eCommerce Fastlane

Free every Thursday · 59,694 Readers

Fastlane Insider Newsletter: What's working for Shopify in the AI era?

Your AI Closed 70% Of Tickets. Half Didn’t Stay Closed.

Your AI support dashboard says 67% of tickets resolved. Your CX team still feels underwater. Both things can be true at the same time, and the gap between them is where Shopify brands doing $10M+ are quietly leaking customer lifetime value.

If you switched on an AI agent this year and your human support costs haven’t budged, this episode explains why.

Sachin Jaiswal is the co-founder and CEO of Kim.cc, an AI-native customer support company that runs support for 200+ Shopify brands, including TRIP, The Perfect Jean and Kahawa 1893. He built Niki.ai, an AI shopping assistant, back in 2015 (years before ChatGPT existed), co-founded the healthcare analytics company Innovaccer, and helped build an AI plus human support system for one of India’s largest food delivery platforms, which he says handled 10 million tickets a month.

In this conversation, Sachin breaks down why roughly half of “AI-resolved” tickets come back, the 90% confidence threshold his system uses before AI is allowed to act alone, and the human-in-the-loop model that stops hallucinations before they ever reach your customer. With BFCM days away, he also shares the one number worth checking before your ticket volume jumps 4x to 5x. Let’s dive in. 👇

What You’ll Learn

✅ Why a 70% AI resolution rate can be misleading: what’s really happening to the tickets your dashboard counts as closed, and why so many of them come back.

✅ The AC repair story every founder should hear: how one AI agent promised a technician the brand never employed, and what it reveals about who owns AI mistakes.

✅ How Kim decides when AI is allowed to act alone: the 90% confidence threshold, and what happens to every ticket that falls below it.

✅ What “Sentinels” actually do all day: the human layer that trains, monitors and catches edge cases before a wrong answer lands in your customer’s inbox.

✅ How support turns into a retention lever for subscription brands: Sachin’s approach to cancellation requests driven by travel, flavor fatigue and other fixable reasons.

✅ Whether it’s too late to fix your setup before Black Friday: a one-week onboarding path, plus a free assessment that shows you the real scope of the problem first.

Episode Sponsor

This episode is brought to you by Walmart Marketplace.

Holiday starts early on Walmart Marketplace. If you’ve got a proven product and you’ve been weighing a second sales channel, this one comes with serious reach: roughly 280 million customers and members shop Walmart’s stores and websites every week. There are no setup or monthly fees, just a referral fee on what you actually sell.

New sellers can unlock up to $75K in New-Seller Savings, including discounted referral fees, up to $2,000 in Walmart Fulfillment Services credits for fast 2-day shipping, and advertising credits to get discovered. Build your seasonal assortment with Customer Favorites, products Walmart shoppers are already searching for, and you could get up to 100% off referral fees on eligible items.

Get holiday-ready with Walmart Marketplace. Visit marketplace.walmart.com/ecomfastlane and sign up today.

*Conditions apply. New-Seller Savings runs through January 31, 2027. See Walmart’s offer details for eligibility.

Episode Summary

Most brands judge their AI support by one number: resolution rate. Sachin Jaiswal thinks that’s the wrong number. Across the brands Kim has analyzed, he sees the same pattern: AI reports 60% to 70% of tickets resolved, yet about half of those come back. The ones that matter, he argues, are the tickets that stay closed. Everything else is AI adding a new line item to your P&L while your human support cost stays exactly where it was.

Sachin breaks the problem into two parts. Part A is AI that makes mistakes. One AC brand’s AI cheerfully promised a customer a repair technician; the customer waited 10 days, and the escalation eventually reached the founder, who had to explain the company has no technicians on payroll. Another AI told a shopper their order could go to a neighbor’s address. It didn’t. Each mistake creates double work (clean up the error, then answer the question properly), and when you’ve spent $500 to $1,000 acquiring a customer you expected to be worth $2,000, a frustrated churn costs far more than the dollar you saved on a ticket. Part B is AI doing lip service: a polite “thanks for reaching out” before handing the hardest, most time-consuming tickets to your team anyway.

