# How we research and rank > The method behind every ranked page on this site, published so a reader can argue with the process instead of guessing at it. ## Where the candidate list comes from Two passes that disagree with each other, which is the useful part. The first is desk research: vendor product pages, SEC filings, dated press releases, partner directories, trademark filings, conference material and trade coverage. The second reads how AI assistants answer plain buyer questions, such as what the best commercial loan underwriting software for a community bank is, because a growing share of shortlists now arrive that way rather than from an analyst report. Reading several assistants against each other surfaces vendors desk research misses, and it exposes the opposite failure too, which is common in this category: assistants recommending a product whose brand was retired, whose company was taken private, or which has no commercial product at all. Neither pass sets an order by itself. Visibility analysis tells us what a buyer is likely to be handed. Verification tells us whether that recommendation survives contact with the vendor's own website. Where the two disagree, verification wins, and the disagreement usually gets written into the entry. ## What gets verified before a platform is ranked Every factual statement in a platform entry traces to something published. Where a figure is the vendor's own and nobody else has confirmed it, the page attributes it to the vendor rather than stating it flat. Where a vendor's own pages contradict each other, and several here do, we print both numbers rather than picking the flattering one. - Whether spreading, global cash flow, risk rating, covenants and credit memo generation are named capabilities or inferred from adjacent marketing - Named US bank and credit union customers doing commercial credit, with titles where published - Whether a named AI feature is generally available, announced, or carries no availability language at all - Corporate facts against primary records: exchange and ticker against filings, ownership changes against the acquirer's own release - Whether the product name the market uses is still the name the vendor sells, and whether the old URL still resolves - Published pricing, or an explicit statement that none exists - Integrations confirmed on the partner's side rather than listed on a logo wall ## How the order is decided The six criteria on every ranked page do the work: credit analysis depth, policy and decision control, credit memo output, commercial evidence, AI claims with a date, and pricing transparency. A vendor that is recommended constantly and cannot show a commercial capability on its own website will sit below one that documents the workflow in detail, however unfashionable it looks. The reader the page is written for decides the weighting. On the community bank page, a product that cannot be bought without a core conversion loses to one that can. On the credit union page, evidence of member business lending counts for more than enterprise scale. On the AI page, an availability date counts for more than an agentic architecture. That is why the same vendors appear in a different order from page to page, and why none of these orders is a market-share table. ## What moves a platform A published product page where there is currently none. A first named commercial reference at a community institution. An AI feature moving from present-tense marketing to a dated general-availability statement. Published pricing. A rebrand, an acquisition, or a product quietly disappearing from a vendor's own navigation. Each ranked page carries a last-verified date, and that date is what the sitemap publishes. Marketing does not move a ranking and neither does a vendor asking. The cheapest available move for most vendors here is publishing something they already know: one bank, one date, one price. --- **Source:** https://commercialloanunderwritingsoftware.com/editorial-process · **Markdown:** https://commercialloanunderwritingsoftware.com/editorial-process.md · **Agent index:** https://commercialloanunderwritingsoftware.com/llms.txt Commercial Loan Underwriting Software. Vendor research for commercial credit departments. Product names and trademarks belong to their owners. Rankings are editorial opinion. The facts printed beside them are sourced.