The AI Reversal: How The Cost Of AI Forced Companies To Take A Step Back

Introducing Michael Georgiou and Imaginovation

Tim (00:28): Welcome to Clutch Conversations, everyone, where we unpack real decisions business leaders face on a daily basis. I'm Tim Condon, the CRO at Clutch. You can't go anywhere without hearing about the latest news regarding AI. At first, it seemed like everything you heard was full speed ahead — at one point, Microsoft said something to the effect that all white collar jobs could be replaced in 18 months. However, we're seeing examples of companies making AI missteps that make this a much more balanced conversation. Joining me is Michael Georgiou, co-founder of Imaginovation, and he's going to talk to us about the future of AI in the workplace. Michael, why don't you give me a little bit of your background and how you guys are using AI today.

Michael (01:27): Thanks Tim and Clutch for having me — I really appreciate it. For me, I oversee the internal marketing, sales, and creative direction for Imaginovation. We've been around for over 14 years, started in 2011 with me and my business partner Pete Peranzo — he's actually my brother-in-law, so we started as a family business. I came back from Australia after my MBA, came back unemployed, couldn't find a job for seven or eight months. Pete basically said, "Hey, want to start a software development company?" I said, "I'm not a coder or anything." He said he had some developers willing to work overseas, thought it could be side income. I said why not — I had nothing to lose.

(02:57): Imaginovation is a premium B2B technology service company. We build personalized custom applications — software, web, mobile, AI — basically anything digital for other companies. We've gotten more specific about who we work with over the years. There are a lot of companies like ours out there, but I feel like the best ones have survived through the economic conditions, COVID, all of it. We hustled, took care of our customers, and reinvested back into the business.

From Humans to AI and Back

Tim (03:45): I would characterize 2024 as the year everybody thought AI was going to replace every job. Then you had examples like Klarna — one of the first to say they killed their customer experience team. Now you're hearing they're quietly rehiring. They went too far, hired back 700 or 800 people. So is AI taking jobs or are we going too far? Is AI over?

Michael (04:35): Right now, AI is in a testing phase. You're getting hit with all this dopamine — fun, cool, amazing tools, save so much time and money. Then for larger companies like IBM, Cisco, Oracle — they've laid off tens of thousands of people in a day because they've embedded AI into their systems and they want higher profit margins and lower overhead. I understand it as a business owner, but you're dealing with people who have to put food on the table for their families. I think the way some of these companies have gone about it is really wrong.

(05:38): Not all tools are good. There's a lot launching right now — the popular LLMs like Claude, ChatGPT, Perplexity, Gemini — and people are still trying to understand how to use and leverage them, especially at a B2B level. A lot of companies really don't know how to leverage AI in their organizations, especially in older blue collar industries like construction, manufacturing, logistics, and transportation. They see the buzzword and the cool factor, but they don't know how to use it. They need a lot of guidance.

How To Utilize AI And Manage The Cost

Tim (07:28): With the clients you work with, are you seeing people use AI to become more efficient and let people go — or are they finding they can use it to be more efficient without reducing headcount?

Michael (07:55): The way I'm seeing it, the clients we're dealing with want to incorporate AI into their operations — not to get rid of people. It's more about creating more value within the company, whether that's a new product to generate more revenue, or fixing manual processes, automating them, digitizing them, and allowing their team to save time and money and just be more efficient holistically. That said, I am seeing on the other side that larger companies are reducing staff — looking at something like administrative tasks that you can automate through AI and saying, "We have 3,000 admins. Let's get rid of them and have a couple managers handle it." That is happening on a much larger scale. And it's scary.

Tim (09:39): The other big thing I'm hearing about is the skyrocketing costs of tokens. There's one story about an Airbnb developer who blew through their budget in a week. Do your clients need real handholding on how to be efficient with AI usage? Is that a fixable problem?

Michael (10:14): Yes — they are seeking a lot of guidance from us on the AI tools being used within the applications we're building, which obviously exhaust tokens. Those costs are getting really expensive. We're using multiple tools here at the company, and sometimes you have so many that it can be unmanageable. You have to stay on top of it like you would an employee — you can't just let AI run and do its thing once you think it's trained. That's the problem. A lot of companies are making that mistake. They trust it too much.

(11:06): We have a quote in the company: trust but verify. Give it trust — that's the door opener — but you've got to verify and stay on top of it. I also think LLM companies like OpenAI and Anthropic are going to have to adapt and lower their fees, because spending $50,000 a month in tokens is crazy. You've got to stay on top of them and understand how to use them.