You’ll get the real story on how Kim’s model is built differently. Kim sits on top of the helpdesk you already use (Gorgias, Zendesk, Shopify Inbox) and takes accountability for the outcome. AI only acts alone when its confidence is above 90%. Everything else goes to Sentinels, trained CX pros who teach the AI scenarios like the 64 different flavors of “where is my order,” monitor effectiveness by ticket type, and arrest risky answers live. Asked whether you ship to Australia with no policy in place? A Sentinel catches it, gets the answer from your team, and trains it back in.

Steve also digs into the economics: $0.70 per resolved ticket with the human team included, a 60% margin target instead of the 90% typical in SaaS, and human involvement that Sachin says drops by around 90% after the first 90 days. Plus a subscription deterrence system designed to save customers who reach out to cancel.

And with BFCM pushing volume up 4x to 5x, Sachin puts the stakes plainly: if your AI gets 100 out of 1,000 tickets wrong today, what happens when that’s 400 or 500? This isn’t a pitch to rip out your helpdesk. It’s a blueprint for making AI support actually pay for itself.

Strategic Takeaways

👉 Track tickets that stay closed, not tickets closed. A resolution rate only counts the first reply. Pull the reopen rate on AI-handled conversations over a 30-day window and compare it with your headline number. If you’re under 1,000 tickets a month, you can spot-check by hand in an afternoon; at 10,000+, sample by ticket type so one strong category doesn’t hide a weak one.

👉 Treat every promise your AI can make as a policy decision. The AC technician story happened because nothing told the AI it couldn’t commit to something the business doesn’t offer. Audit what your agent is allowed to promise: technicians, address changes, refunds, delivery dates, exceptions. Where no written policy exists, the AI shouldn’t be improvising one.

👉 Price AI against your full support P&L, not the vendor invoice. If you pay per ticket touched, including partial work, and your team still finishes most of those tickets, you’ve added a cost rather than replaced one. Ask any vendor directly: do you charge for tickets attempted or tickets resolved, and who finishes the rest?

👉 Measure AI by disposition, not in aggregate. Where is my order, returns, replacements and subscription changes each succeed or fail at different rates. Sachin’s team trains for dozens of variations of WISMO alone. Breaking performance out by category tells you exactly where AI has earned autonomy and where it still needs a human watching.

👉 Put a human judgment layer on high-value decisions. If your average order is $1,000 (furniture, appliances, premium goods), approving or rejecting a $100 return isn’t a script, it’s a call. Route those moments to someone who understands your margins, your customer and your risk tolerance, even if AI handles everything around them.

👉 Turn cancellation tickets into save attempts. A customer leaving for a two-month trip wants a pause, not a cancel. A customer tired of one flavor wants a swap. Make sure your support flow (human or AI) offers the fix before processing the cancellation, because a vanilla “done, you’re cancelled” throws away everything you spent to acquire that subscriber.

Guest Spotlight

Sachin Jaiswal
Co-founder and CEO, Kim.cc

Sachin leads Kim.cc, an AI-native customer support company serving 200+ Shopify brands, including TRIP, The Perfect Jean and Kahawa 1893. Kim pairs AI agents with trained human Sentinels, plugs into helpdesks like Gorgias and Zendesk, and charges per resolved ticket rather than per AI attempt. In June 2026, Kim closed an institutional venture round led by Stellaris Venture Partners to accelerate its U.S. expansion.

Sachin has been building with AI longer than most. An electrical engineering grad from IIT Kharagpur, he started as a product manager at Oracle, co-founded healthcare analytics company Innovaccer, and in 2015 co-founded Niki.ai, a conversational AI shopping assistant, years before ChatGPT made the category mainstream. He later helped build an AI plus human support operation for one of India’s largest food delivery platforms (think DoorDash scale), which he says processed 10 million tickets a month.

That combination of early AI conviction and hard-won operational scars is what makes his take valuable. He’s not arguing against AI. He’s arguing that someone has to own the outcome.

Links & Resources

Featured in This Episode:

  • Kim.cc: AI-native customer support with human Sentinels for Shopify brands
  • ACE Challenge: free assessment that scores how reliably your AI support actually resolves tickets
  • Listener Offer: 20% off when you sign up with Kim, plus a 30-day free trial (valid even if you start after BFCM)
  • Sachin Jaiswal on LinkedIn: connect with Sachin
  • Email Sachin: [email protected]

Kim Customer Brands Mentioned:

  • TRIP
  • The Perfect Jean
  • Kahawa 1893

Helpdesks Mentioned:

  • Shopify Inbox
  • Gorgias
  • Zendesk

AI Support Tools Mentioned:

  • Gorgias AI Agent
  • Yuma AI
  • Decagon

Thanks for Supporting the Pod!

