for a seller of mysteriously specific fortune cookies. We’re going to try to figure out what the future looks like for customer support teams.
Individual support professionals have their own paths to navigate, but support managers and leaders need to steer a whole department. Which direction is safe, and which is risky? Here are some changes to expect.
New teams will include AI as standard
Founders and directors who are setting up their first customer service systems will assume that AI tools, whether directly responding to customers or working in the background, will be a core element rather than an add-on to an existing structure.
That shift in attitude will affect their hiring and training plans, their tagging and reporting choices, their quality metrics, and more. Starting from a clean slate will give those teams a degree of advantage over those of us trying to retrofit processes to include AI.
What they do not have is a set of high-quality, tried and tested answers that their AI can be trained on. They will still need the mind and skills of customer-centric, thoughtful people to set that standard.
Support teams will need new AI skills
As generative AI becomes common in support, the management of AI tools will be taken on by the support teams themselves (perhaps under a support operations role.) Tweaking the models and the prompts, managing AI-driven workflows, and setting and measuring AI quality standards will all require the combination of customer service and AI knowledge.
This represents a new opportunity for career growth, a more technical role which will create scalable impact and may replace one or more frontline support roles in some teams.
The cost/benefit ratio of support will change
If AI capabilities continue to improve and it is able to resolve enough customer requests with high-quality responses, then the cost-per-interaction of customer service will come down.
Companies will have a choice whether to reinvest that money into higher-touch service options, such as one-to-one account management, personal buyers, and dedicated support lines, or to take the savings and invest elsewhere in the business. The return on investment that a support team creates will need to be recalculated.
Support leaders should be prepared to make their own cases for how that changing cost should be best invested. Is it time to set up a new onboarding team for your largest customers? Should you create a paid services tier for the people who want additional help? Perhaps it is time to refocus some of your team on scalable customer enablement work.
Warm up your pitch decks!
Fewer support staff will serve more customers
Support workloads are driven by the contact rate: the percentage of active customers who request human help in a given month. As AI-enabled self-service options handle more of those requests, the rate will decrease.
As knowledge bases, chatbots, and other help desk tools in the past have done, AI tools will amplify the efforts of the support agents, enabling them to serve more people more quickly.
However, we should expect that the requests that AI self-service cannot answer will be more complex, and that may mean there is more of a shift in the type of work than in the overall volume. Support folks will serve fewer customers but more deeply and with greater impact on customer success and loyalty.
New queue triage and management processes will arise
Machine learning is a powerful tool for understanding, categorizing, and reporting on incoming customer support requests. Rather than having staff inconsistently tagging and sorting, expect to see more AI-driven tagging, prioritization, and distribution of workloads.
Your teams will be able to spend more time working on the most important questions and less time worrying about missing a fire in the queue.
Self-service will be more expensive (but more important)
For AI to provide accurate, helpful answers to your customers, nothing is more important than high-quality, accessible documentation. While AI models still struggle with hallucinations, tools are becoming available to restrict it to using known, quality information you provide.
It will be more expensive for AI to answer customers based on your own knowledge base than for customers to find documentation on their own, but it will still be significantly cheaper than that same person writing into your human support team.
Companies who have invested in creating and maintaining really great self-service documentation have a significant head start in 2024. The best time to audit and upgrade your documentation is right now.
Support tier staffing will be rebalanced
The lines between tier 1 basic support and higher tiers will need to be redrawn as some portion of the workload is handled via AI-driven self-service answers or draft suggestions.
With fewer people needed on tier 1, existing teams may find they can shift people into more documentation, customer-enablement, support engineering, and customer success roles or that they have the ability to service more support channels.
For larger companies, a key decision will be deciding when and how to escalate contacts to a human.
A wider range of people could be hired into support
Historically, the time and skill required to onboard new customer support staff, particularly in more technically complex roles, has restricted the pool of candidates.
