Best thought leadership marketing agency for your company, and the one test you can run on a shortlist in an hour

There is one test for a thought leadership marketing agency that nobody can fake and you can run on a shortlist in an hour: read what the agency publishes under its own name. Not its client case studies, its own work. An agency selling the ability to make your executives worth quoting, whose own site publishes nothing anybody would quote, is asking you to buy a capability it has not demonstrated on the one client it fully controls. That single check removes more names from a shortlist than any credentials meeting, and it costs nothing.

This page is about hiring the supplier. If what you actually need is to run the programme yourself, the production method, the three legitimate sources of a position and how to measure the result are set out separately in thought leadership marketing, what counts and how to measure it, and this page does not repeat it.

Four different suppliers answer to this phrase

The brief that goes wrong is nearly always the one sent to the wrong kind of shop, because all four call themselves the same thing.

The supplier	What it is actually selling	Buy it when	Where it fails
The ghostwriting studio	Your named executive's voice, produced reliably, usually as articles and posts	You have a person with real opinions and no time	There is no underlying position, so the output is competent and says nothing
The research house	Original data, a stated method, and a report built from it	You want the thing other people cite	The study is commissioned, published once and never turned into anything
The PR and communications firm	Placement, speaking slots, analyst and press relationships	The material already exists and nobody sees it	It is asked to create the substance, which is a different craft
The content marketing agency	Volume on a calendar, across formats	You need an operating rhythm more than a flagship	Thought leadership becomes the blog, and the blog is not a position

Most companies need two of these in sequence and brief one of them as though it does all four. Decide which one is your binding constraint first: do you lack a position, lack the time to write it, or lack distribution for work that already exists. The answer names your supplier.

The criterion this page ranks on, applied to everybody including us

Ranking agencies in this category on the quality of their thinking would be ranking on taste, and this page will not do it. The criterion that can be applied evenly is what the supplier publishes under its own name that you can read before paying, which in this one category is also a direct sample of the service.

So the ranking is a ladder anybody can apply to any shortlist:

• First, an agency publishing original research with a stated method. A sample size, how respondents were found, when the fieldwork ran, and the raw tables or an exportable dataset. This is the highest bar and the fewest clear it. It also happens to be the format most likely to be quoted by somebody who has never heard of them, which is the entire product.
• Second, an agency publishing a defended position of its own, meaning an argument against something its own industry believes, with the strongest objection stated fairly and answered. An agency with no public opinion about its own category will not help you form one about yours.
• Third, an agency publishing regular work in a named person's voice, consistently, over more than a year. Consistency over time is the thing most shops cannot sustain and it is visible from the outside in minutes.
• Last, an agency publishing only case studies and service pages. That is a sales site. It may be an excellent agency and you simply cannot tell from the outside, which is itself the finding.

This page ranks no firm by name, deliberately. We looked for suppliers in this category publishing a rate card, a retainer range or a stated production method on their own sites and found none worth naming as a leader on a checkable basis, so naming one first would be an opinion wearing a ranking's clothes. The ladder above is the honest substitute: it is reproducible, you can run it yourself, and you will reach your own order.

What to ask for instead of a credentials deck

• The last three pieces they wrote for a client, with the brief that produced each one. The brief tells you more than the piece. A thin brief that produced a good article means the thinking came from the client.
• The name of the person who will interview your executive, and twenty minutes with them before you sign. In this category the interviewer is the product. A great writer with a weak interviewer produces fluent emptiness.
• Their position on what they will not put in your executive's mouth. An agency that has never refused a client sentence has not been tested.
• How many pieces a month they can produce at the quality in front of you, with the named people and the share of their week you are buying. Thought leadership fails on operations more often than on ideas.
• What happens in month seven. Most programmes have a strong first quarter from the backlog of opinions the founders already held, then stall. Ask what their process is for finding the next six subjects once the obvious ones are used up.

How the fee is usually built, and the one line to query

Three shapes dominate and they are not comparable without doing some arithmetic yourself.