Over 9 seasons, I’ve been incredibly fortunate to chat with some of the brightest founders building amazing Shopify brands, as well as the partners shaping the app and marketing ecosystem. Every conversation has taught me something new, and I’m grateful for the chance to learn alongside you.

What matters most is that this podcast helps you solve real challenges and discover new ways to grow. Your support, feedback, and stories have made this journey truly special. Thanks for tuning in, sharing your wins and losses, and being part of the eCommerce Fastlane community.

Stay Connected: Leave an Honest Rating/Review on Apple Podcasts or Spotify. Follow & Subscribe on YouTube for new episodes.

Get in Touch: Email | LinkedIn

Like Reading? Here’s the Full Episode Transcript 👇

Click to Expand Transcript ▼

Steve Hutt: Hey there, welcome back to eCommerce Fastlane. I’m your host, Steve Hutt. Today we’re talking about AI customer support and the number on your dashboard that might be lying to you. If you’re a brand doing $10 million or more, you’ve turned on some kind of AI support this year, and your resolution rate looks great while your team still feels underwater, today’s recording is for you. I did some research before recording, and my guest’s take is that most brands are reporting around 60% to 70% resolution and calling it a win, while about half of those tickets come back. What matters, he says, isn’t how many tickets your AI closes, but how many actually stay closed. That’s important, right? Having a human in the loop as needed so the ticket actually stays closed. My guest is Sachin Jaiswal, co-founder and CEO of Kim.cc. He built one of the first AI shopping assistants back in 2015, before ChatGPT was even a thing. Eventually, they had to shut that platform down, but his team now runs support for more than 200 Shopify brands. It’s really interesting what he’s built with Kim.cc and his team. By the end of today’s recording, I’m hoping you’ll know whether it’s too late to fix this before Black Friday, which I don’t think is the case, and which number on your dashboard you should be thinking about. Sachin, welcome to eCommerce Fastlane.

Sachin Jaiswal: Thanks a lot, Steve, and thank you for hosting me on the podcast. I’m really excited to share a story that every brand should be talking about. Everyone is excited about AI today. Everyone wants to implement it, but how do you ensure it’s working reliably and delivering the ROI you expected? Most brands trying AI end up adding it as another line item, while their human costs stay the same. They question whether the ROI is actually working out for them. That’s the story I want to present: what the problem is, the extent of it, and the system we’ve seen work. For everyone listening, I’m Sachin, one of the co-founders of Kim. As Steve mentioned, I started an AI company back in 2015. It was an AI shopping assistant called Niki. I also started a company called Innovaccer, selling into healthcare to hospitals like Mercy. It’s a unicorn now. In the interim, I built a customer support platform for an ecommerce company called Swiggy that used to get 10 million tickets a month. That’s not $10 million in revenue. They were getting 10 million support tickets a month. We built a similar system using AI with humans in the loop to ensure reliable support for customers. Having seen that at the scale of 10 million tickets a month, I thought it was a playbook we could share with other ecommerce companies. Swiggy is like DoorDash for India. They deliver food. That playbook could be relevant and help a lot of ecommerce brands. I wanted to share those learnings, and I’m really excited to talk about them today.

Steve Hutt: It’s pretty amazing. I did some research before we started because a lot of people are using Google for work. They’re in the early stages, using Gmail, and they might have a light CRM that’s a Gmail plugin or another notable CRM solution. There are a lot of interesting players out there. Even Shopify has Inbox and Shopify Email. People aren’t unaware of those options. But something interesting I learned from you is that roughly half of AI-resolved tickets get reopened. That brings us back to ROI. There’s a cost associated with having an AI system apparently respond successfully to an inquiry, but the reality is that so many of those tickets are being reopened. What’s actually going on here?

