Competitor intelligence for small business: what AI can and cannot see

Most large companies have someone whose job includes watching competitors. Crayon’s 2026 State of Competitive Intelligence survey (a vendor report, and one that mostly surveys software companies big enough to employ a dedicated intelligence team) found that 80 per cent of those teams now use AI to produce sales-facing competitive content, up from 61 per cent the year before.

Most small businesses have nothing. Competitor intelligence at the SME level is usually a customer mentioning a rival’s quote, a mate spotting new signage on a ute, and an anxious scroll through a competitor’s website the night before you set your own prices.

That gap used to be rational. Watching competitors properly took hours of tedious reading every week, and an owner’s hours were better spent on customers. AI has changed the labour cost of the reading. It has not changed what is knowable, and it has not changed the law. This post is about all three.

What is AI-assisted competitor intelligence?

It is the routine monitoring of what your competitors publish in public, with AI doing the reading and a human doing the judging. There are no secrets and no informants involved. The raw material sits on the open web, and there is more of it than most owners realise.

Five public sources do most of the work for a small business:

Pricing pages. What a competitor charges, how they package it, and what quietly changes. A price rise, a new “from” price, or a service tier that disappears is a strategic decision made visible.

Positioning. The words on their homepage, the services they add or drop, the case studies they publish. A business tells you where it is heading through what it chooses to say about itself, and changes over time say more than any single snapshot.

Reviews. Google reviews of your competitors are a free, timestamped record of what their customers praise and complain about, and of how the business responds. Recurring complaints in a rival’s reviews are a map of what you could credibly promise to do better.

Tenders won. If your competitors serve government, their wins are on the public record. Under the Commonwealth Procurement Rules, Australian Government contracts worth $10,000 or more are published on AusTender, generally within 42 days of signing. In this state, Tenders WA lists recently awarded WA government contracts. That is the winner’s name, the buyer, and the contract value, published as a matter of policy.

Job ads. A competitor advertising for a second estimator, a service coordinator, or their first marketing hire is telling you about capacity and direction. Job ads are written to attract, so they are unusually candid about what a business is building.

The AI’s role is the part that used to make this unaffordable: reading those sources on a schedule, comparing them with last month’s versions, and producing a short brief. What changed, what it might mean, what is worth verifying. The judging stays with you.

What can it honestly tell you?

Direction and pattern, not ground truth. One month of monitoring tells you very little. Six months tells you which services a competitor is pushing, which way their prices are moving, whether they are hiring for a capability you do not have, and which complaints keep recurring in their reviews.

The practical value is usually positioning. If three local rivals all carry review complaints about slow quotes and missed callbacks, then answering every enquiry the same day is not a platitude, it is a documented gap you can build your pitch around. That connects intelligence to something measurable: enquiries, quotes, and wins, not a folder of screenshots.

From our own experience: I have built intelligence briefs on Perth small-business markets entirely from public sources, and the value was almost never one dramatic discovery. It was the aggregation of boring detail that nobody had bothered to assemble in one place. Owners consistently knew less about their competitors than they assumed, and the surprises were mostly about pricing structure and service scope, not strategy.

The same public surface is now being read by machines on the buyer’s side too, which is worth understanding in its own right: Google is no longer the only way your customers find you. Everything you can see about a competitor, an AI answering a customer’s question can see about you.

What can it not tell you?

Anything the competitor has not published. Margins, actual revenue, win rates, client lists, whether they are profitable, and whether that new hire signals growth or someone quitting: none of it is in public sources, and any tool or consultant implying otherwise is guessing.

The public signals also mislead in specific ways. Pricing pages go stale. Job ads describe aspirations, not reality. A competitor with 14 reviews is a sample too small to say much about their service. And award notices tell you a tender was won, not whether it was won profitably.

There is a subtler failure mode: overfitting to competitors instead of customers. If a rival drops their prices, monitoring tells you it happened. It does not tell you that matching them is right, and for many small businesses it is not. Competitor intelligence is an input to judgement. Treated as a steering wheel, it produces businesses that copy each other into the same crowded position.

Where is the legal and ethical line?

Public information, gathered honestly, is the line. The three rules that follow from Australian regulator guidance are specific enough to act on:

Public means no logins. Reading pages a business publishes to the open web is what publishing means. Creating an account on a competitor’s client portal, sharing a mate’s login, or pulling data from behind any authentication wall is on the wrong side, and typically breaches the site’s terms of use before you get near other law.

Watch the business, not its people. In their October 2024 concluding joint statement on data scraping, the Office of the Australian Information Commissioner and 16 international privacy regulators were blunt: personal information does not lose privacy protection just because it is publicly accessible. Monitoring a competitor’s pricing and positioning is business information. Bulk-collecting profiles of their staff or customers is personal information, and that is where privacy law starts to bite.

No misrepresentation. Phoning a competitor pretending to be a customer, submitting fake quote requests, or approaching their staff under a false identity is deception, whatever it yields. The ACCC’s guidance on false or misleading claims is written mostly about advertising, but its core principle applies to conduct in trade or commerce generally: honesty is not optional, and intent is no defence. Perth adds its own enforcement mechanism. This is a small market, and you will eventually meet these people at an industry event or across a networking table.

The working test I use: if you would be uncomfortable explaining to the competitor’s face how you obtained a piece of information, do not use it.

Do you actually need this yet?

Not always, and it is worth being honest about when. If you are booked out for months, your constraint is capacity, and competitor intelligence will only tell you things you cannot act on. If you have not yet fixed how you answer your own enquiries, fix that first, because it compounds and this does not.

Competitor monitoring earns its place when you are actively competing for the same customers: quoting against known rivals, bidding on tenders, or deciding where to position a new service. In that situation, a sensible setup for a small business is deliberately modest. A written list of competitors and sources, a monthly AI-generated brief of what changed, and one standing question: does anything here change what we do next month? If the answer is no for six consecutive months, stop doing it. An intelligence process that never changes a decision is a hobby.

Competitor intelligence is one of the four capabilities covered on our growing revenue with AI page, alongside customer targeting, customer insights, and campaign intelligence, and it works best feeding the other three. If you want to talk through whether it would earn its keep in your business, start with a conversation.

Frequently asked questions

What is competitor intelligence for a small business?

It is the routine monitoring of what competitors publish in public: pricing pages, website positioning, customer reviews, government tender awards, and job ads. AI reads those sources on a schedule and produces a short brief of what changed, and the owner judges what it means. It uses no secrets or insider information, only material already on the open web.

Is it legal to monitor a competitor’s public website and prices in Australia?

Reading what a business publishes openly is legal and routine. The legal risks sit elsewhere: accessing anything behind a login, bulk-collecting personal information about a competitor’s staff or customers (Australian privacy regulators have stated that publicly accessible personal information is still protected), and misrepresenting yourself, such as posing as a customer to extract information.

Can AI find out a competitor’s revenue or client list?

No. Public-source monitoring can only surface what a competitor has published. Revenue, margins, client lists, and win rates are not public for private companies, and any tool claiming to reveal them is estimating or guessing. What AI genuinely does is read the sources that are public far more consistently than a busy owner ever could.

How often should a small business check on competitors?

Monthly is enough for most small businesses. Public signals like pricing changes, review patterns, and job ads move slowly, and the value comes from comparing months over time, not from daily alerts. The useful test is whether the monthly brief ever changes a decision; if it has not for six months running, pause the exercise.