AI won’t fix your business. It’ll scale the mess.

Ty Givens has spent 25 years building customer support infrastructure for fast-growing companies. Her position is unambiguous: most small businesses lose their next sale before they go looking for a new customer. They pour everything into acquisition and fumble the experience so badly after the purchase that those customers don’t come back, and don’t send anyone else either.

Givens is the founder of CX Collective, a consultancy that has spent decades helping brands work out what goes on between “thank you for your purchase” and “I’d like to speak to a manager.” Her read on small business growth is direct: the cheapest thing you can do is treat the customers you already have so well that they do your marketing for you.

She also has things to say about AI, and not the standard version. If your customer service process is broken, adding AI won’t fix it. It will run the broken process faster, at scale, to more people. Her line: add AI to a broken process and you don’t fix the mess, you scale it.

Customer experience is your cheapest marketing

Givens’ case is straightforward: retention is cheaper than acquisition. What is less commonly said out loud is that for small businesses without the capacity to outspend competitors on ads, the post-purchase experience is not a separate department. It is the marketing. Referrals and repeat purchases are downstream of a single question: did you do right by the person who already trusted you with their money?

She points to a specific failure she sees often. A customer contacts the business, something has gone wrong or they have a question, and they wait 45 minutes to get a response in a chat window. By the time someone replies, the decision not to buy again has already been made. The interaction might resolve the issue on paper, but the relationship is gone. That’s not a customer service problem in isolation. It’s a marketing problem. The acquisition cost to replace that customer is higher than whatever it would have taken to respond in a reasonable time.

The starting point she recommends is not hiring more support staff. It’s mapping what your customers experience after they buy, specifically where the friction points are that send them looking elsewhere. For most small businesses, that map reveals cheaper wins than any paid channel would.

Why AI makes a broken process worse

The episode title comes directly from Givens’ framing on AI in customer service: add AI to a broken process and you don’t fix the mess, you scale it. That is a more useful observation than most of what is currently being said about AI in business.

The version she has seen play out is this. A business has an inconsistent customer service process. Staff give contradictory answers and no one is clear on what the return policy says. They implement an AI chatbot to handle volume. Now the inconsistency runs faster and reaches more customers. The chatbot is efficiently delivering a bad experience to people who might have been fine waiting a little longer for a human who knew what they were doing.

She also raises a detail that catches people off guard: the content on your own website can break an AI implementation before it starts. If your FAQ is outdated, your product descriptions are inconsistent, or your policies aren’t written clearly, an AI trained on that content will give unreliable answers. The chatbot isn’t the problem in that case. The documentation is. Fix what you’re feeding the AI before you worry about which AI to use.

What to automate and what to keep human

Givens is not arguing against automation. Her view is that most businesses default to “automate the things that take the most time” when the better question is “automate the things where a human adds no value.” Those are not the same list.

Routine, predictable queries are good candidates: order status, hours of operation, standard FAQs. These are interactions where the information is fixed and the customer just wants a quick answer. Handling them with automation frees up human capacity for the situations where it counts.

Where she is clear about keeping humans in the loop is anything emotionally charged. A complaint from a customer who has had a genuinely bad experience is not a volume problem to be managed with a bot. It needs someone who can acknowledge what happened and respond in a way that doesn’t feel scripted. A templated reply to a real grievance doesn’t resolve anything. It confirms to the customer that no one is paying attention. Could this interaction end in a refund request or a public review? Those are the ones a person should own.

The jargon that sank her first launch

Early in running CX Collective, Givens positioned her services using the language she had grown up with inside corporate environments. The terminology was precise to anyone who had worked inside a large customer experience team. To small business owners who had never had such a team, it was incomprehensible. She launched, and it didn’t land.

The fix was not a rebrand. It was a translation. She changed how she described the work to match how her prospective clients talked about their own problems. Not “customer experience infrastructure” but “what is costing you repeat customers and what would stop it.” The core offer stayed the same. The language around it became legible to the people she was trying to reach.

It’s a specific failure mode for anyone who has been inside their industry long enough that the jargon has become invisible. The language that signals expertise to a peer often registers as noise to a buyer, and the longer you’ve been in it, the harder it is to hear. Givens fixed it by listening carefully to how her clients described their problems, then using that language back to them.

Why she turns down clients who’d happily pay

Givens is specific about the work she won’t take on: engagements where the fit isn’t right, even when the client would happily pay. The reason isn’t philosophical. It’s operational. When a client’s approach doesn’t line up with what she’s there to fix, or when their willingness to act on her recommendations is limited, the outcomes she can deliver are limited too. In customer experience work, where success depends entirely on the client implementing changes, a resistant client doesn’t just make the engagement harder. They make it impossible to show results. Those limited outcomes go on her track record regardless of whose fault they are.

She walks through a specific situation in the episode where the misalignment became clear enough that continuing wasn’t worth it. The money would have come in, but the cost in time and reputational risk was higher than the revenue justified. Most advice about saying no to bad-fit clients frames it as a values conversation. Givens treats it as a business calculation, which makes it considerably more useful.

The burnout warning sign she ignored

On burnout, Givens is candid: she ignored the signs for longer than she should have, and it took an employee naming it directly before she acknowledged what was happening. The pattern is recognisable. The pace that was necessary when building the business becomes normalised. You stop noticing it isn’t sustainable because it has been the baseline for long enough.

The point she makes that most burnout conversations skip: if the business only functions when you’re running beyond capacity, that is a structural problem, not a personal failing. Willpower and better habits won’t fix it. The model itself needs to be redesigned so it doesn’t depend on you being unwell to keep it running. By the time someone else has to name what’s happening, the problem has usually been building for months.

Givens’ point about coordination applies past customer service. Marketing has the same failure mode: bolt more channels onto a function no one owns and you scale the disconnect, not the revenue. McKee Creative’s Economy Package runs paid ads, email, SEO, and strategy as one coordinated function for exactly that reason — one strategist across the whole picture, not four contractors scaling four separate messes.

About Ty Givens
Ty Givens is the founder of CX Collective, a consultancy focused on customer experience strategy for growing businesses. She has 25 years of experience building customer support operations for fast-growing brands.

Website: cxcollective.com

LinkedIn: linkedin.com/in/tybryantgivens

Want to be a guest on The Jo Show?
If you’re a business owner with a story worth sharing, Jo would love to hear from you.
Apply to be a guest here.