Sachin Jaiswal: I’ll answer those two parts separately. First, there are different types of tools, which we call help desks, in the market today. Shopify Inbox is one, Gorgias is another, and Zendesk is another. These are systems that help you organize work. We don’t displace or replace them. We sit on top of them to help you do the work with a unique AI-with-human-in-the-loop model. There are really two approaches people use to do the work. Some companies give you tools. Gorgias AI, Yuma, and Decagon give you an AI tool you can implement for customer support tickets. Then there’s a company like us. This category is called an AI-native service, a term used by Y Combinator and other organizations. We take accountability for AI outcomes. We use AI with humans in the loop to make sure that if you send us a ticket, we get it resolved. This is a huge problem in the AI industry today. For the last year, boardrooms have been pushing AI into workflows because it’s one of the most remarkable inventions of our century and one of the biggest technology breakthroughs we’ve seen in some time. But after going through that journey, we’re still seeing implementation ROI fall short. Most brands that implement AI add a new cost line item while their existing human costs remain on the P&L. They start wondering, “Of course I need AI, but why isn’t my cost going down? Why is it only going up?” The data doesn’t tell the full story. We’ve analyzed many brands, and we have something called the ACE Challenge that helps customers uncover the reality of their customer support AI by analyzing their tickets. The problem today is that AI makes mistakes, or it only does half the work before passing the baton to humans. When AI makes mistakes, it creates double the work for your customer support team. First, they have to clean up the mistake.

Steve Hutt: Right.

Sachin Jaiswal: Then they have to do the actual work of answering the ticket. Most of the brands we work with say AI is effective only half the time. That’s causing costs to increase without generating the efficiency they expected. Half of those tickets are getting reopened because AI is making mistakes or isn’t completing the work. I can give you real examples. We were talking to an air-conditioning brand whose customer reached out and said, “I bought an AC from you, but it needs a repair. Can you send a technician?” The AI said, “Sure, I’m going to send a technician.” The customer waited 10 days and reached out again: “I’ve been waiting for the technician. Where are they?” The matter got escalated to the founder, who said, “Hold on. We don’t have a technician on our payroll. What’s happening?” That’s a hallucination, and it happens. As Sam Altman has said, it’s a feature, not a bug. AI is trying to be helpful and can make mistakes in the process. Who owns the outcome when it makes that mistake? That’s one example, but there are many others. Someone might ask, “Can I get this order delivered to my neighbor?” The AI says, “Sure, you can do that,” but the order still gets delivered to the customer’s address. Those frustrations pile up and can lead to customers churning. Imagine you’ve spent $500, sometimes $1,000, acquiring a customer, and you expect a lifetime value of $2,000. Because of a frustrating experience, the customer churns. Multiply that across your customer base. You could be spending millions of dollars on acquisition, but because you’re trying to save a dollar through AI efficiency and have taken it a little too far, you’re losing customers. You could lose thousands of dollars per customer because you wanted to save a dollar. You have to be mindful of the customer experience you’re delivering so your customers don’t churn. These mistakes can also become a PR nightmare for your brand. We’ve read stories about McDonald’s chatbots, Spirit Airlines, and other companies that implemented AI and ended up facing million-dollar lawsuits. Those mistakes affect your brand. In our case, we use AI with humans in the loop, so we take accountability for those outcomes. We know AI can make mistakes, but we don’t let those mistakes reach your customer. That’s one part of the problem: AI makes mistakes. The other part is that sometimes it’s just paying lip service. A customer reaches out, the AI says, “Thank you for reaching out,” and hands it over to a human. Maybe it gives a slightly more sophisticated answer, but it still hands the work to a human. Your existing support team then has to finish what AI left behind. Most of the tickets that roll over to humans are more complex and require more time per ticket. As a result, the actual work your human team does isn’t really coming down. You want vendors that take accountability for end-to-end resolution. I’ve seen companies say, “I’m going to charge you a dollar per ticket,” and that includes tickets where AI has only done part of the work. They’ve spent money on tokens, so they charge you for those tickets too. But you don’t recover those costs because your human team still has to take care of the ticket at the end. There’s a more efficient model, like our AI-native service. Whenever there’s remaining work to be done, we deploy our own humans in the loop. You don’t end up with an extra cost line item, and accountability for the AI’s work stays with us.

Steve Hutt: You have something that I think is called a confidence score. Apparently, your AI only acts alone if its confidence score is above 90%. How does it know what it doesn’t know? The technology behind that is interesting to me. You give it a confidence score, and if it’s above 90%, it can do the work digitally as an AI agent. If it’s below that, there are checks and balances, which eventually bring in the human in the loop. I guess that’s the accountability layer. But how does the AI agent know it’s at 90% or higher and is unlikely to hallucinate or make a mistake with the customer?

