← All insights
Getting StartedBy Albany AI Consulting

What AI Can Actually Do for Albany Small Businesses (No Hype Edition)

A plain-English look at the AI use cases that reliably pay off for Capital Region small businesses — and the ones that don't.

If you run a business in the Capital Region, you've heard some version of "you need AI" roughly a thousand times this year. What almost nobody tells you is which AI, for what, and whether it will actually pay for itself.

Here's the honest version.

First, the number everyone quotes is probably wrong

Depending on which survey you read, somewhere between 15% and 58% of small businesses are using AI right now. Both figures are real. They're measuring different things, and the gap between them is the most useful thing in this article.

The U.S. Chamber of Commerce reports that 58% of small businesses say they use generative AI, up from 40% the year before. But the JPMorganChase Institute, measuring AI actually embedded in producing goods and services, puts overall firm adoption at about 17.7% — and only 15.1% among the smallest firms. The U.S. Census Bureau found that fewer than 20% of firms with four or fewer employees report using AI, and that adoption among firms under 20 employees didn't grow measurably over a recent six-month stretch, even as larger firms kept climbing.

That gap is the difference between someone here uses ChatGPT sometimes and AI is doing a job in our operation. Almost everyone is in the first category. The businesses getting real returns are in the second, and getting there is a much smaller step than the hype implies.

The use cases that reliably pay off

1. Answering the same questions over and over

Every business has a set of questions that make up most of its inbound calls and emails: hours, pricing, availability, "do you service my area?" An AI assistant on your website or phone line can handle these instantly, around the clock, and hand off anything complicated to a human. For most service businesses, this alone recovers hours per week.

There's good evidence about who benefits most. In a study of 5,179 customer support agents published through the National Bureau of Economic Research, researchers found that access to an AI assistant increased issues resolved per hour by 14% on average — but by 34% for novice and low-skilled workers, with minimal impact on the most experienced ones. That pattern matters for a small business. AI does not make your best person better. It brings everyone else closer to your best person, which is exactly what you want when your best person is the bottleneck.

2. Paperwork that follows a pattern

Invoices, intake forms, work orders, insurance documents — anything where a person reads a document and types the contents into another system is a strong automation candidate. Modern document AI reads these with high accuracy, and the time savings are easy to measure: count the hours your team spends re-keying information today.

The benchmark numbers here are unusually concrete. Research from Ardent Partners puts the average cost of processing a single invoice at $9.40, taking 9.2 days — while best-in-class organizations do it for $2.78. If you process a few hundred invoices a month, that spread is a real line item, and it is the kind of thing you can calculate for your own business in about twenty minutes.

3. First drafts of anything

Quotes, follow-up emails, job postings, social posts, review responses. AI shouldn't send these unsupervised, but it can produce a solid first draft in seconds that a person approves or edits. The math is simple: a task that took ten minutes now takes one.

4. Never letting a lead go cold

Speed-to-response is one of the strongest predictors of winning local business. The most-cited research here is the Lead Response Management study led by Dr. James Oldroyd — worth knowing it dates to 2007, though nothing since has overturned it. Across more than 15,000 leads, it found the odds of qualifying a prospect dropped 21-fold when response time slipped from five minutes to thirty, with a fourfold drop between five and ten minutes.

You do not need AI to respond in five minutes. You need something to respond in five minutes, and at 7pm on a Friday that something is not going to be a person. This is usually the cheapest workflow to automate and the fastest to show a return, which is why it's often where I start.

The use cases that usually don't (yet)

  • "Fire and forget" content marketing. AI-generated content with no human review and no strategy behind it doesn't rank and doesn't convert.
  • Fully autonomous customer service. Letting AI resolve complaints without oversight is how you end up in the local news. Draft-and-approve beats full autonomy for anything sensitive.
  • Big-bang AI transformations. This is the expensive one. Reporting on MIT's Project NANDA research found that 95% of enterprise generative AI pilots failed to produce measurable P&L impact, with only 5% of custom tools surviving the jump from pilot to production. Those are enterprise numbers, not small-business ones, but the lesson transfers cleanly: the projects that die are the broad ones with no specific job to do.

How to think about the decision

Ask three questions about any process you're considering automating:

  1. Does it happen often? Weekly or daily, not quarterly.
  2. Does it follow a pattern? If you could write instructions for a new hire, AI can likely follow them too.
  3. Can you measure the time it takes today? If yes, you can measure the payback — and payback is the only justification that matters.

If a process passes all three, it's worth a serious look. If it fails the third one, that's not a reason to skip AI — it's a reason to spend a week measuring before you spend a dollar building.

If you'd like help running that analysis on your own operation, book a 20-minute working call. Tell me which workflow eats the most staff time and I'll tell you honestly whether it's worth automating — including when it isn't. If you'd rather see the whole picture before committing to a build budget, the AI Operations Audit ranks and costs every workflow in your operation and hands you a written plan you own.

Sources

Wondering what AI could do for your business?

Twenty minutes. One workflow. A straight answer.

Book a 20-minute working call