The verdict in three sentences
In San Francisco, corporate real estate heads, landlords and investors now ask AI assistants like ChatGPT questions they used to put to their network: which advisor represents tech tenants, who knows the SoMa market, who can handle a sublease. A firm shows up in those answers when its public information is consistent, factual and picked up by third parties, and it drops out when its site is vague or stale. The market charges USD 3,000 to 10,000 for an audit and USD 2,500 to 8,000 a month for ongoing work (2026 ballpark), and its value can only be shown through a dated record of the answers.
What buyers ask AI assistants in 2026
San Francisco's office market is going through high vacancy and widespread lease renegotiation, which pushes companies to compare more advisors than before. The questions look like this:
| Profile | Question put to an AI assistant | What the answer needs to be able to cite |
|---|---|---|
| CFO of a scale-up | "Which firm represents tenants for 20,000 square feet of office space in San Francisco?" | Tenant representation practice, deal sizes |
| Tech real estate lead | "Who can help me sublease part of our office in SoMa?" | Sublease experience, neighborhoods covered |
| Landlord of a Financial District building | "Which brokers know the Class B office market?" | Published market research, named team |
| Family investor | "Who advises on buying lab buildings in the Bay Area?" | Life sciences specialty, geographies |
| Founder | "Is now the right time to renegotiate our San Francisco lease?" | Recent, dated market notes |
| COO | "Independent advisor or national brokerage, which should we pick?" | Clear positioning, how conflicts are handled |
| Procurement lead | "Which commercial real estate advisors in San Francisco hold a DRE license?" | Verifiable mention of the team's licenses |
Why some firms appear and others do not
AI assistants build their answers from public information. They cite a firm more readily when its name, address, specialties and team are described the same way on its site, its Google profile, LinkedIn and professional directories. Third-party mentions weigh heavily: regional real estate press, professional associations, conference speaking slots. Clear, factual pages (who you represent, which asset types, which neighborhoods) summarize better than a tagline. Structured data helps machines read that information, and freshness counts: a market note dated 2023 does not help in 2026.
Three limits specific to the sector: listings must be real and current, transactions can only be cited if public or approved by the client, and broker communication falls under state real estate licensing rules (in California, the Department of Real Estate). On data, the CCPA as amended by the CPRA applies to many businesses that collect data from California residents, website accessibility is a frequent subject of ADA-based suits (a WCAG 2.1 or 2.2 AA target is common practice) and CAN-SPAM governs commercial email. This is not legal advice.
What the work covers and what it costs
In broad terms, the work covers a baseline of current answers, making public information about the firm consistent, clear reference pages on services and team, work on third-party mentions and regular tracking. The market is young: the ranges below are 2026 ballparks drawn from typical Bay Area agency quotes and public agency rate pages.
| Service | 2026 range | What moves the price |
|---|---|---|
| One-off AI visibility audit | USD 3,000 to 10,000 | Number of questions, services and markets covered |
| Monthly program | USD 2,500 to 8,000 a month | Tracking frequency, content volume |
| Rebuild of reference pages | USD 4,000 to 15,000 one-off | Number of pages, market research to write |
| Specialist PR | USD 3,000 to 7,000 a month | Regional or national targeting |
| Quarterly tracking only | USD 1,000 to 3,000 a quarter | Number of questions tracked |
At Kolonell, this work is on quote, after a diagnostic tailored to your specialty.
How to measure without wishful thinking
Measurement rests on a dated record: a fixed list of questions, asked on set dates at regular intervals, with three indicators. The share of answers that mention the firm, the accuracy of what is said (specialties, neighborhoods, team) and the sources used. Add inbound enquiries that name an AI assistant as their starting point. Nobody can promise an AI assistant will recommend your firm: a serious provider commits to a scope and transparent measurement.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Mini case study
Illustrative example: "Embarcadero Tenant Advisors", a 14-person firm focused on tenant representation. Its baseline record of 30 questions cites it in 3 answers, or 10%, and two answers wrongly credit it with a property management business. It commissions a USD 4,500 audit followed by six months of tracking at USD 3,500, so 4,500 + 21,000 = USD 25,500. With a hypothetical USD 120,000 fee on a mid-size lease, the spend is covered by 25,500 ÷ 120,000 ≈ 0.21 of a mandate. The firm will repeat the same record at 3 and 6 months.
FAQ
Do commercial real estate clients really use AI to choose an advisor?
Increasingly for shortlisting, especially in Bay Area tech companies. The final decision stays human, but a firm missing from an initial list of 3 to 5 names often never gets the call.
How long before the answers change?
Allow 2 to 6 months after public information is corrected, with no certainty. Answers also vary from day to day, which is why tracking 20 to 40 fixed questions matters.
Can we cite our deals to get picked up more often?
Only deals that are public or approved in writing by the client. A page of 6 to 10 verifiable references beats a long list nobody can check.
What budget should we start with?
A one-off audit at USD 3,000 to 10,000 gives the baseline and the priorities. A monthly program only makes sense once public information has been corrected.
Can an agency guarantee that ChatGPT will recommend us?
No. AI answers cannot be controlled; any promise of that kind is a red flag.
Find out how AI assistants describe your firm. We record what AI assistants like ChatGPT say about your firm today, side by side with your San Francisco competitors. 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.
