If your website provider proposes blocking AI crawlers while your marketing team proposes an advertising budget, you have separate decisions to make. As of November 2026, we recommend reviewing those proposals separately. The provider documentation reviewed here describes advertising tools and website-access controls; it does not establish that buying an ad determines whether an independent answer cites your firm.
The operating question is straightforward: how should your firm manage search access while making separate decisions about advertising and training use? OpenAI documents independent search and training settings. Cloudflare describes controls for crawlers that serve both purposes. Neither source establishes that allowing access will produce a citation, recommendation or inquiry.
Short answer: separate the budget decision from the access decision
We recommend assigning separate decisions to advertising, public-page search access and training preferences. The managing partner should be able to approve one without accidentally approving the others. A request to keep attorney biographies available for search should not become an unexamined decision about training use or paid campaigns.
Our interpretation is that combining these decisions into a single question about “AI visibility” could obscure what the firm is buying or changing. An advertisement has creative material and a destination. A crawler setting addresses access or a stated use. An observed citation records what appeared in a particular response. We recommend keeping those records separate.
What the dated provider evidence establishes
This article uses provider product documentation and vendor guidance accessed on November 11, 2026. Its source class is public provider documentation, not first-party outcome evidence. It does not report an AlphaForge deployment, a test of a firm's website or measured client demand. Publication dates were not supplied for the two OpenAI excerpts. Cloudflare's guidance was published on September 15, 2026.
Advertising documentation describes creative and destinations
The retrieved Ads | OpenAI Developers documentation describes ad creation and management through its application programming interface, or API. Its multiple-image workflow attaches uploaded image identifiers to an ad, with a title, body and destination URL applying to every image. It also describes product-ad templates that receive their image and destination URL from a selected feed item.
Our interpretation is that this excerpt supports a description of advertising infrastructure, but does not establish how independent answers rank law firms, whether payment affects organic citations or how a sponsored conversation behaves. In our reading, the returned creative type named chat_card does not, by itself, establish any of those behaviors.
We recommend requiring an advertising proposal to identify the actual creative, destination page and proposed measurement method. Evaluate that proposal on its documented terms. The existence of an Ads API is not evidence that the firm must pay to remain discoverable.
OpenAI documents separate search and training settings
In its Overview of OpenAI Crawlers, OpenAI says website owners can manage OAI-SearchBot and GPTBot independently through robots.txt tags. These are website instructions addressed to crawlers. OpenAI describes them as a way to manage how sites and content work with AI. Its example allows OAI-SearchBot for search results while disallowing GPTBot to indicate that crawled content should not be used for training its generative foundation models.
OpenAI also says that when both bots are allowed, it may use results from one crawl for both purposes to avoid duplicate crawling. It notes that search systems can take roughly 24 hours to adjust after a robots.txt update. That is a provider statement about its systems, not an AlphaForge response commitment.
Our interpretation is that these controls could support different firm choices for search and training. They do not establish that an accessible page will be selected or cited in any particular answer. The excerpt supplies no law-firm citation rate or business-outcome measurement.
Cloudflare addresses crawlers with more than one purpose
Cloudflare's September 15, 2026 guidance, Have it both ways: stay discoverable in search while disallowing AI training, describes a mixed-use crawler as one crawler serving both search and training. Cloudflare announced a Disallow AI Training setting intended to preserve search access while communicating a training restriction.
Cloudflare says its Accountable designation combines capabilities available at publication with commitments to deliver others. It describes Apple, Google and Microsoft as combining available capabilities with time-bound commitments for those still in development. Its proposed controls over how much content appears in AI summaries were also described as a future goal.
Cloudflare further warns that selecting Block for mixed-use crawlers stops Applebot, Bingbot and Googlebot from reaching the site, including for search. This is vendor guidance about its controls. It is not independent proof that any configuration will improve a law firm's visibility.
A practical example: one firm, three separate approvals
Illustrative operational example: Consider an estate-planning firm reviewing public attorney biographies and probate articles. Its website contractor proposes stricter bot controls, while a marketing vendor proposes a paid campaign. This is a constructed example, not a client account or evidence of actual demand.
We recommend that the firm first document which public pages it wants available for search. Ask the contractor to identify the relevant provider controls and record how the proposed configuration implements that choice. “Allow AI” is too imprecise to serve as the approval.
We recommend recording the firm's training preference separately. If it chooses to restrict training, ask the implementer to identify the applicable provider setting and its stated limitations. Review each provider's documented controls rather than assuming that a preference expressed to one provider governs every provider.
We recommend reviewing the advertising proposal as its own purchase decision. Identify what the ad says, where it sends the visitor and how the firm proposes to distinguish campaign activity from other visits. The supplied Ads API excerpt does not establish the suitability or availability of a particular campaign for this hypothetical firm.
Our interpretation is that this separation could make troubleshooting more useful. If a public page cannot be fetched, the team can investigate access. If an accessible page is absent from a sampled answer, that observation alone does not identify the cause. Activity attributed to an advertisement does not establish an organic citation.
Recommended checks before accepting a website change
We recommend using the following as proposed acceptance checks for a scoped implementation. They are recommendations, not tests performed for this article or promises of discovery results.
- Record the intended policy. List the public pages, search-access decision and training preference. Name the person responsible for approving each choice.
- Review the effective configuration. Ask the website administrator to compare robots.txt with hosting and network-level bot controls. Record the actual settings rather than relying on the label of a dashboard switch.
- Check representative public pages. Ask the implementer to document whether the intended crawler can retrieve selected biographies, practice pages and articles. Preserve the date, method and observed result.
- Keep access findings separate from answer observations. Record what a successful retrieval check establishes within its method and conditions. Do not treat it as evidence that an answer will mention or cite the page.
- Preserve a change record. Record what changed, who approved it and how the previous configuration can be restored. Recheck provider documentation before applying a later change.
For related implementation detail, AlphaForge's guide to separating search, training and user-request crawler policies organizes policy around those purposes. Use it as a related implementation guide, rather than evidence that a particular firm has a visibility problem.
What to measure after the access review
We recommend maintaining three separate records: technical access observations, sampled AI-answer observations and first-party inquiry outcomes. Our interpretation is that combining them into one visibility score could hide which part of the process was actually observed.
For answer observations, we recommend saving the exact question, date, provider, available product settings, response and cited URLs. Mark questions written by the team as synthetic probes: they are research instruments, not evidence that prospective clients ask those questions. Record a firm mention separately from a clickable citation.
For business outcomes, we recommend using the firm's own inquiry and matter records where attribution is available. Keep a reported discovery source separate from a verified referral record. An AI mention is not an inquiry, and an inquiry is not a retained matter or collected fee.
The excerpts do not establish how independent answers select law firms, how often a firm will appear or what any appearance will produce. Our interpretation is that a documented access policy could answer a narrower, useful question: whether the firm's chosen settings match its intended treatment of public content. We recommend measuring advertising performance and organic discovery separately, without treating either as proof of the other.