Somewhere in Minnesota tonight, somebody with a legal problem and no money is going to type it into an assistant instead of calling a lawyer. That’s been true for a couple of years. What’s new is what happens next. The assistant doesn’t hand back ten blue links for the person to sort through. It writes an answer, and to write it, it goes and takes passages off pages like mine.
That changes who I’m writing for. And “optimize for AI!” isn’t an answer. That slogan has already spawned a whole consulting industry selling the same thin content it sold last decade under a new three-letter acronym.
Here’s my claim, and you can check it: the reader that matters most for a law firm’s published writing is now a retrieval system, and a retrieval system rewards depth, verbatim accuracy, pinned authority, stable addresses, and explicit statements of what an authority does not cover. In other words, it rewards what makes legal writing good and punishes what made SEO content bad. Writing for machines that read on behalf of people pushes you back toward real quality. That’s a happy accident, and it’s the most interesting thing about this whole shift.
Search asks which page. Grounding asks what can be said.
Don’t take my word for it. The clearest explanation of the change I’ve read came from Bing’s own engineering team in May 2026, and if you publish anything, read the whole thing.
Here’s how they frame it. Traditional search asks “which pages should a user visit?” Grounding an AI answer asks “what information can an AI system responsibly use to construct a response?” Those sound like the same question. They aren’t, and the post lays out the split in a table. In traditional search, the “unit of value” is the document (page). In grounding, it’s “groundable information (discrete, supportable facts with clear provenance).” In traditional search, “imperfect ranking is tolerable; recovery is easy,” because a person scans the results, skips the bad ones, and course-corrects. In grounding, “errors can compound across reasoning steps.”
Then comes the line that should stop any lawyer cold. Among the valid outcomes for a grounding system, Microsoft lists “answer when supported; abstain when evidence is insufficient.” And on freshness: in search, “stale content degrades ranking”; in grounding, “a stale fact produces a misleading response.”
Read that as a lawyer and it sounds familiar. A system that has to decide whether the evidence supports the assertion, flag conflicting sources instead of quietly picking one, and know when to keep its mouth shut — that’s not a search engine’s set of values. That’s a research memo’s. Bing’s own summary: “search indexing was built to help humans decide what to read. Grounding indexing is being built to help AI systems decide what to say.”
If the machine reading your page is deciding whether your sentence is safe to repeat, everything about what makes a page valuable flips.
Your new reader is a hostile cite-checker
Walk through what that reader actually wants. Every litigator already knows the list.
Provenance. Microsoft’s version: grounding requires “high-quality source identification and attribution so users (and downstream systems) can verify what was used and follow the evidence when needed.” Ours: a pin cite. A page that says “under Minnesota law, you generally have six years” gives a retrieval system nothing to stand behind. A page that names the section and the subdivision and quotes the operative clause gives it something it can commit to and attribute.
Factual fidelity through chunking. This is the piece worth understanding, and Microsoft doesn’t sugarcoat it: the processes of “breaking content into retrievable chunks and transforming it for fast lookup can distort page substance in ways that never appear in any ranking signal.” Nobody reads your page. It gets cut into pieces and the pieces get ranked. So a paragraph that needs three paragraphs of setup to be accurate will eventually get quoted wrong. That gives you a drafting rule: each paragraph should survive being lifted out of the page. Microsoft’s own guidance says the same thing from the other side — content is eligible for selection when it has “self-contained phrasing: sentences that make sense even when pulled out of context.”
Lawyers have a name for a sentence that changes meaning when you separate it from what came before. We call it a quote taken out of context, and we spend our careers objecting to it. Now the retrieval layer does it to every page on the internet, mechanically, at scale, and the only defense is writing paragraphs that are true standing alone.
Conflict. A grounding index “must detect and represent conflict; silent arbitration risks confident wrong answers.” This is the part I like best, because conflict is where legal writing actually lives. The published Court of Appeals decision that cuts against the trend, the 2026 amendment that hasn’t made it into the codified text yet, the split between districts — a content farm flattens all of it into one confident sentence. A page that says “the Court of Appeals has said X; the supreme court has not addressed it” is more useful to a system trying to figure out what it can responsibly say, not less.
Thin content was always bad. Now it’s invisible.
The old game rewarded a 600-word page titled “What Is a Personal Injury Claim in Minnesota?” that never cited a statute. Ranking was a contest between documents, and a document could win on keywords, links, and freshness signals without having anything in it.
Under grounding, that page has nothing to grab. There’s no discrete, supportable fact with clear provenance to pull out. It can rank and still never get cited, because the thing being selected isn’t the page.
