HBR Asked 230 Businesses About AI and Talent. Here Is My Answer.

January. October.

Nine months ago, none of what I run today existed. Not as a plan, not as a budget line.

In January, finding new retailers for a brand meant people, lists and time. Someone downloaded a trade show exhibitor list, someone checked websites one by one, someone looked for an email, someone typed it into a spreadsheet. Our procedures lived in long documents, hard to use on the floor. I have run gift and lifestyle brands for more than 20 years. I do not code. Software of our own meant developers, an agency quote, and a year.

Today I run my brands on QVINTA SUITE, a platform I built myself with Claude doing the engineering in chat. No developers. No agency. Seven modules: the CRM, the retailer map, sourcing, stock, the shop, product specs and cash. One sign-in for the team.

Here is what one part of it, SCRAPE, shows for one of my brands. Just one. And it is not even a month old.

Another 1.7 million photos are waiting to be read. From November, with more storage in place, these numbers grow at a much bigger scale. In January, the first column alone would have been a team's year.

So when Harvard Business Review asked 230 small and midsize businesses how AI is changing the way they hire and train, I read it closely. Here is what they found, in plain words, and my answer.

What HBR Found. The Short Version.

Almost everyone is adding AI. Almost nobody feels ready for what it does to their people.

HBR Analytic Services surveyed 230 people at US businesses with 1 to 999 employees in October and November 2025, all familiar with how their company hires and trains. 76 percent say their company will use more AI in the next 12 months. Only 3 percent expect AI to change nothing.

The number that matters most: 79 percent agree AI is driving the need to upskill the people they already have. Not replace them. Upskill them.

And yet only 19 percent say their company is highly prepared to get the AI skills it needs. 56 percent expect trouble even working out which AI skills they need. That is the real gap. Not the tools. Knowing what to ask of people.

The changes are already showing. 56 percent expect to train people differently, 37 percent to look for different skills in candidates, 36 percent to keep people differently. 78 percent are working out where AI fits in the company, 49 percent expect existing roles to change, 47 percent are revisiting privacy and security rules.

The report opens with Vizrt, a live production software company based in Bergen, Norway. Melanie Grasso, a global HR partner in their San Antonio office, says AI helped her team cut 300 job applications down to a shortlist of 20 top candidates in minutes.

Organizations really need to build more muscle in forecasting the talent implications of technology improvements.

That line is from Orsolya Kovács-Ondrejkovic, a partner and associate director at Boston Consulting Group in Zurich. Her point: the question is not only "can we do this faster with AI" but "what does that do to our people."

Sources: HBR Analytic Services, The New Talent Playbook for Small and Midsize Businesses in the Age of AI, sponsored by TriNet, survey fielded Oct. to Nov. 2025, published Mar. 2026

The Real Cost Is Time. Not Money.

HBR says training is too expensive for most small businesses. From where I sit, money is not the problem.

The report is honest about the barrier. 49 percent expect trouble training their own people on AI. 43 percent say their employees want to learn AI skills but lack the resources or the chance. Courses from universities and vendors cost thousands, plus the time away from work. Grasso at Vizrt puts it simply: if they see value, they pay for a course, but it is a one-off, and the advanced ones are expensive.

Others found a cheaper path. Catherine Rymsha, a lecturer at the University of Massachusetts Lowell, says LinkedIn Learning, lunch-and-learn sessions and YouTube are a quick way to learn for businesses that cannot pay training experts. AnnMarie Brockhouse, director of operations at the Blockchain Association, says her team simply had to jump in and play around with AI, then moved on to free courses.

I agree with Brockhouse. The tools now teach you while you use them. What they cost is attention. Mine. Someone has to sit with the tool, explain the business, check the work, and say "no, that is wrong" many times. That cannot be delegated, and it is the real price.

Our procedures are the best example. We had 90 written procedures, long and hard to follow mid-task. Today they are 113 short guided cards on the phone, each one checked line by line five times before the team saw it. AI did the rewriting in hours. The checking was the work. Nobody took a course. People learn inside the tool they use every day.

Sources: HBR Analytic Services report, as above

Industry Experience Beats AI Skills.

This is the finding I agree with most. And the one most people will skip.

HBR asked: if you could hire one person, who would you pick? 52 percent chose someone with a lot of industry experience. 34 percent chose someone with equal industry and AI experience. Only 7 percent chose someone with a lot of AI knowledge.

The report calls it the "wisdom economy". AI can find patterns in huge amounts of data, but it struggles with emotional intelligence, ethics and judgment. The people who matter are the ones who can see where AI is wrong. Kovács-Ondrejkovic gives the cautionary story: a company used AI to handle simple customer requests and send the complex ones to people. The bot could not resolve the simple ones either. Calls from angry customers to human agents roughly doubled, the company lost some of its best people, and in the end the investment "wasn't worth it."

Grasso looks for soft skills like empathy in candidates, because you do not want to lose your culture. Rymsha says AI forces leaders to understand their business, their people and their competitors more deeply.