Sachin Jaiswal: Part of that scoring model is proprietary, but it’s also part of the system design behind our AI accountability layer. If a ticket comes to us, we take responsibility for resolving it. Having accountability from the AI system is paramount, and we’ve built a system to uncover that. That’s not necessarily the case with AI tool providers. If they show low confidence in a particular resolution, they can’t make money from that work. It’s not necessarily in their interest to make that visibility available. For us, because we take accountability for outcomes, we need to know exactly where AI is falling short or isn’t sure it can resolve a ticket with complete confidence. We’ve designed it that way because of the approach we’ve taken to solving the problem.

Steve Hutt: I love that accountability. A lot of people are either all in on AI or they’re not, and I think that’s the problem. Some of the more basic AI chat solutions on websites are simply scraping the site’s content. You might point them toward a few FAQs: “Do you ship to Canada? What’s the lead time? Are duties and taxes included? What’s your return policy?” Those things can be ingested and made available through a chat solution, and maybe that’s the starting point. But that isn’t true accountability. What’s happening behind the scenes? Are people auditing those solutions to see whether the right answer was given? That’s the flywheel I find interesting. Your agent can answer when it has high confidence. If it can’t, there’s a workflow back to a human in the loop who can do the work without an additional cost beyond the fee for resolving the ticket, including when a live agent is involved.

Sachin Jaiswal: Part of the belief here, Steve, and I think the consensus people are arriving at, is that AI needs human accountability. For example, when you post content, you might use Claude to brainstorm or ideate, and you might get it vetted by ChatGPT or Perplexity. But you’re still the human in the loop. You’re the one taking accountability for whatever goes out from your account. AI might give you the ability to do 10 times more, but the accountability layer stays with you. Why is customer support any different? If you’re deploying a tool to your customers and it makes mistakes, where is that accountability layer? Who owns it? If something goes wrong, is your vendor just saying, “I’m sorry,” and moving on? Or are they making sure those mistakes don’t reach your customers? If your AI is passing work to your existing team, is it really giving you ROI? If not, your vendor should answer why it isn’t doing the complete work. This isn’t coding, where there’s still a human in the loop. These are inquiries from your end customers. If you’re allowing AI to do end-to-end work and it makes mistakes, it can harm your brand’s reputation. You need a better accountability layer around the AI you’ve implemented. It can’t happen after the fact, because by then the mistakes have already reached your customer.

Steve Hutt: You also use an interesting term: sentinels. I always think of a movie like The Matrix when I hear that word. But you call your people sentinels. Can you talk about what that actually means? I’d like to understand it because it’s a term you use as part of the Kim platform.

Sachin Jaiswal: Their job is to guard the work being done by AI. Our mission is to deliver reliable customer support AI. To do that, we need someone to take accountability for the outcomes, and that’s what our sentinels do. Their job has several parts. First, they train the AI. If your customers ask, “Where is my order?” or say, “I need to make a return,” the sentinel makes sure the AI understands how to handle those scenarios. They might train it to handle 64 different “Where is my order?” scenarios, and as they encounter more, they keep retraining it. Second, they monitor the AI’s work. In the initial days, they review attempted resolutions across different ticket categories. How effective is it at handling order-status inquiries? How effective is it with returns, replacements, or subscription questions? They monitor the work so that low-confidence responses don’t reach your customer. As each category becomes more effective and accurate, they reduce the monitoring while still making sure low-confidence inquiries are reviewed. If a new scenario comes up, they monitor those edge cases and feed them back into the AI. In the process, they ensure your customers don’t end up with a suboptimal resolution. They don’t just train the AI on those edge cases. They also make sure those scenarios are corrected live. For example, someone might ask, “Can you ship to Australia?” If there’s no defined policy, someone needs to find out the answer and feed it to the AI so it doesn’t hallucinate or make a mistake. They stop that ticket from being answered incorrectly, learn from your team how to handle the situation, and train the AI so it knows the answer next time. The customer gets the correct answer because the sentinel prevented the AI from making a promise that couldn’t be fulfilled. The third part is handling edge cases and escalations, depending on the risk involved. If you’re selling air conditioners or sofas and your average order value is $1,000, most of the work might be done by AI, but someone may still need to review whether to approve or reject a $100 return. We have CX experts training and working with AI, so they handle those escalations rather than handing the tickets back to your team. If you send us a customer inquiry, we make sure it’s answered well. Naturally, these sentinels are plugged into your existing ecosystem, including your Slack channels. They work with your team, learn about your brand, and answer the way they’re supposed to. Because of this model, we deliver both ROI and quality outcomes.