How to Introduce AI Into Your Company the Right Way

Tim (14:19): We're still in the early innings and there are a lot of companies that haven't really introduced AI into their business yet. How would you advise a leader to introduce AI while not making the mistakes others have made? You don't want to be so cautious that your competitors beat you, but you also can't bankrupt yourself.

Michael (14:59): The way we've done it at Imaginovation is we embedded AI into our culture from top to bottom — giving all of our employees access to OpenAI, Anthropic, Gemini, depending on their roles and the strengths of each LLM. But we did it slow. If you go too quickly — 100% all in — you're going to have a lot of consequences. I'd recommend going step by step. First, understand the purpose and the why behind it. What pain points is it going to solve? Why do you want to include AI in your sales processes, your QA, your product management?

(17:06): If a CEO just says, "Here's access to all these LLMs, stay under this budget, that's it" — that's a recipe for disaster. People can leak data. I saw a post on LinkedIn a couple months ago from someone who got laid off because their company went out of business after all their data got leaked. They trusted AI too much, didn't manage it, didn't oversee it. That's a big issue.

Tim (17:59): It sounds like what you're saying is it's no different than when somebody brings you in to build software. The first question is, what are you trying to accomplish?

Michael (18:12): Exactly. You have to be very intentional about it. Create a living document. Build out the process — the why, the how, the where, the when, who's going to guide you through it. You can't go from zero to a hundred. Go zero to one, one to two, step by step. And not every company needs AI for everything. You can also spend a lot of time on it and waste time. Make sure you have intention behind it.

Tim (19:15): To the company that feels like they're behind — are they behind, or are we still early? And what does it look like five years from now?

Michael (19:30): Instead of asking if they're behind, what I'd do is a business process audit. Understand first what they're trying to do and why. Then audit the internal processes — where are the leaks? Where do things need to be fixed? Most thriving companies have processes and infrastructure in place. You audit that environment, understand where the problems are, and figure out if they're ready for the possible repercussions if they don't address them. And not every company needs AI for everything. Use it with intention and purpose, and you'll start to see real value in how you're using it — because every company uses it so differently.

Rapid Fire: Personal AI, Safe Jobs & the Most Underrated Tool

Tim (21:28): Put costs aside — what everyday non-work task would you use AI to automate?

Michael (21:43): In my personal life, I've automated home tasks that my AI recommends based on what I'm teaching it. It's learning. I have an LLM on my computer with specific permissions — it gives me daily updates on things to focus on at home. For example, I've told it my wife loves when I take the garbage out on Mondays before the truck comes. It'll remind me: "Mike, you're finishing work around 5:36 — take the garbage out, spend time with your son." It even references things I told it last week. That's pretty cool.

Tim (23:22): Roles you think are safe from AI replacement five years from now?

Michael (23:29): Anything physical. I used to do boxing training — AI can send me tutorials and videos, but it's not going to replace Coach D who taught me how to actually throw a hook properly without hurting myself. On a higher level, anything requiring human ingenuity. A lot of competitors in the software development space have told me they're afraid of going out of business because of AI. My response: we just have to adapt and evolve. Think about vibe coding — a lot of startups think they can just vibe code their way to a scalable product. Studies are showing that's not realistic. You need human guidance. You need people who actually know how to do it.

Tim (25:36): Most underutilized feature of AI tools right now?

Michael (25:44): AI Slack assistants — and a lot of people don't even know these exist. This has been a game changer for us. We've built a Slack agent that creates our sales proposals. The design matches our branding, it's consistent with our brand personality, it pulls in the right case studies automatically. That used to take my team hours. Now it saves my guy five to seven hours a week when we're busy. That's a lot.

Where to Find Michael & Imaginovation

Tim (26:35): Michael Georgiou, thank you very much. Where can people find you?

Michael (26:42): Check out our company at imaginovation.net. We also launched a product called Magic Task — a gamified task management platform for gamers, entrepreneurs, and small businesses. Over 7,000 users now. Find it at magictask.io. And for me personally, search Michael Georgiou on LinkedIn — you'll find the bald-headed guy at Imaginovation.

Tim (27:20): Awesome. Thanks for doing this — great conversation, and we'll talk soon.
 

About the Author

Tim Condon Chief Revenue Officer at Clutch

Tim Condon is the Chief Revenue Officer at Clutch, the leading global marketplace of B2B service providers. Prior to Clutch, Tim served as the Chief Revenue Officer at Homesnap, the top-rated real estate app built for agents, which CoStar Group acquired in 2020. During his tenure, Homesnap grew its paying user base from 0 to over 80,000 clients. In addition, he previously served as the Director of New Ventures at The Washington Post. In this role, he built several new businesses, including The Capitol Deal, which became the third largest deal site in the DC metro area and was known for giving away 100,000 pizzas

See full profile