The alternative is expensive, and the only way I can show you what it costs is to show you what this firm actually built. As of today the news library on this site is 321 published articles running to roughly a million words. Three hundred of those 321 cite at least one Minnesota statute; the corpus has 3,093 separate references to “Minn. Stat.” The statutory quotations come from the Revisor’s raw HTML, not through a summarizing fetch layer, for reasons I’ve written about separately — a summarizer paraphrases, and a paraphrase inside quotation marks is a fabrication. Case quotations get read from the opinion, with pin cites resolved against star pagination. Every article closes with a sources block naming each authority and the exact proposition it supports, down to the page.
That’s not a content strategy. It’s the ordinary standard of care for legal writing, applied to publishing. The only reason to bring it up is that for ten years the market paid you not to do it, and the market has changed its mind.
Say what the statute doesn’t say
This is the move I’d push hardest, because it’s the one thin content can’t make.
Search this site’s articles and the pattern is everywhere: 22 places that say a statute “does not define” a term, 80 that say some authority “does not say” something, three that describe a statute as silent, four that report finding no Minnesota case on a point. One article works through the honesty-testing statute and stops to note that the prohibition reaches “any test purporting to test the honesty” of an employee, that the statute doesn’t define “test,” and that the phrase therefore names a purpose, not a technology. Another reports that the Court of Appeals “found no Minnesota case allowing recovery in the absence of direct physical injury to the spouse.”
To somebody skimming, admitting a gap looks weak. That’s exactly why marketing copy never does it. To a system deciding whether it has enough to answer, a clean statement of where an authority stops is one of the most valuable sentences on the page. It’s the difference between a source that lets the assistant answer confidently inside a limit and a source that invites it to overstate.
It’s also how you talk to a client who deserves to know the answer is unsettled.
A URL is a citation format
If your page is going to get cited, its address has to behave like a citation.
Articles here live at dateless slugs — /news/<slug>/ — with the publication date in frontmatter instead of the URL, so revising a guide in 2027 doesn’t mean moving it. When something has had to move anyway, a 301 goes into the redirects file. There are 41 rules in it now, most of them pointing old addresses at current ones.
The sneakier piece is lastmod. The sitemap integration will happily stamp every URL with the build time, and the comment I left in the config when I fixed it says why that’s a problem: it “tells a crawler that 212 articles changed every time one of them is edited — and it learns to stop believing the file.” So the sitemap’s lastmod comes out of each article’s own frontmatter — the updated date when a piece has been revised, otherwise its publication date — and the same date drives dateModified in the article’s schema.org markup. Topic hub pages carry CollectionPage and ItemList schema listing their members, so a crawler sees nine subject clusters instead of 321 unrelated pages, and every article’s publisher points at a single firm entity node declared once instead of restated 321 times.
None of that is clever. It’s bookkeeping. But freshness is one place where grounding and search really do differ in what’s at stake. Stale ranking is an annoyance. A stale fact in a composed answer is a wrong answer with your name on it as the source. And a lastmod field a crawler has learned to ignore is worse than none.
The plumbing, and what it actually buys you
Everybody overclaims on this part, so I’ll stick to what’s documented.
IndexNow is a real protocol with published mechanics. You host a text file at your root whose name and contents are both your key. You POST JSON with host, key, and a urlList to a participating endpoint, and you may “submit up to 10,000 URLs per post.” A 200 means accepted, a 202 means received with key validation pending, a 403 means the key didn’t check out. The protocol’s own documentation is careful about what success means: “The HTTP 200 response code only indicates that the search engine has received your URL.” The participating endpoints listed at indexnow.org today are the IndexNow global endpoint, Amazon, Bing, Naver, Seznam.cz, Yandex, and Yep — and a submission to any one of them gets shared with the others. Google isn’t on the list.
Deploying this site ends with a script that submits the new URLs. It checks that the key file is actually reachable before it submits anything, because IndexNow rejects the batch if it can’t fetch the key, and otherwise you’d never know.
What it buys you is where people get sloppy. Microsoft says IndexNow “helps ensure that AI systems reference the most current version of a page when generating answers,” and separately that “Powered by Bing’s search index, experiences like Microsoft Copilot, Microsoft Start, and others handle billions of queries each month.” Microsoft’s new AI Performance dashboard reports citations across “Microsoft Copilot, AI-generated summaries in Bing, and select partner integrations.”
Notice what Microsoft doesn’t say: which partners. You’ll read confident claims that Bing’s index is what backs this or that third-party assistant. I can’t find a primary source for that, so I won’t write it. What I can verify is narrower and still worth acting on: pinging IndexNow gets a new page into Bing’s index in minutes instead of weeks, and Bing’s index is, on Microsoft’s own statement, what Copilot and Microsoft Start run on.