Steve Hutt: I was thinking about brands listening today that are in the subscription business. They like the predictable revenue and lifetime value that come with it, especially in supplements and sports nutrition. There are even subscriptions for socks and underwear. It’s amazing what’s available now. I read somewhere, possibly in your support documentation or case studies, that one subscription brand was saving around 21% of cancellation requests because the agent had been trained to understand the reason. Is it flavor fatigue? Do you need to change the address? Are you going on vacation? Do you want to pause rather than cancel? You have some interesting technology to help subscription brands. That’s one example, but can you talk about it? I hadn’t really thought about how support could turn into a retention tool with positive ROI.

Sachin Jaiswal: What we’re seeing is that most brands don’t want customer support to be just a cost line item. They want it to generate revenue and prevent revenue churn. There’s a big need among brands that receive subscription-related inquiries, so we’ve designed a system called subscription deterrence. A customer might reach out to cancel because they’re traveling and don’t know how long they’ll be away, or because they don’t like the taste of a product. You’ve spent so much money acquiring those customers. You don’t want AI to do a basic job of simply canceling the subscription. You want an attempt to save the customer from churning. We build systems and solutions around that. Given our team’s experience building this at scale, we’ve figured out design approaches to solve these problems. It’s about giving the customer support bot the ability to prevent subscription churn while ensuring customers get the quality of experience they want. Subscription deterrence is a focused solution we build for subscription brands.

Steve Hutt: It’s amazing. Pricing is the next thing on my mind because people are price-sensitive and want the best ROI from their AI investment. I noticed your pricing starts at around 70 cents per resolved ticket, but the important part is that it includes a human team. Most others in this category, with an AI component in their CRM, seem to be closer to a dollar or more, and that’s often software only. There’s no human included in the overall process. Can you talk about that math? It’s a significant saving when you’re getting both components: enterprise-quality AI resolution and, when the AI isn’t confident, routing to a live human for resolution. That’s interesting for only 70 cents.

Sachin Jaiswal: Unlike some big platforms trying to maximize revenue from the interest in AI, our focus is different. We’re trying to build a solution that guarantees ROI for customers. We want to make sure you’re implementing AI for the right reasons: reducing costs, delivering the savings you’re looking for, and ensuring you get the quality you want. In our operating model, unlike a typical SaaS business trying to make 90% margins selling AI, we’re happy with a 60% margin, and we optimize it on the back end. In the initial days of implementation, there’s more human effort involved in training and monitoring the AI. Typically, after day 90, our human involvement comes down by 90%. Being based in lower-cost centers also lets us deliver this human-in-the-loop package cost-effectively. We’ve made design choices that enable that. For example, the sentinels who train, monitor, and handle edge cases are based in India and the Philippines. We’re not trying to maximize the profit we can make. We want to keep it at a level that’s sustainable for both you and us.

Steve Hutt: Depending on when you’re listening to this recording, we’re about 11 days out from Black Friday, the Super Bowl of commerce. If a brand called you today, is it too late to implement Kim? Or is there a reasonable onboarding process that can happen quickly and create a seamless but improved customer experience?

Sachin Jaiswal: Let’s first look at the overarching problem. Your AI can handle 1,000 tickets, but what happens when it gets 100 wrong during BFCM?

Steve Hutt: Yeah, right.

