A practitioner field guide, openly published by the team that builds Jarvis AI

Jarvis AI is our product. ASCENDING Inc., the company that owns this site and pays the salaries of the people who write here, also builds Jarvis. Every comparison table that includes Jarvis is flagged. Every page that discusses Jarvis opens with a disclosure banner. The /jarvis page is a product page, not editorial.

The standard we apply is the Federal Trade Commission's endorsement guide at 16 CFR Part 255 [1]. The rule requires a clear, conspicuous disclosure of any material connection between an endorser and an advertised product, in the same medium as the endorsement. Our connection is simple. We make Jarvis. We say so at the top of every relevant page. We repeat it in the footer. We flag it in every comparison row. The FTC's 2023 update [2] tightened the "clear and conspicuous" language and expanded liability to intermediaries. We read it. We apply it.

We chose to write openly rather than pretend to be independent. Three reasons. It is honest. It is better for trust, because the readers we want are smart enough to search our LinkedIn profiles in ten seconds. And it lets us be useful in a way anonymous editorial cannot. We can tell you what we've seen break in production, with the specifics, because we were the ones fixing it.

ASCENDING Inc. is headquartered at 2751 Prosperity Avenue, Suite 240, Fairfax, VA 22031 [3]. We are an AWS Advanced Consulting Partner listed in the AWS Partner Solutions Finder [4]. Since 2023 we have focused increasingly on productised AI, building Jarvis AI (see the Jarvis product page and the AWS Marketplace listing [5]), working with public-sector and regulated-industry customers, and publishing open research.

The Explore Agentic masthead is a rotating cast of ASCENDING staff and a small number of external contributors; each byline links to a LinkedIn profile readers can verify in ten seconds. Our audience is CIOs, platform engineers, AI governance leads, and the people inside procurement who have to make the pick. We publish for them because we are trying to be chosen by them.

Every piece is anchored to public sources we can link to. Vendor documentation. Standards-body publications (ISO, NIST, Linux Foundation). Analyst commentary (Gartner, The Futurum Group, McKinsey QuantumBlack, Forrester). Peer-reviewed academic papers. Comparisons cite every column of every table. Pricing pages are dated. When our reading is directional rather than authoritative, we say so on the page.

The method borrows from three public standards: the International Fact-Checking Network's five-commitment code, covering nonpartisanship, non-advocacy, transparency of sources, transparency of funding, and an open corrections policy [6]; the Reuters Trust Principles on integrity and freedom from bias [7]; and the FTC's clear-and-conspicuous disclosure standard at 16 CFR Part 255 [1]. We are not an IFCN signatory (the IFCN accredits fact-checking organisations, which we are not). We reference the code because its source-transparency and corrections commitments are the right baseline for editorial that intersects with a commercial product.

Where our experience as Jarvis builders informs a claim (and it often does, because we see production issues most commentators don't) we flag it inline. We use AI tools in drafting. Every piece is reviewed by a named human editor before publication.

When Jarvis appears in a comparison table, the disclosure banner runs above the fold. We list categories where Jarvis loses. We publish our weaknesses on the same page as our strengths. Anyone evaluating us can already figure them out by reading the docs. A comparison table that shows our product winning every category is not a comparison. It is marketing we don't respect.

If we get something wrong, we publish a correction. In place, logged at the bottom of the piece. If the error is big enough to change the conclusion, we date-stamp the correction and update the citation trail. Corrections are not deletions. Follows the IFCN commitment to an "open and honest corrections policy" [6]. Signatories publish their corrections policy and follow it scrupulously. So do we.

To flag a factual error, email editors@exploreagentic.ai. Include the URL, the specific claim, and a source that contradicts it. We acknowledge within three business days and publish the correction with a timestamp when it ships. Disagreements about framing are different from factual errors. We welcome both. Only the second produces a dated correction note.

Story tips, corrections, and tips about our product: editors@exploreagentic.ai. If you are evaluating Jarvis or a competitor and want to talk under NDA, we'll make time. If you are a writer or researcher who wants to contribute a guest piece, we publish bylines from outside the company. Write in with a pitch.