AI is already showing promise as an effective real-time assistant to support teams, creating suggestions and drafts based on the whole corpus of your previous support answers.
That type of AI assistance can build confidence and reduce the time to competency, which may open up those roles to people who may not have previously qualified. These folks might include parents who have had a career break for childcare or people who grew up without free access to technology.
Smart companies will be able to identify future leaders and high-value contributors and bring them on board with less risk and at less cost than before.
Quality assurance will be expanded and elevated
When draft answers — or entire customer interactions — are led by generative AI, there must be a new system for maintaining quality. AI will not “learn from its mistakes” in the way a new staff member can be expected to.
Expect to see a growth in QA tools, processes, and roles that take responsibility for the quality of AI answers, identifying areas that need improvement and managing risks appropriately.
Customers will have new and higher service expectations
Customers bring their experiences with other companies and tools into every service interaction. As they become used to instant AI answers from other companies, they will be less willing to wait for help.
Conversely, the inevitable poor service experiences some companies will create using AI may create demand for access to competent “real” people. Whatever research and knowledge you have about your customers and prospective customers will need to be updated to understand how their expectations and experiences have changed.
Do they want more AI self-service, or has their exposure to AI increased the value they place on human conversation … and are they more willing to pay for it? There may be new and different segmentation of customer profiles that should be factored into your service offerings.
The value of human touch will increase
Customers today often express a desire to “talk to a real person.” Such a request tends to be less about craving human interaction and more a means of expressing frustration with a system that is not getting them what they need.
As AI improves, some of those problems will be solved. The technology will just do a better job, in the way that most of us now prefer the convenience of 24-hour cash from an ATM to waiting for an available bank branch teller.
However, there are always edge cases, and as more and more interactions are mediated by AI, access to a human will become less common. Therefore, we expect that scarcity will make human contact more valuable.
Choosing when and where to insert people into interactions with customers will be how companies differentiate the quality of their service.
Teams will become more scalable and flexible
Many businesses have large spikes in their customer service volumes, such as seasonal businesses and educational institutions. Those businesses face the challenges of handling volume spikes without maintaining large and under-utilized teams all year or scrambling to train new people for every rush season.
AI-based onboarding, training, and ongoing assistance will reduce the cost and effort of bringing new people onto a team without the same risk of huge dips in quality.
We may see smaller teams that can move between multiple products or services according to demand rather than having to stay specialized because of the time cost of switching.
What AI won’t change about customer support
After such a long list of changes, perhaps you feel like throwing everything out and starting again. That would be a mistake. The arrival of AI is disruptive, but it does not have to be destructive. If you already have a solid, working service structure, build on it.
AI cannot make companies care about their customers or their customer service. Being customer-centric is not something you can purchase from a vendor or download as a software update.
It is inevitable that within a couple of years, essentially every online service team will be using AI in some form. Some will be using it to create better experiences than ever, and others will be merely providing the same terrible service by spending even less than before.
Great support — the kind that creates long-term customer loyalty and delivers customer insights to the rest of the business — is about much more than answering questions. It’s about developing an understanding of who your customers are and what they are using your service to achieve.
Great service means asking better questions and sometimes answering a different question than the one asked of you. There will still be plenty of roles for those with empathy, understanding, communication skills, and an interest in other people.
Remaining exactly the same as a service organization is not a realistic option, but the teams that succeed through this AI transition will preserve their core values, finding new ways to express them with new tools.
Creating the AI-enhanced customer support team
Now is the time to pay attention. Notice the patterns you are in as a team, whether deliberate or unconscious. What have you been doing just because that’s the way it was set up or that’s the way the old tools worked? What does “good service” mean in your company?
Take some time to review your old staffing plans, and revisit those headcount calculations. Rework the schedules and imagine what might be possible if the AI promise becomes reality.
Encourage your team to chart their own paths and take advantage of the opportunities that a big disruption always creates, and redraw your ideal team structure to reflect the world of tomorrow — or perhaps merely next year.