The verdict in three sentences
For an international arbitration and corporate law firm in Melbourne, AI assistants like ChatGPT have become a shortlisting step: a general counsel, a board or a foreign firm asks them who has experience with a seat, an institution or a sector. A firm appears when its public information is consistent, factual and echoed by third parties, and drops out when its site is vague or outdated. The only way to judge is a dated record of answers to the same questions, tracked over time.
What clients now ask AI assistants
Buyers in this niche are general counsel at mining, energy or infrastructure groups, funds, sometimes states or public entities, and foreign firms looking for an Australian counterpart. Their questions look like these.
| Question put to the AI | Who asks it | What the AI must be able to find |
|---|---|---|
| 'Which Melbourne firms handle investment arbitration in the mining sector?' | General counsel of a mining group | Sector expertise pages, partners involved |
| 'Australian firms experienced in ICC and SIAC arbitration for construction disputes' | Board, project owner | Institutions practised, dispute types, no confidential detail |
| 'Who can represent a Chinese or Japanese company in an arbitration seated in Melbourne?' | Foreign firm | Working languages, teams, network |
| 'What is [firm name]'s reputation in international arbitration?' | General counsel doing checks | Consistent public profiles, third-party mentions |
| 'Lawyers specialising in investor-state energy disputes in Australia' | Public entity, fund | General experience described, publications |
| 'How much does an international arbitration represented from Melbourne cost?' | Exporting SME | Clear page on fee models, no promises |
| 'Which Melbourne partners sit as arbitrators?' | Opposing counsel, institution | Up-to-date partner profiles |
None of these questions is about a slogan: they are about verifiable facts.
Why some firms appear and others do not
AI assistants like ChatGPT build answers from what they find consistent and repeated online. For an arbitration firm, that covers:
- precise expertise pages by sector and dispute type, rather than a generic 'Litigation' page;
- up-to-date partner profiles (institutions practised, languages, arbitrator appointments where public);
- third-party mentions: professional publications, conference talks, bylined articles;
- consistency between the site, professional profiles and law society directories;
- structured data describing the firm, its offices and its partners;
- freshness: a partner who left two years ago and is still cited signals disorder.
Professional conduct rules for Australian lawyers restrict advertising, and the Australian Consumer Law prohibits misleading claims: no success rates, no invented rankings, no confidential matters. This is not legal advice: your ethics partner remains the judge of what can be published.
Kolonell built the website of TR Signing Group, a signing services company in New Jersey: an example of professional services where clear public information matters, with no result figures to share.
What this work covers and what it costs on the market
In broad terms, AI visibility work for a firm includes a baseline record of answers, identifying false or missing information, aligning public content (site, profiles, directories) and regular monitoring. The market is young; the amounts below come from typical agency quotes and are converted approximately (1 AUD ≈ 0.65 USD).
| Service | Range in AUD (2026 ballpark) | Approximate USD equivalent | Note |
|---|---|---|---|
| Initial diagnostic (20 to 40 questions) | 4,000 to 15,000 AUD | 2,600 to 9,800 USD | Depends on languages and partners tracked |
| Monthly answer record | 1,000 to 3,000 AUD per month | 650 to 1,950 USD per month | Same question list, archived |
| Content alignment | 2,000 to 6,000 AUD per month | 1,300 to 3,900 USD per month | Expertise pages, profiles, structured data |
| Expert legal writing | 800 to 2,500 AUD per piece | 520 to 1,600 USD per piece | Reviewed by the partner concerned |
| Combined SEO and AI support | 5,000 to 12,000 AUD per month | 3,250 to 7,800 USD per month | Multi-office firms |
At Kolonell this work is on quote, based on the number of questions, languages and partners.
How it is measured, and what cannot be promised
Measurement relies on a fixed list of questions, put to AI assistants like ChatGPT on a fixed date, with the answers archived. Three things are read: the share of answers naming the firm, the accuracy of what is said (partners, institutions, sectors, offices) and the visible sources. What counts is the trend over six to twelve months. Nobody can promise that an AI will recommend a firm; what can be done is to make public information clearer and check, record after record, what changes.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Data collected through site forms falls under the Privacy Act 1988; the small business exemption below 3 million AUD turnover has exceptions, to be checked with your adviser.
Mini case study
Illustrative example. Collins Lane Disputes, 9 partners in Melbourne, handles construction and energy arbitrations. The baseline record on 30 questions shows the firm named in 3 answers out of 30 (10%), one of which presents as a partner a lawyer who left in 2024.
- Diagnostic: 6,000 USD.
- Monthly record: 12 × 1,000 = 12,000 USD.
- Content alignment for 6 months: 6 × 1,500 = 9,000 USD.
- Year 1 total: 6,000 + 12,000 + 9,000 = 27,000 USD.
The goal is measurable without any promise: the partner error disappears from the records, and sectors are described accurately in the answers that name the firm.
FAQ
Do arbitration clients really use AI assistants to choose a firm?
Mostly for a first list and a quick check, before peer recommendations. An inaccurate answer can be enough to drop a firm from a list of 3 to 5 names.
Should we name our matters to be cited more?
No, unless they are public and the client agrees. General descriptions (sector, institution, order of magnitude of the amount where publishable) are often enough.
How long before we see a change?
Usually 3 to 9 months, depending on how quickly the sources used by AI assistants are updated. A monthly record on 20 to 40 questions makes it visible.
What budget should we plan for year one?
A diagnostic of 2,600 to 9,800 USD, then monitoring of 650 to 1,950 USD per month, plus any content (2026 ballpark). A mid-sized firm often spends 15,000 to 35,000 USD over the year.
Can a provider guarantee that ChatGPT will recommend us?
No, and such a promise should worry you. A provider can improve the clarity of your public information and measure change on a dated record.
Find out how AI assistants describe your firm. We record what AI assistants like ChatGPT say today about your firm, partners and sectors, then cost the priority corrections. Request a diagnostic quote: reply within 48 hours.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.
