The verdict in three sentences
General counsel, boards and public entities looking for an arbitration firm in Toronto now put part of their shortlisting questions to AI assistants like ChatGPT, often before they open a legal directory. A firm shows up in those answers when its public information is consistent, factual and confirmed by third parties, and a partner's reputation inside the profession does little if it is not written down anywhere. The work runs at a 2026 ballpark of 1,500 to 4,000 USD for a diagnostic, then is measured with a dated record of answers to the same questions, and no serious provider can promise a recommendation.
What do buyers ask AI assistants before they call a firm?
The client base of a Toronto arbitration practice is rarely local in the strict sense: TSX-listed mining companies, Canadian pension funds investing abroad, European groups with an Ontario subsidiary, sometimes states or public bodies. Their questions to AI assistants are long and contextual, mixing seat, institution, sector and language. They rarely look like a three-word Google search.
| Question asked of the AI assistant | Who asks it | What the answer needs to rest on |
|---|---|---|
| "Which Toronto firms have an ICC arbitration practice in mining?" | General counsel of a listed miner | Mining and energy sector page, institutional experience described |
| "Which Canadian firm represents investors in ICSID arbitration against a state?" | Counsel at an infrastructure fund | Investor-state arbitration page, named partners |
| "Toronto or New York as the seat for an energy supply contract?" | CFO of an industrial group | Signed, dated analysis on seat selection |
| "Which Toronto partners argue arbitrations in both French and English?" | French group with an Ontario subsidiary | Bios with working languages stated plainly |
| "How much does an international commercial arbitration run from Toronto cost?" | CEO of an exporting company | Page on fee models and the stages of a proceeding |
| "Which firm can get a foreign award recognized in Ontario?" | Foreign correspondent counsel | Content on recognition and enforcement of awards |
| "Corporate law firm in Toronto for a cross-border acquisition with an arbitration clause" | Board of directors | Visible link between the M&A and arbitration practices |
Two findings come up again and again in diagnostics. Questions often name an institution (ICC, LCIA, SIAC, HKIAC, ICSID) that the site mentions only inside a PDF brochure. And the working language, a deciding factor for a French-speaking client, appears only in the partner's downloadable CV.
Why some firms appear and others do not
AI assistants like ChatGPT cross-check what they find. A firm whose website, directory listings, professional profiles and publications tell the same story is easier to cite correctly. A firm whose lead partner carries three different titles depending on the source is much harder.
| Signal | What the AI assistant can draw from it | Common state of law firm websites |
|---|---|---|
| Partner bios consistent everywhere | Identity, specialty, office | Titles and dates that differ between site and directories |
| Pages by type of arbitration and institution | Actual scope of experience | A single generic "Litigation and Arbitration" page |
| Third-party mentions (publications, conferences, directories) | External confirmation | Very few traces outside the firm's own site |
| Signed, dated, updated analysis | Current expertise | Last article published more than two years ago |
| Structured data (organization, people, offices) | Machine-readable facts | Missing or limited to the address |
| Working languages stated clearly | Fit for a foreign client | Languages buried in a PDF |
| Accessible, fast pages | Content that can actually be read | Key content locked in scanned documents |
Professional conduct rules weigh on this work. The Law Society of Ontario regulates the marketing of legal services: no misleading statements, care with past results and with the word "specialist". Content has to stay factual, with no invented rankings and no promised outcomes. On the data side, contact forms fall under PIPEDA, newsletters under Canada's anti-spam legislation (CASL), and accessibility of Ontario websites under the AODA. This is not legal advice: sign-off belongs to the partner in charge of compliance.
What the work covers and what it costs on the market
In broad terms, it means checking what AI assistants say about the firm today, fixing inconsistencies in its public information, making the factual pages (partners, practices, institutions, languages) clear and current, strengthening its presence with credible third parties, then tracking the change. The market is young: the ranges below reflect typical quotes from specialist agencies in North America, with no official benchmark.
| Service | 2026 ballpark (USD) | Approximate local equivalent (CAD) | Note |
|---|---|---|---|
| AI visibility diagnostic (30 to 60 questions) | 1,500 to 4,000 | approx. 2,000 to 5,500 | First dated record |
| Upgrade of partner and practice pages | 4,000 to 12,000 | approx. 5,500 to 16,500 | Depends on number of partners |
| Structured data and consistency of external listings | 1,000 to 3,000 | approx. 1,400 to 4,100 | Often paired with the diagnostic |
| Monthly support (content, tracking) | 2,000 to 6,000 per month | approx. 2,700 to 8,200 per month | 6-month commitment is common |
| Quarterly record only | 800 to 2,000 | approx. 1,100 to 2,700 | To track without producing |
Indicative conversion at roughly 1.37 CAD to 1 USD. At Kolonell this work is on quote, after a diagnostic that sets the scope. Kolonell built the website of TR Signing Group, a signing services company in New Jersey, for instance: a case where clear public facts matter as much as design.
How to measure without fooling yourself
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Serious measurement rests on a dated record: the same questions, asked under the same conditions, at regular intervals. Three indicators come out of it: the share of answers that mention the firm, the accuracy of what is said (right partner, right office, right institutions) and the sources cited. A firm can be mentioned often and described badly, which is a problem in itself for a client who checks. Inbound inquiries that name an AI assistant as their starting point round out the picture, provided someone asks on the first call.
Mini case study
Illustrative example: Northbay Arbitration Partners, 14 lawyers in Toronto, with commercial and investor-state arbitration practices. The initial record covers 40 questions; the firm appears in 3 answers, or 7.5%, and two of them place a partner in an office he has left. The budget under consideration: 3,000 USD for the diagnostic and 8,000 USD to upgrade the pages, so 11,000 USD. If a single arbitration mandate represents 150,000 USD in fees, it covers that budget about 13.6 times. The assumption guarantees nothing; it gives the scale of the decision.
FAQ
Do AI assistants like ChatGPT replace legal directories?
No, they add to them and often use them as a source. A general counsel may ask an AI assistant 5 or 6 questions, then check the resulting list against directories and peers.
How long before the answers change?
As a ballpark, allow 3 to 6 months after the pages and external listings are fixed. The pace depends on the sources AI assistants draw on and how often they refresh.
Can we be sure of being recommended?
No. No serious provider can promise that; the work targets the likelihood of being cited accurately, measured on a record of 30 to 60 questions.
Do Law Society rules limit what we can publish?
Yes, lawyer marketing in Ontario has to be accurate and not misleading. Factual content on institutions, seats and languages is generally compatible, subject to internal sign-off.
What budget should we plan to start?
A diagnostic alone sits between 1,500 and 4,000 USD as a 2026 ballpark. It lets you decide afterwards whether a 4,000 to 12,000 USD page upgrade is justified.
Find out how AI assistants describe your firm. We record what AI assistants say today about your firm, your partners and your arbitration institutions, then price the corrections worth making. 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.
