Save time and money: real examples of AI in use with the numbers, honest limits where it won't work, and one step you can take this week.
Written by Maksym Zaletskyi, 20+ years as an operations manager and programmer, building and automating processes in real businesses. In the issues I break down IBM, IKEA, Klarna, C.H. Robinson and others. I write from the position of someone who has put changes like these in place himself.
A main story with numbers and company names. Sometimes it's a breakdown of what another company did. Sometimes it's a shift in the market you haven't noticed yet. Sometimes it's a loud headline taken apart fact by fact. Always something looked at closely, never general advice.
The real limits of AI. Where it genuinely helps, and where you're better off leaving it alone, so you don't waste time, money or your reputation.
New AI tools with prices. I skip whatever is loudest right now. I focus on what you can actually try this week and judge for yourself whether it fits a small business.
Typical email example
IBM automated 94% of its HR routine work. CNBC's headline for this: "Companies replaced workers with AI, and now they regret it."
How can both things be true at once?
Imagine a restaurant that buys an industrial dishwasher and lets go of two human dishwashers. A month later it turns out the machine can't get burnt-on grime off pots - those still need soaking by hand. They hire one dishwasher back.
Is the dishwasher a failure? Quite the opposite - it's still washing 90% of the dishes. The owner just didn't think ahead about what it can't do, and who covers that. AI didn't fail here - they drew the line in the wrong place.
After a similar experience, CBA bank said it plainly: "we misjudged the roles" - not that the AI didn't work. In other words, they didn't fully understand what their call center employees actually did. And once the AI bot couldn't keep up, they realized the people were already gone.
In my experience, this is the typical case. Most owners don't know what their people actually do all day. They know the work gets done, more or less. And that invisible 6% of activities (often more) is exactly what AI can't handle.
IBM automated employee Q&A, standard procedures, repeat requests - everything with one clear-cut answer. What's left is the 6% of situations with no single right answer.
AI replaces anything with a clear answer or a standard procedure:
"How much does delivery cost?"
"What's the return procedure?"
"Send me a confirmation"
A person is needed where the answer depends on the situation:
an upset customer - are they genuinely angry, or working an angle?
a new situation the script never covered
a customer who doesn't need an answer, just needs to be heard
For big companies, the mistake hurts but isn't fatal - lay off 2,000, hire back 300, write a press release. For a small business, the cost is disproportionately higher. Let your front desk person go, the AI bot can't keep up, customers leave - and you can't undo that.
Two practical rules before you automate:
First, understand what your people actually do. Have them log every task for a week and sort it into "always the same" and "lots of judgment calls" buckets. The first bucket is your automation candidates.
Don't let anyone go before the pilot results are in. Run the AI, give it a few months, look at the numbers - then decide. (I know about this kind of premature layoff from friends too.)
As for headlines like "companies regret it and are bringing people back" - I wouldn't rush to that conclusion. 94% of routine work automated isn't a failure. It's a new division of labor. Just not the one anyone planned for. We're living through a time when "delegating" is getting a whole new meaning.
How IKEA didn't lay off 8,500 people - and made €1.3 billion. When IKEA's AI bot took over nearly half of all customer calls, the obvious move was cuts. The company did the opposite: they asked what the bot couldn't do - and found out customers wanted help with home design. They retrained the people who would've been laid off into design consultants, and that new line of business now brings in a €1+ billion. Klarna (in the same article) went the other way - cut staff, got a wave of complaints, and quietly started hiring back.
Only 1 in 5 companies that added AI to service cut headcount. Gartner surveyed 321 customer service leaders and found an uncomfortable truth: AI frees up an average of 5.5 hours per agent per week - but most companies never replanned what agents should do with that time. So the time gets saved, and nothing else changes. As one ops director puts it in the piece: 'AI frees up capacity, but nobody redefines what to do with it.'
HappyFox closed $1M in deals with an AI agent that cost under $20. The company has 4 salespeople and $20M in annual revenue. When they realized closed support tickets were full of hints about customers' new needs, they built an agent called Rex to read them and flag the team. No salesperson would ever comb through those tickets that way - there's just no time for it. If you have a support desk, a customer inbox, or a chat, there's probably something similar sitting in there.
A beauty salon put AI on the phone - and started getting more 5-star reviews. Every Saturday at 11am, the front desk is juggling the register, the client in the chair, and a ringing phone - and the phone clearly isn't the priority. Two salons in Florida fixed this with an AI receptionist that answers instantly and books straight into the schedule. +$9,000 in revenue, 16x ROI. And the unexpected part: customers started praising in reviews how easy it was to get through on the phone.
Microsoft 365 with Copilot. If your team already works in Word, Excel, and Outlook, an AI assistant has been built in by default since July 1. Not a separate add-on - a standing feature at $23.50 per user per month. So: drafting an email, summarizing a spreadsheet, fixing formulas, writing up meeting notes - all of that now lives in the tools you already have open.
Canva AI 2.0 with Connectors. Canva now pulls data straight from your Gmail, Zoom, or HubSpot and puts together a post, presentation, or proposal in your style on its own. So instead of "I should really make that meeting summary look nice" - it's already done. Canva Pro from $15/month, rolling out in June-July.
AI Overviews in mobile Google Drive. Search from your phone for "what's in our catalog" or "where are the files with the new supplier prices" and instead of a list of 15 documents, Gemini gives you a straight answer pulled together from several files. If you have Google AI Pro/Ultra or Workspace Business/Enterprise, it's already rolling out to your account - nothing to install.
Mark by Airtop. Give it your website's link, and it studies your business, customers, and competitors, then in a day proposes a marketing plan and launches agents that go find leads, run Google Ads, and track competitors on their own. For a business with no budget for a marketing agency, this isn't a replacement for a marketer - it's a way to get your first structured campaign without hunting for a contractor. Free to start, then from $26/month.
Have you already sketched out which of your employees' tasks are always the same, and which have a lot of exceptions? I'm curious what you find - write and tell me what you're thinking.
Maksym.
A new issue every one to two weeks, so it stays worth reading and doesn't clutter your inbox. Takes 3-5 minutes.