A Front Desk That Never Sleeps

Automated Functions

The customer service task that companies tend to automate first is the handling of frequently asked questions (FAQs). Machuca explains, “Answering the same ten questions repeatedly drains staff time without adding real value, and Alaska’s geography makes it worse; customers in rural communities or visitors from Outside are inquiring at all hours.”

Scheduling is the second-most automated function, especially for service-based businesses like healthcare providers, salons, and tourism operators. Beyond that, automated booking and reminders are low-risk and high-reward areas for AI-assisted customer service.

Increasingly, the value is in the integration of AI—not the plug-in, according to Jake Chambers, CEO of DenaliTEK, which has been providing IT solutions in Alaska since 2001. A chat widget that only answers FAQs is useful, but an AI connected to a calendar, customer relationship management, and phone system can actually book appointments, look up accounts, or route calls to whomever is free. “That connected setup is where it goes from ‘nice’ to genuinely useful,” he says.

Companies should think of AI as a “front desk that never sleeps,” says Steve Vick, owner of Web 907, a website design team in Fairbanks. “It isn’t a replacement for your team… It’s an extension that protects your customer experience when you’re not available to talk.”

For small and mid-sized organizations, the two most practical AI entry points are conversational website chatbots and AI voice agents for phones. Many people still think of AI as the frustrating, rigid “canned-response” bots of years ago, but modern AI is different. It can read directly from the business’ knowledge base and answer in a conversational tone. “For instance, in a tourism business, the chatbot doesn’t just list excursions; it converses with visitors to ask about their party size, preferred tour duration, and expected level of adventure,” Vick explains. “It then recommends tours based on this interaction, qualifying a lead exactly like a front-desk person would.”

Real-time chat is nice, yet AI voice agents can handle similar work over the phone. For example, Web 907 deployed one to help an emergency veterinary service manage its rotating after-hours, on-call schedule. Instead of worried pet owners searching through social media posts at midnight to find the open clinic, they call one number and receive an immediate answer from a calm-voiced AI. “Under the hood, the AI is just reading a spreadsheet, but to the caller, it is the difference between panic and immediate relief,” Vick says.

Specific Expertise Required

Web 907 client Rod Pangborn uses AI to support various functions for his North Pole tour business, Rod’s Alaskan Guide Service. The company, which offers authentic adventures just outside Fairbanks, placed a chatbot on its website last October with impressive results. “It’s scary-good,” Pangborn says. “It answers questions like a human would.”

The chatbot fields the same kind of repetitive questions that the company’s employees used to handle. This has significantly reduced the number of incoming phone calls and emails, freeing staff to address other customer needs. “About 5 to 10 percent of our total volume of customers use the chatbot,” Pangborn says. “At the end of the day, it’s increased our revenue.”

How much expertise do business owners need to implement AI for customer service? They can start by experimenting with tools like ChatGPT, Claude, or Microsoft Copilot to test scripts, responses, and the outcomes they want to achieve, according to Phillip Jackson, chief information officer at Arctic IT.

However, setting up a more complex AI chatbot or phone assistant requires greater technical expertise. These tools must be designed to answer questions from a company’s knowledge base. “To operationalize a chatbot or phone assistant—to give it memory, a voice, and a ‘heart’ that reflects your business values—that typically requires a technical partner who can build it correctly and integrate it into your environment,” Jackson says.

Today, many AI tools are no-code solutions, so a business owner can stand up a basic chatbot or scheduler in an afternoon without any programming experience. In fact, Chambers recommends using AI to set up AI. “Open Claude or Copilot and ask it to walk you through building an AI chatbot or phone assistant for your business,” he explains. “It’ll give you a surprisingly good roadmap.”

The “real” work with implementing AI, Vick says, is training, testing, and troubleshooting. “Out of the box, an AI model does not know what it’s not supposed to say,” he says. “If you don’t set clear boundaries, it will happily pull random data from the open internet, write three-paragraph essays nobody wants to read, or confidently make up an answer.”

Phone and voice AI are slightly more technical, and they require specific expertise. As Vick explains, “You have to connect them to your phone system and map out a natural vocal tone for your brand. But once the lines are connected, voice AI needs the same disciplined training and guardrails as a website chatbot.”

Take Your Bot to School

Training AI relies heavily on context, according to Jackson. The more information the tool has about a company, the better it can respond. Alaska’s complexities can be built into the system through the rules, examples, and business knowledge it receives. “For example, there isn’t a road to every town, so logistical queries may require a person to validate before AI provides an answer to a customer,” Jackson says.

To keep AI from filling Alaska-specific gaps with Outside assumptions, Web 907 trains it on local realities, including written rules, pricing spreadsheets, seasonal calendars, and regional FAQs. During testing, the company asks the web chatbot or phone agent awkward, hyper-local questions until it proves it can reliably navigate the unique geography and logistics of Alaska flawlessly.

