The Google knowledge panel is the most misunderstood piece of real estate in search. Founders want them. Vendors sell them. Most of the advice written about them is either wrong, outdated, or a disguised pitch for services. This post is the real process, based on what actually works across dozens of panels I’ve watched trigger over the past four years.

What a knowledge panel is

The knowledge panel is the box that appears on the right side of Google search results (on desktop) or at the top (on mobile) when you search for certain entities. Type “Stripe” into Google and you’ll see a knowledge panel with the company logo, founding year, founders, headquarters, and a handful of other facts. Type “Patrick Collison” and you’ll see a personal panel with his title, company, and a short description.

The panel is Google’s way of saying “we know what this entity is.” It’s pulled from Google’s knowledge graph, which is a structured database of entities and their relationships, built primarily from Wikipedia, Wikidata, and Google’s own web crawling and entity extraction.

For a business, a knowledge panel is three things at once. First, a credibility signal — users see the panel and assume the entity is legitimate. Second, an AEO asset — language models trained on the web are far more likely to correctly identify and describe an entity that has a knowledge panel, because the panel itself is part of the training data. Third, a brand anchor — the panel centralizes the facts about your company and makes it harder for incorrect information to dominate search results.

The qualification question

Not every entity qualifies for a knowledge panel. Google’s knowledge graph has internal thresholds for what counts as a notable entity worth displaying. The thresholds aren’t published, but the observable pattern is that an entity needs to meet some combination of the following:

Companies that meet all four criteria reliably trigger panels within months. Companies that meet three of four sometimes trigger panels and sometimes don’t. Companies that meet two or fewer almost never do.

The work of knowledge panel building is essentially the work of moving your entity from the “meets two criteria” category to the “meets four criteria” category.

The four-stage path

From experience, the path to a knowledge panel goes through four stages, each of which takes time and can’t be skipped.

Stage 1: Establish the entity on structured data sources. This is the foundation. You need to create or update the canonical structured data about your entity in the databases Google’s knowledge graph reads.

For a company, that means: claim and complete the Google Business Profile if applicable, create a Crunchbase profile with complete founding details, create a LinkedIn company page, add structured data (JSON-LD) on your own website with Organization schema, claim or create a Wikidata entry.

For a person, that means: LinkedIn profile with complete work history, Crunchbase personal profile if applicable, author pages on publications where you’ve contributed, Wikidata entry.

Wikidata is the single highest-leverage structured source. It’s an open, editable database that Google’s knowledge graph reads directly. Creating a Wikidata entry takes about 30 minutes and the entry persists indefinitely once accepted. Most companies and founders don’t have Wikidata entries and adding one is the single fastest-moving lever in the whole process.

Stage 2: Build citation density in authoritative sources. Google won’t trigger a panel from structured data alone. It needs to see the entity mentioned in multiple authoritative sources (news articles, trade publications, established blogs, educational institutions, industry reports) with consistent descriptions.

This is the work that overlaps most with traditional PR and press outreach. Every piece of earned media coverage contributes to the citation density. The target is not volume of press but variety of authoritative sources with consistent framing.

Five news articles in five different publications, each describing your company the same way, are worth more than 50 articles in the same publication. Diversity of sources is what the knowledge graph looks for.

Stage 3: Wikipedia (if warranted). Wikipedia is not required for a knowledge panel in 2026, but it remains the single most powerful signal when it’s possible. A Wikipedia article about your company or founder will trigger a panel on its own most of the time.

The catch is that Wikipedia has strict notability guidelines. You can’t create an article about your own company. If you do, it gets flagged as promotional and deleted. What you can do is wait until you’ve accumulated enough independent coverage that an uninvolved Wikipedia editor would consider your company notable, then either wait for someone to create the article organically or engage an experienced Wikipedia editor to review the notability case and create a draft for community review.

Most companies reach Wikipedia-notability territory only after they’ve done Stages 1 and 2 thoroughly. Stage 3 is optional and should not be chased prematurely. Companies that try to force Wikipedia articles before they’re notable end up with deleted articles and flagged accounts, which actively harm their entity profile.

Stage 4: Wait, monitor, and request corrections. After the first three stages are done well, knowledge panels typically trigger within 3 to 12 months. The trigger itself is silent — one day the panel just appears.

Once the panel appears, the work shifts to monitoring and corrections. Check the panel regularly for factual errors. If the panel shows incorrect information (wrong founding year, wrong founder name, outdated description), there’s a “Suggest an edit” link on the panel itself. Submit corrections with citations to authoritative sources. Google reviews and updates panels based on these submissions, though not immediately.

The Wikidata deep dive

Because Wikidata is the single highest-leverage move, it deserves more detail.

Create an account at wikidata.org. Search for your company or your own name to see if an entry already exists. If yes, improve it. If no, create a new entry.

For a company, the core properties to add are:

The important thing is to add references for each claim. Each property should cite a source — a news article, a Crunchbase profile, a corporate filing, the company’s own about page. Uncited claims get removed by other editors. Cited claims stick.

A well-referenced Wikidata entry with 15 to 20 properties is usually enough for Google’s knowledge graph to start treating the entity seriously. Combined with the other stages, it’s often the tipping point that triggers the panel.

What doesn’t work

Things you’ll see sold or suggested that don’t actually work.

Paid knowledge panel services that promise guaranteed panels. Most of these either submit spammy Wikipedia articles that get deleted, fabricate citations, or simply take payment and do nothing. There is no legitimate paid path to a knowledge panel. If the panel triggers after you pay a vendor, it’s because the work they did overlapped with real citation building, and you could have done it yourself.

Keyword stuffing your company description everywhere. The knowledge graph cares about entity-level signals, not keyword density. Repetitive identical descriptions across ten press releases don’t help more than one well-written description in five authoritative sources.

Buying Wikipedia articles. Illegal under Wikipedia’s terms, frequently detected, and a fast way to get your company flagged as promotional. The deleted article is worse than no article.

Social media profiles alone. LinkedIn, Twitter, and Facebook profiles contribute minor signal but are not sufficient on their own. The knowledge graph needs third-party authoritative coverage, not just self-published profiles.

Timeline expectations

Realistic timeline for a company starting from zero entity recognition:

Personal knowledge panels (for founders, authors, experts) take longer — usually 12 to 24 months — because Google’s notability thresholds for people are higher than for businesses.

The cost-benefit honest take

Getting a knowledge panel takes real work and real time. For a company still searching for product-market fit, it’s probably not the right investment. For an established business with $1M+ in revenue and a credible founder story, the panel is worth the six to twelve months of effort because the AEO and credibility dividends compound for years afterward.

The companies I’ve worked with who got the most value from knowledge panels are the ones that treated the panel as a downstream benefit of legitimate PR and entity work, not as the goal. They did the press outreach because it was good PR anyway. They created the Wikidata entry because it was easy. They fixed their schema markup because it was hygienic. The panel showed up as a byproduct.

The companies that treat the panel as the goal and try to force it through shortcuts usually either fail to get a panel or get one that gets suspended months later. The real process is slower, harder, and more reliable.