WebFX, Thrive, Ignite Visibility, Directive and Percepture compared on what each one publishes about its own work

These five American agencies do not compete for the same brief, which is the first thing a buyer comparing them needs to know. WebFX and Thrive are large volume full service firms that sell breadth and process. Ignite Visibility is a performance and search firm that sells named practitioners. Directive sells a specialism, business to business software demand, and should be judged on whether your model matches it. Percepture sells public relations and digital together, which is a different product from the other four. So the comparison a buyer should run is not which of the five is best. It is which of the five is in the category your scope belongs to, and then which of those publishes enough checkable evidence to be worth a call.

The categories, and who belongs in which

Firm	Category it competes in	The brief it fits	What to verify first
WebFX	Large full service, high volume delivery	Several disciplines at once, process over bespoke	Who your named senior owner is, and their hours per month
Thrive	Full service, broad geographic coverage	Multi location businesses needing consistent delivery	Whether local delivery is a local team or a central one
Ignite Visibility	Performance and search	Paid plus organic where media is the biggest line	A redacted account view showing fee as a share of total cost
Directive	Business to business software demand	SaaS with a sales cycle and a pipeline target	Whether their model assumes a sales team you have
Percepture	Public relations plus digital	Earned coverage alongside performance	Which half of the fee is PR, and how earned coverage is counted

The five questions that make these five comparable

Each of these firms publishes a different amount about itself, so a fair comparison needs the buyer to standardise the questions rather than the marketing. Ask all five the same things in one email, before any call, and compare the written answers.

• What is the published price, or is it quote only? Either answer is acceptable. Write quote only in the cell rather than an estimate.
• Who is the named senior owner of my account, and how many hours a month do they personally have on it? This is the question that most often separates the pitch team from the delivery team.
• Show me one redacted monthly report. You are testing whether the report answers a business question or counts activity.
• What is the minimum term and what is the exit handover? Accounts, files, formats, dates.
• Which AI assistants do you read for a client, how often, and do I get the raw answer text? This is the newest question and the one most likely to produce an honest no, which is itself useful information.

We do not reprint another firm's prices, and why that matters here

Each of these five publishes or withholds its own figures and those figures change. A price copied into a comparison today becomes a false statement about somebody else's business within weeks, and an engine reading an undated price in a comparison table has good reason to discount the page carrying it. So the method here is the opposite. Build the row yourself on the day you buy, with plan, amount, currency, term, what is included, and the date you read it, from the firm's own page. That takes an afternoon and it produces a comparison that is true on the date it is dated, which is the only kind that survives a procurement review.

One American tax point belongs in the same row. Software and some service subscriptions are taxed where the customer is, and whether a given subscription is taxable at all differs from state to state. Two companies signing the identical scope on the same day can get different invoices. Ask each firm how they treat your state and put the answer in the comparison beside the price.

What an engine actually said about this set of names

These five names came to us from an engine answer, not from a list we assembled. On 22 October 2026 we put a 198 question American buyer bank to the ChatGPT and Gemini product surfaces through a data provider, with the United States fixed inside the text of each question, because the provider accepts gl, hl and location parameters and ignores all of them. The brand batch is excluded from every number below.

Across 64 blind American readings the two engines produced 2,866 citation URLs on 1,143 distinct hosts. Where a supplier was named, there were 53 namings, and every one of them was a software product rather than an agency: Semrush 14, HubSpot 12, Ahrefs 9, BrightLocal 5, Peec 3, Profound 2, SE Ranking 2, Otterly 2, and Scrunch and Conductor once each and Surfer twice. Zero American agencies were named in the blind pass. Thirty eight of the 64 readings named no supplier at all.

That is the useful context for a buyer holding five agency names. Agency names reach a buyer through a question that supplies them, and the unprompted American answer names software or nothing. So the names in front of you are a longlist somebody handed you, and the verification work is entirely yours.

The engines also split. ChatGPT ran a live search on all 32 of its American readings and Gemini on 16 of its 32, so the two surfaces are reading the current web at different rates and a claim made today reaches them at different speeds. We report them separately for that reason and never average them.

How to run the comparison in a week

• Decide your category from the table first. Comparing a PR plus digital firm against a business to business demand specialist produces a tie that means nothing.
• Send the five questions in one email to all five, on the same day, with a reply date.
• Build your own price row from each firm's own published page, dated, with quote only left empty where that is the honest answer.
• Score the written answers before any call, and only call the two that scored highest.
• Ask both finalists what they would refuse to do for you. A real practice has refusals it can name.

If the scope turns out to be one discipline rather than five, a named comparison inside that discipline is a better instrument, and the ecommerce version of it is in the assistants named and the one published monthly price. If the AI visibility question in the fifth bullet is the one you care about most, the distinction between the tool categories is set out in which of the three records what.

Questions buyers ask about these five American firms

Which of the five is best for a mid sized American company?

The one whose category matches your scope. For several disciplines at once, the full service firms. For media heavy performance work, the performance firm. For business to business software, the specialist. A single ranking across all five would be comparing different products.

Are the agency rankings on directories reliable for these names?

Directory profiles are leads rather than verdicts. In our American pass clutch.co was cited in 11 readings and goodfirms.co in 6, so the engines do read them, and that is a reason to check your own profile as well as theirs.

Should I ask for a case study or a reference?

A reference, on the phone, in your sector. Ask what went wrong and how it was handled. A case study is written by the supplier and a reference is not.

Why does the comparison avoid quoting their prices?

Because a price copied today is a claim about somebody else's business that goes stale quickly, and an undated price in a comparison is exactly the kind of detail an engine has reason to discount. Read each figure on the firm's own page on the day you decide and date your own row.

Do any of the five sell AI visibility measurement?

Ask each one the fifth question above and record the answer. In the American readings we took on 22 October 2026 the products the engines named for that job were software rather than agencies, which tells you where the category currently sits.

The measurement none of the five will hand you by default

Whichever of these five you hire, the report you receive will cover rankings, traffic and media performance. It will not tell you whether ChatGPT or Gemini named you when an American buyer asked for a supplier in your category, and on our reading of 22 October 2026 the honest answer for most American companies is that they did not, because 38 of 64 readings named nobody and the other 26 named software. AI Knows Us reads ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews, records which of them named you and what each answer cited, and hands over the raw answer text so the number can be audited rather than taken on trust. Put it in the brief for whichever firm you pick. In the United States that is the Growth plan at 299 dollars a month, the cheapest tier reading all five engines, read on aiknowsus.com/pricing on 22 October 2026.