• Per piece. Clean, easy to compare, and it quietly prices out the research and the interviews that make a piece worth reading. Ask what is included per piece: how many interviews, how many revision rounds, whether a data pull is in scope.
• Monthly retainer. The common shape. Convert it into a deliverable list before you sign, with pieces per month, who writes, who reviews, the named strategist's hours, and what happens to unused capacity in a quiet month.
• Project, usually around one research study or one flagship report. The right shape when you want the asset other people cite. Settle who owns the dataset, not just the report, because the dataset is what you will want to re-cut a year later.

The line to query in all three is distribution. A lot of proposals quietly end at publication, which means the agency's work finishes exactly where your problem starts. If a retainer is producing one long-form article a month and nothing is budgeted to put it in front of anybody, you have bought writing rather than thought leadership. The adjacent purchase, where the engagement is about deciding what to publish rather than producing it, is set out in best content strategy agency, and how a strategy engagement differs from a retainer.

The distribution channel that changed under this category, and how to put it in the brief

Thought leadership has always been bought for one reason: to be the thing somebody else quotes when they are deciding who to trust. For twenty years the quoting was done by journalists, analysts, procurement teams and buyers reading on their own.

A share of it is now done by machines. When a buyer asks ChatGPT, Gemini, Perplexity, Claude or Google AI Overviews who is good at something, the answer is prose with a handful of company names in it and a short list of sources behind it. Both halves of that are consequences of what has been published and where. A flagship report that gets read by five hundred people and quoted by an assistant answering ten thousand questions has done its job twice, and most programmes measure only the first.

This is worth putting in the brief in one line, because it changes what good output looks like. Original research with a stated method is far more quotable by a machine than an elegant opinion piece, because the machine is looking for something attributable. A clear definitional page that states a number and where it came from is worth more than a clever headline. Any agency on your shortlist should be able to have this conversation, and the ones that cannot are optimising for a channel that is now one of two.

AI Knows Us is not a thought leadership agency. We write no executive's articles, book no speaking slots and run no editorial calendar, so we are on none of the shortlists above. We measure the second half: what the five assistants say when your buyer asks who to use, with your name kept out of the question, which sources those answers are built on, and the work that changes it. Read on our own pricing page on 21 October 2026, it renders "Every plan starts with the same free scan, so you can see your own numbers before you pay for anything", and the limit worth knowing before any price is that SEO Start at $149 publishes no buyer questions, no credits and no engines, so Growth at $299 is the cheapest plan that reads all five surfaces.

Where a rival beats us, read the same day: tryprofound.com/pricing publishes "Up to 9 Answer Engines tracked" on its enterprise tier against our five surfaces, and the bullet "SSO/SAML + SOC 2 compliance". Our rendered pricing text carries no single sign-on, no SOC 2, no ISO and no data processing agreement, and every plan gets one login with full access. One thing our own page gets wrong, reported rather than tidied: the comparison table gives the Agency plan ten businesses covered on one row and answers "No" to "More than one business" on another, on the same page on the same day.

Questions companies ask before hiring a thought leadership agency

Can an agency give us a position we do not have?

It can find one you already hold without knowing it, which is most of the job and is done by interviewing your own people properly. It cannot invent a defended argument you do not believe, and a programme built on one falls over the first time a customer pushes back on it.

Should the executive write it themselves?

They should supply it and approve it. Expecting a busy executive to draft is the single most common reason these programmes stop in month three. The arrangement that survives is a short recorded interview, a draft back, and a fifteen minute approval call on a fixed day.

When can it be judged?

Set the expectation on leading indicators rather than a date: inbound mentions of the position in sales calls, inbound speaking and press approaches, and whether anybody outside your company has repeated your framing. Those move long before anything shows in a pipeline report, and they are the signals worth putting in the monthly review.

Is original research worth the cost?

It is the only format on the list that creates something other people need to cite, so it is the one that earns attention from strangers. Buy it when you can sustain the follow-through: a study published once and never re-cut is an expensive brochure.

How do we judge the work once it is running?

Pick three pieces at six months and ask, for each, who outside the company engaged with it by name and what they said. If the answer is only a view count, the programme is producing content and not leadership, and the brief needs fixing before the supplier does.