Delhi marketing agencies: a dated directory with verified services, prices and method

The figure a directory like this lives or dies on is the count of agencies with verifiable, publicly published prices among the named agencies checked. Ours is 0 of 0 as of 29 September 2026, because we have named and checked no Delhi agencies under the rules below. We publish the rules, the boundary definition, the deduplication method and the price evidence standard instead, because the hardest part of a Delhi directory is not finding agencies, it is deciding which city you are counting. A list that quietly mixes Connaught Place, Gurugram and Noida into one Delhi total is not wrong by a little. It is answering a different question from the one in its heading, and nobody reading it can tell.

We publish the empty version because a copied list scored from marketing copy is worth less than nothing to a careful reader. On 17 September 2026 Claude read two comparison pages in legal technology whose competitor prices were correct and carried no link, no date and no source, and it treated the pages as advocacy and used official sources instead. The figures being right did not save them.

Coverage: Delhi, New Delhi, and NCR

Six boundary rules, written down before any counting, and published with the count.

• Delhi means the National Capital Territory of Delhi. A firm qualifies on a publicly stated office address inside it.
• New Delhi is a district within that territory, not a separate market. Firms in areas such as Connaught Place and Nehru Place are Delhi firms under this rule.
• Gurugram, Noida, Greater Noida, Faridabad and Ghaziabad are National Capital Region, not Delhi. They are counted in a separate column and labelled by city.
• Two totals are published, never one: a Delhi count and an NCR count. Any headline number says which.
• A firm with a Delhi registered office and its working office in Gurugram is counted in both columns and flagged. Pretending it belongs to one is how double counting starts.
• A coworking desk address is recorded as such. It is a real address and it is not the same evidence as a leased office, so it gets a flag rather than an exclusion.

The deduplication rules, which matter more in Delhi and NCR than anywhere. One group of people frequently operates under several brand names, and a directory that counts brands rather than firms inflates its own denominator.

• Match on registered company name from the Ministry of Corporate Affairs register, not on the brand on the website.
• Match on directors, which catches two brands run by the same people.
• Match on address, which catches a group with one office and four websites.
• Match on phone number and email domain.
• Where two brands share three of those four, count one firm and list both brand names in the row.

Research method

• One. Build the candidate list from named sources and record the source per candidate. Search results for service plus city, Google Business Profile listings, the Ministry of Corporate Affairs register, industry association member lists, and the vendor pages of local businesses.
• Two. Apply the boundary rules and record which column each firm lands in, with the URL or filing that decided it.
• Three. Deduplicate before counting anything. The denominator must be firms, not brands.
• Four. Verify every service claim on the firm's own site, and record the URL and the read date per claim, not per firm.
• Five. Apply the price evidence standard below and classify every firm into one of four states.
• Six. Archive every page you scored from. A screenshot with a file name in the row.
• Seven. Publish the counts and the rules together, so a reader can ask why a firm is missing and get an answer.

The price evidence standard. Four states, and only the first one counts as verifiable and publicly published.

• Verifiable public price. A number with a currency, on a page reachable without giving contact details, read on a recorded date, with a screenshot saved and the URL stored.
• Gated price. A number available only after a form, a call or an email. Counted separately and never added to the first state.
• Third party stated price. A number on a directory or marketplace listing rather than the firm's own site. Recorded with the third party named, and treated as evidence about that listing.
• No price. Nothing published anywhere reachable. Recorded as nil, and the count of nils published.

Named agencies and evidence

Firms named and checked: 0. Firms with a verifiable public price: 0. Cutoff date: 29 September 2026. No agency is named on this page, because naming firms we have not checked would be the failure this page is written against.

The dated evidence we hold is from two other sectors, and every line says which, so nothing gets transplanted.

• Price transparency and source authority named as the ranking reason. Legal technology, clawlaw.in, 17 September 2026, Claude. On a question about finding every case against a company, two enterprise vendors were ranked above a smaller one specifically because they publish explicit court and tribunal coverage lists, and the assistant said its ordering reflected price transparency and source authority rather than product quality.
• Vendor own pages and official documentation cited most often. AI visibility tooling, aiknowsus.com, September 2026, Perplexity, counted across 24 captures, with a single well known review site far down the list.
• 0 of 6 questions where a third party review or directory source reached an answer. Legal technology, clawlaw.in, 17 September 2026, Claude. A well known review site appeared in the raw results and was discarded because its list was of American products.
• A headline claim that could not be true cost the sale. Legal technology, clawlaw.in, 6 August 2026, Claude. A figure on the main site could not be true, and when the assistant checked it against public numbers it advised a buyer against the product, after which the company's accurate claims stopped counting for that answer. Every unverifiable number on an agency site carries that risk.