“Simply put, you have to teach the AI how to use your local resources, and just as importantly, how not to use them,” Vick says.

Chambers has a similar view, saying AI is only as good as what you feed it. For instance, a generic bot will not know that a “weather day” in Utqiaġvik means something very different than in Anchorage—but one trained on the company’s operations will. Therefore, it’s important to build that reality into the knowledge base. “You don’t train it on ‘Alaska,’” he says. “You train it on your Alaska—your specific company, solutions, industry, market, and client set.”

Affordability of AI

For a small business, AI customer service is not an expensive option, but doing nothing is, Vick says.

The cost generally breaks down to a one-time setup fee and a predictable monthly fee that scales with usage. A basic website chatbot that answers FAQs and routes users to the right contact person typically costs from $600 to $1,000 for setup and $150 to $300 a month for ongoing usage. Setup for a more complex AI agent—whether for chat interactions or handling phone calls—usually starts at $1,250 to $2,000, with monthly plans starting at $300 and scaling based on call or chat volume.

“For the vast majority of local small businesses and nonprofits, a basic monthly plan offers more than enough volume,” Vick says. “When you do the math, an AI tool that saves you hours of work each week or catches a late-night call covers its own cost incredibly fast.”

Using AI for customer service and support can work for organizations of all sizes. However, Jackson’s best advice is to start small, prove it, and then expand. “At the low end, that may be as inexpensive as a paid subscription to an LLM [large language model] for each person using it, which can run about $30 a month,” he says. “Full AI implementations can vary widely on price, depending on what you want the tool to do, how many systems it needs to connect to, and how much of the customer-service workflow you want to operationalize.”

He adds, “It’s important to note that businesses should always choose a paid AI service over a free version, so you don’t run the risk of sensitive data leaks.”

Maintaining a Human Touch

AI can perform well with high-volume, low-stakes, and repeatable questions where the answer is knowable: business hours and location, appointment scheduling, order status, or FAQ-type questions. But humans should stay involved for the exceptions, such as upset customers. “The goal is not to replace the human touch, but rather to let AI handle the routine tasks, so people can handle the exceptions,” Jackson says.

Wild Alaskan Company, which sells premium wild-caught seafood online, is taking a careful approach to leveraging AI. It closely evaluates which support tasks can be automated for customers while preserving the human touch of its member experience specialists where it matters most. “We’re using AI to respond to members’ inquiries via chat and email fully when possible and, if not, to gather information up front before handing it off to our team,” says Meg Christman, vice president of member experience. “We’re also developing some uses that can support our specialists, such as performing tasks for them mid-conversation, that they currently need to do manually. This is intended to help free them up to keep their focus directly on the conversation.”

Thanks to AI, Wild Alaskan Company has interacted with more customers while its support team is offline. At the same time, the company is working to improve the quality of its AI-generated responses. “We continually review both the answers it provides and the knowledge it draws from, ensuring conversations are routed to a specialist before a customer gets stuck in a loop,” Christman says. “In our industry, we prefer our highly trained member experience specialists to handle areas such as cooking advice and food-safety guidance. And whenever someone wants to speak with a person, we want to make that connection; that’s what we’re here for.”

Deploying AI without a thoughtful human backstop can damage trust in ways that are difficult to repair, Machuca says. This is especially important for businesses serving Alaska Native communities or rural customers, where cultural responsiveness is essential. That’s why community context should be intentionally built into the data AI draws from.

Machuca offers this simple rule for deciding when AI needs human oversight: “If getting it wrong would cost you a customer permanently, keep a human in that loop.”

AI is not a customer-service replacement strategy; it’s a customer-service “capacity” strategy, Machuca emphasizes. “The businesses winning with AI right now aren’t the ones who handed everything to a bot,” he says. “They’re the ones who used AI to free up their best people to do the work that actually requires human judgment, relationship, and expertise.”

Today’s customers expect personal, attentive service and are often wary of cold, robotic interactions, Chambers says. The trap many businesses fall into is using AI to replace the human touch. But when used well, it does the opposite. “It clears away the friction, the phone trees, the waiting, the repeating yourself, so your people are freed up for the moments that actually build loyalty,” he says. “The goal should not be fewer humans but humans doing what only humans can do, with the busywork lifted off their plate.”

Chambers continues: “Do it thoughtfully, and it makes your service more personal. Do it carelessly and it makes you feel like every big, Lower 48 company we all dread calling. That difference is entirely in the setup—which is why I’d tell any Alaska [business] owner to treat AI in customer service as a way to protect their personal touch, not trade it away.”