Sachin Jaiswal: We all know volume can go up four or five times during BFCM. Those 100 mistakes become 400 or 500. Those misses could mean a lot of your customer acquisition spending leads to more churn. If you’re interested in working with us, our onboarding process takes about a week. We’ll make sure we get started quickly enough to handle the initial volume. Having said that, if you don’t want to unsettle your current setup, at a minimum you should understand the scope of the problem. How effective is your AI? Is it making mistakes? How many? Is it actually answering those tickets? If your dashboard shows 70% resolved by AI, is it really 70%, or is most of that work going to a human? If there are mistakes, what’s happening as a result? Are customers getting frustrated? Customers not saying anything can also be a big problem. You can uncover that through our ACE Challenge. We help you look under the hood of your existing AI setup and see how well it’s tackling the work. The tool is completely free. You can install it and get a report in Slack about how well your AI system is designed. Maybe we can add a link to the ACE Challenge with this podcast so listeners can see how well their AI is doing and become better prepared, informed, and aware of the problem. Whenever you’re ready to talk, we’re happy to help quantify the scope of the problem, assess your current AI setup, and explain what’s possible and what isn’t.

Steve Hutt: The ACE Challenge is interesting because I went and took a look at it. I’ll put it in the show notes. It’s kim.cc/ace-challenge. I ran one of my friends’ leather handbag companies through it because I was curious to see what it would show. It’s a free assessment, and it really understands the business. In my opinion, it measures whether AI is reliably resolving your customer service challenges, not just how fast it replies. Is it effectively set up? That’s an interesting framing and a bit of a contrarian take. It helps you find gaps in your support. You put in your name, email address, current monthly ticket volume, and role in the company, and then you get your ACE score. From there, you can make decisions about what to do based on the results. Once someone gets that number and sees what ACE has found, what do you believe their next steps should be?

Sachin Jaiswal: Usually, the report gives you an analysis of how well your AI is handling different ticket categories. For example, how well is it handling “Where is my order?” inquiries or returns? What scenarios haven’t you thought of that you should consider? If you find issues you want fixed, we can fix them completely free of charge. If you want a long-term solution, not just for BFCM but beyond, we’re happy to see how we can help. Independently of that, the ACE Challenge will uncover things that may surprise you.

Steve Hutt: That’s good. People should head over to Kim.cc. I asked you before recording, and the CC is an acronym for customer care. I think it’s a great name. What do you believe the next steps should be? Should people reach out to you directly, go to the website, or sign up through the ACE Challenge? I also know you have an offer for eCommerce Fastlane listeners.

Sachin Jaiswal: For anyone interested in learning more about how we can help make your AI more effective, you can reach me at [email protected]. That’s S-A-C-H-I-N at kim.cc. You can also head to the ACE Challenge to learn how effective your AI is and understand the scope of the problem. For those who want to sign up now, we’re offering a flat 20% discount. If you sign up now but want to get started after BFCM, that will still apply. We’re happy to help in any way. Our mission is to make great customer service accessible to growing brands. Anything we can do to help you toward that is what we want to work on.

Steve Hutt: That’s amazing. I’ll put your email and LinkedIn in the show notes so people can get in touch. People should try the ACE Challenge, and thank you for the 20% off the first month. There’s also a 30-day free trial, so it’s a no-brainer to at least try the platform. You don’t know what you don’t know. The fact that hundreds of brands have already decided to use Kim shows it’s working for a lot of businesses. When I heard about you, I knew I had to have you on the show because I wanted to learn more. I love progressive new businesses solving old-school problems with technology. Adding a human in the loop gives you a unique solution that I don’t think is rivaled by anybody in the marketplace right now. You get a bit of a first-mover advantage by offering this, and that’s why I wanted to have you on today. Thank you so much for sharing.

Sachin Jaiswal: I’ll add that some really great brands are working with us. TRIP Drinks, which recently signed Kendall Jenner as an ambassador, is one of our clients. The Perfect Jean is another, along with Shark Tank brands like Hava Coffee. As Steve mentioned, hundreds of brands work with us because they understand the problem we’re solving is real, and they want a solution that works for them. If you’re curious or want to learn more about how this could work for you, we’re happy to discuss it and help.

Steve Hutt: It’s amazing. I’ll put links to those customer stories in the show notes. I can see the TRIP Drinks story about how they scaled their customer support. It’s really cool. You have a great solution. Thank you again for recording today. I wish you continued success, and thank you for helping all these Shopify-powered businesses. This is a necessary service, and the human-in-the-loop component is the flywheel that makes it unique. Thanks again for recording.

Sachin Jaiswal: Thanks a lot, Steve. Thank you for taking the time.

Steve Hutt: All right, have a great afternoon.

FIND US ONLINE

WEEKLY DTC INSIGHTS

TRUSTED BY THOUSANDS

TRUSTED PARTNER

Choose a language