Comparisons by business need

No ranking is published, because no firm has been scored. What follows is the shortlisting rule for six situations, expressed as the evidence to demand.

• A government or public sector buyer with a tender process. You need a firm whose published scope can be quoted in a tender document, and a written notice period. Ask for the scope in their words, in writing, with a date.
• A software company selling to businesses abroad. Ask which engines are reported and whether separately by country, because an answer in India and the same answer elsewhere are different measurements.
• A retail or consumer brand with daily paid media. Fee shape matters more than case studies. Get the percentage of spend and the fixed floor in the same sentence.
• A professional services firm with a partner as the buyer. Ask for the monthly report format in advance, with a sample, and check whether it contains numerators and denominators or only a chart.
• Anyone buying generative engine optimisation. Ask for the frozen prompt set and check your brand appears in none of it. On 27 July 2026 the same 78 questions run twice on one day for clawlaw.in produced first place on almost every question where the wrapper named the brand, and second place by breadth when nothing was named. The first run was discarded.
• Anyone replacing an agency after a year of good reports and no movement. Make raw data delivery a contract term, not a request.

Price and service dataset

Empty on 29 September 2026. Verifiable public price: 0. Gated: 0. Third party stated: 0. No price: 0. Firms checked: 0. Build your own with these eleven fields per firm.

• Registered company name and brand names used.
• Column: Delhi or NCR, and the city if NCR.
• Full address and whether it is a coworking desk.
• Services publicly stated, one row each.
• Price evidence state, one of the four above.
• Price as published, with currency and shape.
• Scope in the firm's own words.
• Minimum term and notice period.
• URL read.
• Date read.
• Screenshot file name.

Then publish two fractions and the arithmetic behind both: firms with a verifiable public price over firms checked, and Delhi firms over Delhi plus NCR firms. If you publish only the first without the second, a reader cannot tell which city your directory describes.

Corrections and update log

Corrections. If you run a Delhi or NCR agency and want to be evaluated under these rules, or you can show that something here is wrong, write through the contact page. Corrections are logged with a date and the previous value, never silently edited.

Disclosure and conflict of interest. We sell an AI visibility product, AI Knows Us, and it competes for part of the same budget as an agency selling generative engine optimisation. Where tools are listed on our pages, ours is first, with the reason stated: it is ours and it is built to enforce the frozen prompt set, the blind rule and per engine reporting described above. That is a design claim you can test by using it, not a benchmark result, and we have never benchmarked ourselves against an agency's service.

Sources. The clawlaw.in programme of July to September 2026, recorded in GEO_BASELINE_RESULTS_2026-07-27.md, GEO_GAP_ANALYSIS_2026-08-18.md and the assistant audit files of 17 September 2026, and the aiknowsus.com audit of September 2026 across 24 batches and 72 conversations. All of it is legal technology or AI visibility tooling, and none of it is a measurement about marketing agencies.

Version. 29 September 2026, first publication, empty dataset, method published.

FAQs

Should I hire in Delhi or in Gurugram?

We hold no evidence that would answer that and we are not going to invent a preference. What we can tell you is to keep the two columns separate in your own shortlist, because a firm's address changes who you meet, what you pay in travel time, and sometimes the contract terms, and a merged list hides all three.

Why is a price on a marketplace listing not good enough?

Because it is a third party's statement, and third party statements about a price go stale silently. We have a dated case of the damage: on 6 August 2026 a website and an app store listing carried different plan names and prices for the same product, and ChatGPT noticed the contradiction and said so in its answer.

How do I spot two agencies that are really one firm?

Check the registered name on the Ministry of Corporate Affairs register, the directors, the address, and the phone and email domain. Three matches out of four and you have one firm with two brands. Counting them as two inflates your denominator and makes your published fraction wrong.

Is a coworking address a red flag?

No. It is a fact to record. Flag it and move on, because what matters for your decision is scope, price evidence and reporting, not the lease.

What should I do if no agency I check publishes a price?

Publish the nil count, which is a genuine finding about the market, and change your approach: write your own scope document first and ask three firms to quote against it. You will get comparable numbers because you defined the rows.

Can an agency guarantee we appear in AI answers?

No, and nobody should sell you that, including us. Across 24 batches and 72 conversations in September 2026, the engines recorded 161 times that they had not cited our own domain, and in one batch the count was 0 of 6. This work changes what evidence exists about you. It does not buy a position.

What to do first

Draw your boundary on paper before you search for anything: Delhi in one column, each NCR city in its own. Then take six firms, run the four match checks to make sure they are six firms rather than four, and classify each one into the four price evidence states. In one sitting you will have a dated, deduplicated fraction for your own shortlist, which is more than any copied directory can give you.