This is exactly my experience building SUITE. Claude wrote the code. I brought the rules: how buyers think, how gift products sell through, how retailers charge back, how factories close a production run. I made the calls: what to build, what to cut, when it was wrong. I brought the taste: the brand, the screens, the words. And I made the clicks only an owner can make: domains, accounts, approvals.

The new bottleneck is knowing your business. The code is the easy part now.

Sources: HBR Analytic Services report, as above

What Claude Did on My Build.

HBR says the rare skills are prompt engineering, machine learning and AI infrastructure. Here is what that looked like for a founder who has none of them.

Kovács-Ondrejkovic says people who can do prompt engineering, machine learning, agents and AI infrastructure are extremely rare, and even large employers fight over them. She is right. A small brand will not win that fight. The good news is you may not need to. On my build, Claude covered every one of those jobs.

How the work ran was simple. I asked for things in plain words in a chat. They were built and deployed the same day. Several Claude agents worked in parallel on separate parts. Health checks run three times a day and the results go to my phone. Every change is tested, logged and can be reversed.

Built the traditional way, the same scope is about 650 person-weeks: 10 to 12 people for 14 to 18 months, around $3.9 million at US agency rates. That is not a talent strategy a gift brand can have. Now it is.

Sources: QVINTA SUITE build review, Oct. 2026 | HBR Analytic Services report, as above

My Answer. One Expert Per Area. AI Does the Rest.

I will hire one expert for each area of the business. AI does the worker bee jobs wherever it can.

Sales and buyers. Operations and logistics. Finance and cash. Brand, design and product. Every area where the job is thinking and controlling, not manual labor. I need a person who knows the trade, makes the calls and checks the work. The time-consuming, mundane tasks go to AI.

Here is one example. Sales. The old way to open new accounts was a team: someone building lists from trade shows, someone checking each shop, someone finding the buyer's email, someone writing the pitch, someone chasing. Today SCRAPE reads the lists and the websites, merges the duplicates, scores the fit and finds the public email. The 2,374 shops above are already sorted. SOLD drafts the pitch. One sales expert reviews, decides which accounts are right for the brand, and spends the day on buyers instead of on spreadsheets.

That expert needs exactly what HBR's 52 percent are asking for: industry experience. Knowing which shop will sell a $40 puzzle and which one will send it back. AI does not know that. It finds the shops. The expert knows which ones matter.

HBR's own data points the same way. 61 percent hope AI streamlines operations, 57 percent hope for more productivity. Only 19 percent point to more revenue. For a brand, revenue is the point. The productivity only matters if the time saved goes to buyers.

Where Small Brands Win.

HBR ends on agility. It is the one advantage a small brand has over a big one, and AI makes it bigger.

"We're small, nimble, and in some ways, can be reactive," Brockhouse says. Those traits work in a small company's favor when new technology is moving fast. The catch, she adds, is finding people with the right skills to lead the charge. Kovács-Ondrejkovic notes the barrier to entry for this wave of AI is extremely low, and adoption is moving extremely fast.

I would put it more strongly. A big company needs approvals, procurement and an IT department to try something new. I need an idea in the morning and a chat window. If it is wrong, it is undone by the afternoon. The constraint is not budget or talent anymore. It is how well the owner knows the business and how much time the owner gives it.

Sources: HBR Analytic Services report, as above

What to Do Right Now.

If you run a small brand, this is the order I would do it in.

• ✓Start with the boring work. Lists, data entry, follow-ups, reports, procedures. That is where AI pays first and where nobody misses doing it by hand.
• ✓Hire for industry, not for AI. One expert per area who knows the trade and can tell when the output is wrong. The tools are learnable in weeks. Twenty years of buyers are not.
• ✓Budget time, not courses. Block hours every week to work with the tools yourself. Explaining your business to the AI is the training, for you and for it.
• ✓Write down your rules. How you price, who you sell to, what a good account looks like, what you never do. AI is only as good as the rules you give it.
• ✓Train inside the work. Turn procedures into short guided steps people use while they do the job. Nobody needs a certificate to follow a card.
• ✓Keep a person on every customer touch. Let AI prepare and draft. Let a person decide and send. HBR's chatbot story is the reason.
• ✓Check everything, every day. Automatic health checks, logs and a way to undo any change. Speed is only safe if you can reverse it.

The Read

HBR is right. AI is changing who small businesses hire and how they train them, and most are not ready. Where I would push further: the answer is not to chase rare AI talent you cannot afford. It is to hire the people who know your trade, give them AI for the repetitive work, and spend your own time teaching the tools how your business works.

The new bottleneck is knowing your business. The code is the easy part now.

In January I could not have built any of this. In October I run it. The difference is not money. It is what I know and the hours I put in.

Source Register.

This article is mostly a plain-language rewrite of the HBR Analytic Services report below. All survey figures and quotes are theirs; the QVINTA and SCRAPE figures are mine, from my own system on Oct. 22, 2026.