The other assistants crawl for themselves, which makes robots.txt an editorial decision, not a technical one. OpenAI’s documentation is explicit: OAI-SearchBot “is used to surface websites in search results in ChatGPT’s search features,” and “sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers.” Anthropic documents three separate agents — ClaudeBot for training data, Claude-User for user-initiated fetches, and Claude-SearchBot for search indexing — and says that disabling Claude-User “prevents our system from retrieving your content in response to a user query.”
This site allows all of them and names them in robots.txt. Funny thing, though: that named list is already out of date. Claude-SearchBot isn’t on it, because it was documented after the file was written. The only reason that doesn’t matter is that the wildcard rule allows everything, and that’s what’s actually doing the work. A list of named crawlers in a robots file is a statement of intent. The default is the policy.
I publish an llms.txt and I don’t buy it yet
The llms.txt proposal — Jeremy Howard’s, first published September 2024 and now in a second version modified August 10, 2026 — asks sites to serve a markdown file that gives an agent a curated map: an H1 with the site name, a blockquote summary, then H2 sections with lists of links and short descriptions. The reasoning holds up. Web pages “wrap information in navigation, ads, and JavaScript,” turning them back into clean text is “difficult and imprecise,” and “every wasted token costs time and money.” The idea is that the file itself “stays small enough to fit in context,” with the detail behind the links.
This site generates one at /llms.txt from the content collections at build time, plus a /llms-full.txt carrying the source markdown of every published page.
Two problems.
The evidence for adoption is all on the publishing side. The spec page documents thousands of sites publishing the file, documentation platforms generating it automatically, and Chrome’s Lighthouse auditing for one. It documents no assistant vendor committing to read it. Google’s John Mueller, asked about it in June 2026, called it “purely speculative for now,” and added the line that lands: “the file has existed for years, yet none of the AI systems use it.”
And my own version breaks the spec’s central premise. /llms.txt here is 158 KB and /llms-full.txt is 6.9 MB. Neither one is “small enough to fit in context” in the sense the proposal means. The v2 spec’s stronger recommendation — serving a clean .md twin of each page at the same URL, with rel="alternate" and rel="describedby" link relations pointing to it — is something this site doesn’t do at all.
So here’s where I land. It costs nothing to generate, it’s generated well enough, it’s unproven, and if I’m wrong about it I’ve wasted forty lines of build code. That’s the whole case. Anybody selling you an llms.txt engagement is selling you something else.
Nobody can measure this yet
The most important thing I can tell another lawyer about writing for retrieval is that you can’t currently tell whether it’s working. Don’t trust anybody who says they can.
Assistant citations don’t show up in your analytics as referrals in any consistent way. The one real instrument I know of is Bing Webmaster Tools’ AI Performance view, released in public preview in February 2026. It reports total citations, average cited pages per day, page-level citation counts, and the “grounding queries” — “the key phrases the AI used when retrieving content that was referenced.” Microsoft hangs its own caveats on every one of those metrics: the citation count “does not indicate placement or presentation within a specific answer”; cited-page averages do “not indicate ranking, authority, or the role of any page within an individual answer”; the grounding-query data “represents a sample.”
And that tool only covers Microsoft’s surfaces plus unnamed partners. There’s nothing like it for ChatGPT, Claude, or Gemini that I can point you to. Bing’s own engineers put it plainly: “What makes this hard is not the technology gap — it is the measurement gap.”
So no chart. You get what I built, why the reasoning holds, and what would knock it down.
What would prove me wrong
I’m wrong if what gets pulled into AI answers turns out to be short, structured, thin content instead of depth with pinned provenance — if a page of Q&A blocks with no authority beats a three-thousand-word piece that quotes the statute, because the extraction layer rewards format over substance and never gets far enough into the page to notice the substance.
That’s not a strawman. Microsoft’s own optimization guidance pulls the other way in places. It says to “avoid long walls of text” because they “blur ideas together and make it harder for AI to separate content into usable chunks,” and it recommends lists, tables, and Q&A pairs, which “assistants can often lift word for word.” It even advises being “cautious with em dashes,” on the theory that overuse “can confuse sentence structure for machines.”
I’m keeping my em dashes. But the real half of that advice corrects me, and I’ll take the correction: the reward is for structure, not for length. Depth without structure loses to shallowness with structure. A long, shapeless essay is as hard to chunk as a short empty one. Headings that name the actual question, one idea per paragraph, the quotation set off where it can be lifted cleanly, the sources at the foot — that’s the format. It’s no concession, either. It’s also how you write something a busy lawyer can use.
