How Do Companies Embrace AI without Losing the Human Connection?
Rachel Orston, Chief Customer Officer at Instructure, on making customers and employees feel more understood.
As companies race to put AI into customer service, a quieter risk is emerging. Tools built to save time can end up putting distance between a business and the people it serves, and between leaders and the employees who do the work.
Rachel Orston has spent more than 25 years thinking about that relationship. A customer success pioneer, she has scaled customer organizations at Fortune 100 companies and high-growth technology businesses. Today she is Chief Customer Officer at Instructure, the educational technology company best known for developing the Canvas learning platform.
In this interview, Orston explains where AI implementations go wrong, which human skills now matter most, and why she believes trust, not efficiency, is the real measure of success.
The Mistakes that Create Distance
As companies rush to implement AI in customer service, what are the most common mistakes you're seeing that actually create more distance between brands and their customers?
One mistake is designing around what AI can automate rather than what is actually helpful for a customer. Success with AI, or any technology, should be measured by whether customers receive meaningful help, feel understood and ultimately keep their trust in the company.
Another is automating the entry-level work that helps people build experience, then wondering why early-career professionals struggle with more complex situations. Confidence and good judgement develop through practice. As AI takes on more routine tasks, companies need to create deliberate opportunities for employees to learn and build that confidence.
Continuous training remains essential, because many customer interactions require understanding, judgement and empathy.
Hiring for Judgement, Not Just Experience
You've built customer success teams for over 25 years. How has your approach to hiring and developing talent changed now that AI can handle many traditional support tasks? What human skills have become more valuable?
Earlier in my career, I put more emphasis on whether someone had handled similar customer issues before. Today, I'm more interested in how they approach an unfamiliar problem. Do they ask thoughtful questions, verify information and use their own judgement?
My experience working at CNN helped shape that perspective. I would never publish information simply because someone shared it with me; I had to do my own fact-checking. Working with AI requires that same discipline.
Critical thinking, communication, sound judgement and adaptability are becoming even more valuable in an AI-enabled workforce. Deep product knowledge remains essential too. When teams combine that expertise with AI that takes routine tasks off their plates, they have more time to understand customers' needs and provide meaningful help.
In developing talent, I want people to learn how to use AI responsibly but, more importantly, to always apply their own judgement when evaluating any decision. That means questioning AI's outputs, verifying information and thinking through the best next step themselves.
We also need to keep onboarding and early-career experiences intact, so that junior team members can gain experience and practice a range of scenarios before taking on much more complex work.
Vera: Making the Customer's Voice Accessible
Can you share a specific example from Instructure where AI has enhanced the human connection rather than replaced it? What made that implementation successful?
We held an AI hackathon where our team built a tool code-named Vera. It brings together signals from customer call recordings and Qualtrics surveys to surface customer sentiment. A product leader, or anyone in the company, can ask Vera a question and quickly understand how a customer is feeling and what they need.
The goal is to make the voice of the customer accessible across the organization, giving everyone real-time insight into customers' experiences with us.
What made this approach valuable was starting with a need our team already understood: making customer feedback easier to access and act on. Vera helps employees prepare for better-informed conversations and decisions. The human connection comes from what people do with that insight, like following up, asking better questions and responding to what matters to the customer.
Empowering Employees, Not Threatening Them
Many employees worry that AI adoption puts their roles at risk. How do you communicate AI strategy to your teams in a way that makes them feel empowered rather than threatened?
I start by acknowledging the concern and being open about what we're introducing AI to do, what it won't do, and how we expect it to change the work.
We position AI as a tool to support employees. That might mean getting a first draft started, finding information faster or reducing repetitive tasks, so people have more time to understand a customer's needs and work through a complex problem.
But that message needs to be backed by training. Employees need time to experiment, learn how to question AI outputs and understand when to rely on their own thinking. They should also help shape how we use these tools, because they know where customers experience friction and where a human conversation matters most. The goal is for our teams to feel empowered and supported.
The Litmus Test for New AI Tools
When you're evaluating new AI tools for customer success, what's your litmus test for whether a tool will strengthen or weaken the customer relationship?
My litmus test is simple: will this help the customer make progress and strengthen their confidence in us?
Then I look at what that means in practice. Is the information accurate, relevant and informed by the customer's situation? Does the tool fit naturally into their experience? Can they easily reach a person when they need one, and will that person have the context so the customer doesn't have to start over?
Transparency and data protection are essential. Customers should understand when AI is involved, how their information is used and what safeguards are in place.
I also ask whether the tool helps our employees make better decisions while keeping people accountable for the outcome. Ultimately, we need evidence of better resolution. Saving our team time is valuable, but the customer needs to feel the benefit.
Where a Human is Non-Negotiable
At what point in the customer journey is human interaction non-negotiable, even if AI could technically handle it? How do you make those decisions?
Customers should always have a clear, easy path to a person when they want one.
I make decisions based on the complexity of the situation and the customer's preference. When a customer is frustrated, trust has been damaged or a decision has significant consequences, a person needs to be involved.
AI can help by gathering information and giving our team the context to respond thoughtfully. But a person should own the conversation when it calls for a unique solution or for accountability.
Our Customer Connectivity Program is another example of where human interaction matters. Through regular meetings and discovery sessions, customers serve as advisors whose perspectives help shape our priorities and decisions. AI can help organize their feedback, but the partnership depends on people listening, discussing tradeoffs and showing customers how their input leads to action. That's how we build trust over time.
The Next Three to Five Years
Looking ahead, what does the ideal blend of AI and human touch look like in customer success? How should leaders be preparing their organizations now?
Over the next three to five years, I expect AI to take on more routine tasks and support increasingly complex workflows, including helping teams recognize when customers need support before they ask. That will give people more capacity to advise customers, strengthen relationships and work through situations that require judgement.
Getting there will require a more deliberate approach to training. AI training needs a complete overhaul. AI literacy has to become an ongoing part of employee development, with opportunities to practise, receive coaching and learn as the tools evolve.
The best blend of human touch and AI is one that makes customers feel understood and confident in the team they're working with.
The companies that succeed will measure progress by how effectively they help customers achieve their goals. Greater efficiency matters, but stronger relationships and lasting trust are the more meaningful measures of success.
Rachel Orston is Chief Customer Officer at Instructure, the maker of Canvas and the LMS market share leader in higher education and paid K-12, serving more than 15,000 institutions, districts and schools in over 100 countries. A serial entrepreneur, technology executive and customer success pioneer, she has spent more than two decades leading post-sales organizations at companies from Series A to public, helping scale growth from $1 million to more than $1 billion. Before joining Instructure in 2024, Orston was Chief Customer Officer at SmartRecruiters, the hiring platform used by organizations such as Visa, IKEA, LinkedIn and Bosch. She is also a founding member of CS Angel, the first investment syndicate of customer success and customer experience leaders who back and advise startups focused on improving net revenue retention.
