top of page

Why Isn't AI Saving My Small Business Time Yet?

Writer: Zh Dast Sari
Zh Dast Sari
Oct 28
6 min read

Why isn't AI saving my small business time yet? If you bought a tool this year, opened it twice a week, and your days feel exactly as crowded as before, you are not behind. You are in the majority the surveys just counted — owners who use AI, and owners whose work actually runs on it, turn out to be two very different groups.


Three separate surveys landed this month, and they tell the same story from three different angles. I run NexAIS (https://www.nexais.ca/) in Richmond Hill, where we put an AI receptionist, AVA, onto small businesses' phone lines. This gap — between having AI and having it wired into the day — is the one I watch owners hit in person, so let me walk through what the numbers say and what closing the gap looks like in practice.


Three surveys, one awkward gap


Goldman Sachs published its latest small business report on October 22. Among the 2,473 owners surveyed — all graduates of its 10,000 Small Businesses programme, so a motivated group, not a cross-section of every shop in the country — 86% now use AI, up from 56% two years ago. Only 15% say it is fully part of their core operations. Read those two numbers together. Nearly everyone has it. Almost no one runs on it.


Thryv's 2026 adoption survey, which questioned 561 owners who are not Thryv customers, puts general adoption lower — 66%, up from 55% a year ago — and then asks the blunter question. Seventy percent of owners say they need more training to use what they already bought. Among those using AI, 92% say it saves them time and 70% say it added revenue in the past year. Those gains are self-reported, not audited, and still the direction is consistent: the tool works when it gets used properly, and most owners will tell you, honestly, that they are not there yet.


The Small Business & Entrepreneurship Council's October check-up adds the last piece. Fifty-eight percent of small business employers now use AI agents — software that does a multi-step job, not just a chatbot window — regularly or occasionally, and the typical AI-using business runs five separate AI tools. Five logins. Five subscriptions. Five things nobody quite owns.


Goldman's own response tells you where it thinks the problem sits. Alongside the report it announced $500 million to expand its small business programme and open a Small Business AI Institute, built with Anthropic, Google, Meta and OpenAI. Half a billion dollars, and the pitch is not new software. It is training and implementation help. When the banks start spending on lessons instead of licences, the lesson is the shortage.


Using AI is a login. Wiring it in is a job description.


Here is the difference as I see it on phone lines, because the phone makes it visible.


A tool that is merely "used" waits for someone to remember it. A person opens it, asks it something, copies the answer somewhere else, and the benefit depends on that person, that day, remembering. Nothing breaks when they forget. Nothing happens either.


A tool that is wired in has four things a login does not: a trigger, a set of rules, an owner, and a number it is judged by. It starts on its own when the event happens — a call comes in, a form is sent, closing time passes. It works from rules somebody wrote down: these are our services, these questions get asked in this order, this kind of call goes straight to a person. One named human is responsible for it. And there is a count, kept monthly, that says whether it earned its money.


That is exactly how we have to build a receptionist, which is why the distinction is not abstract to me. Every AVA is designed around the information that one business gives us — its services, its booking questions, its rules for what gets passed to a person right away. Hand a caller a generic script and you have bought a polite voicemail box. The 15% in Goldman's survey are not using better software than everyone else. They are the ones who wrote the rules down.


The phone line is where you can see the difference in a week


Most AI use is hard to measure, which is part of why it drifts. Did that summary save time? Hard to say. The phone line is stricter, and that makes it a good first place to wire something in properly, because the count fits on an index card:


- How many calls came in after hours this month?

- How many of them became bookings or call-backs?

- How many of those callers would have reached a competitor before they reached you?


If your current setup cannot answer the first question, you do not have a phone system problem. You have the same gap the surveys found, sitting on the one line where customers arrive ready to spend. A missed call can be a missed customer — and unlike a chatbot tab, you can count exactly how often it happens.


Three moves that close the gap in about a month


1. Pick one job, and write its number down before you change anything. Not "use more AI." One job: after-hours calls, quote follow-ups, review requests, the Monday report. Count how it goes this week — how many, how long, how many dropped. Without that first count, every later claim about saved time is a guess, yours and the vendor's alike.


2. Write the rules while the job is still small. Who can see the recordings and the customer records — named accounts, not one shared password taped to the monitor. What the tool may do on its own, and what it must stop and ask a person about first; in our builds, anything that sends a message or moves money waits for a human yes. A small tool like [AFFILIATE: 1Password Business — owner inserts tagged link] costs less than cleaning up one shared login that walks out the door with a departing employee. This is the page of the setup most owners skip, and it is the page that decides whether the other 85% in Goldman's survey ever trust the tool with real work.


3. Count for thirty days, then decide like an owner, not a subscriber. Compare the count from move one. Keep it, fix the rules, or cancel it — all three are respectable answers. What the surveys describe, under the polite wording, is a fourth answer most businesses are currently giving: keep paying, keep dabbling, never count. Five tools at once is how that happens. One job, wired in and measured, beats five logins every month of the year.


What wired-in costs on a phone line


Our prices are flat, and I will put them in the open the same way I ask owners to count their calls. Starter is $197 a month (https://crm.nexais.ca/pricing/ai-receptionist) and is the receptionist on its own. Professional is $399 a month and adds the CRM and app, where bookings, orders and customer history sit in one place — that is where restaurants start, because a restaurant needs the booking system, not just a voice answering. Grow is $599 a month and adds website and Instagram management on top.


The trial is set up to force the measurement question early, because that is the honest test: seven days free, running your own AVA — built from your information, on your line. Your card is taken at signup, nothing is charged during those seven days, and charging starts on day 8 only if you stay. At the end of the week you have a count, not an impression: calls answered, bookings taken, after-hours callers who would otherwise have got your voicemail.


If you would rather start smaller, call the demo line, +1 (437) 529-0400, at whatever hour you read this. It answers at noon and at midnight, and listening to it will tell you more about what "wired in" sounds like than another survey will. Questions about your own line? Our contact page (https://www.nexais.ca/contact) reaches a real person on our team.


Disclosure: Some links in this article are affiliate links. If you buy through them, we may earn a commission at no extra cost to you. We only recommend tools we genuinely believe help small businesses.


Sources


- Goldman Sachs, "The AI Economy Is a Small Business Economy" (survey of 2,473 programme alumni; $500M commitment and Small Business AI Institute announced October 22, 2026), read via Clayocraft's October 24, 2026 action plan — https://clayocraft.com/2026/10/goldman-sachs-small-business-ai-institute-2026-action-plan/

- Thryv, "2026 AI and Small Business Adoption Survey" (561 owners, fielded April 2026), as reported by CPA Practice Advisor and summarised by PureBrain, October 2026 — https://purebrain.ai/blog/66-of-small-businesses-use-ai-70-say-they-need-more-training/

- Small Business & Entrepreneurship Council, "Scale Without Size" — October 2026 Small Business Check Up Survey results (October 14, 2026) — https://sbecouncil.org/2026/10/14/scale-without-size-digital-tools-and-platforms-are-expanding-the-reach-and-power-of-small-businesses/

 
 
 

Recent Posts

See All

Comments


nexais-logo-full-horizontal-C1ABDukH.png

© 2026 NexAIS Inc. All rights reserved.

bottom of page