If, three years from now, the assistants are reliably citing content with no authority behind it and ignoring the primary-source work, I’ll have been wrong about what the retrieval layer picks, and the right move will have been schema tricks and volume. I don’t think that’s where this goes. The whole direction of the engineering — abstain when unsupported, detect conflict, attribute so the user can check — runs toward evidence.
Would an assistant that checks keep citing you?
There’s a symmetry here that I think settles it.
I’ve written elsewhere on this site about how this firm’s own assistant does legal research: it pulls statutes as raw HTML from the Revisor instead of through a summarizer, reads opinions instead of snippets, and runs a separate verification gate before anything gets used. An assistant built that way doesn’t need a law firm’s blog post about a statute. It has the statute.
So the only content worth publishing is content that holds up when a reader with the primary source open checks it. A page that paraphrases the statute loosely is worse than useless to that reader. It’s a liability the moment somebody compares. A page that quotes the operative clause correctly, cites the subdivision, notes that a 2026 amendment changed the effective date, and says which question the statute doesn’t answer — that reader keeps that page.
That’s the standard. Not “will this rank.” Would an assistant that checks the statute keep citing you.
And this is where the whole section comes back around. The person who can’t afford a consultation is going to ask an assistant anyway. Whatever that assistant tells them is going to be built out of what’s on the open web about Minnesota law. Right now a lot of that is lead-generation copy nobody ever checked against anything, and it’s getting laundered through a confident interface into something that sounds like advice.
No gatekeeper is coming to fix that. The only fix is lawyers who know the law putting accurate, cited, honestly bounded versions of it where the retrieval layer can find them. That costs the firm nothing but the work, and it’s worth more to a person with no money than any amount of free-consultation marketing.
That’s why everything in this section is free, and it’s the same reason to write the guides carefully. The audience was always people who couldn’t afford to ask. Now there’s a machine standing between you and them, and it turns out the machine has excellent taste.
Sources
- Bing Search Blog, Evolving role of the index: From ranking pages to supporting answers (May 6, 2026) — traditional search vs. grounding comparison table; “groundable information (discrete, supportable facts with clear provenance)”; abstention as a valid outcome; freshness and conflict-detection differences; chunking distortion; “the measurement gap”
- Bing Webmaster Blog, Introducing AI Performance in Bing Webmaster Tools Public Preview (Feb. 10, 2026) — citations across “Microsoft Copilot, AI-generated summaries in Bing, and select partner integrations”; total citations, average cited pages, grounding queries, page-level citation activity, and Microsoft’s caveats on each; IndexNow and answer freshness
- Microsoft Advertising, Optimizing Your Content for Inclusion in AI Search Answers (Oct. 8, 2025) — “Powered by Bing’s search index, experiences like Microsoft Copilot, Microsoft Start, and others”; parsing into modular pieces; self-contained phrasing; guidance against walls of text, hidden content, and PDF-only information; the em-dash caution
- IndexNow protocol documentation — key-file ownership verification, POST JSON format, 10,000-URL batch cap, 200/202/400/403/422/429 response semantics, and the requirement that participating engines share submissions
- IndexNow FAQ — the list of participating endpoints (IndexNow global, Amazon, Bing, Naver, Seznam.cz, Yandex, Yep); Google is not among them
- Jeremy Howard, The /llms.txt file, v2 (published Sept. 3, 2024; modified Aug. 10, 2026) — format specification, the “small enough to fit in context” design goal, the v2 markdown-twin and link-relation recommendations, and the publisher-side adoption evidence
- Search Engine Journal, Google Says LLMs.txt Is “Purely Speculative” For Now (June 2, 2026) — John Mueller: “purely speculative for now (the file has existed for years, yet none of the AI systems use it)”
- OpenAI, Bots — OAI-SearchBot surfaces sites in ChatGPT search features; “Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers”; GPTBot and ChatGPT-User distinguished
- Anthropic, Does Anthropic crawl data from the web, and how can site owners block the crawler? (updated Apr. 7, 2026) — ClaudeBot, Claude-User, and Claude-SearchBot, and the consequence of disabling each
- Site figures (321 published news articles, ~1,000,000 words, 3,093 “Minn. Stat.” references across 300 articles, 22 “does not define” passages, nine topic hubs, 41 redirect rules, 158 KB
/llms.txt, 6.9 MB/llms-full.txt) were counted from this site’s own repository and live URLs on August 20, 2026.
Commentary on legal publishing and practice technology; the opinions and predictions are the author’s, offered with failure conditions so they can be tested rather than relied upon. Not legal advice, not ethics advice, and not a recommendation of any vendor or optimization service. Nothing here reports measured results — as stated above, the measurement tooling for assistant citations barely exists. No client information appears in this article. Questions about anything here: Send us a message or 612-470-6529.