Frequently asked.

Four ways to put the intelligence to work, all powered by AskDPG. Tell us whether you are evaluating this market or competing in it. Choosing? You get a neutral analyst view. Competing? You get a clear-eyed read of where you lead and where you are exposed.

• Vendor Briefing Pack. An analyst-grade write-up of any vendor on demand: what they do, where they are strong, where they are thin, how they have moved. Go deeper for funding history, twelve months of material events and the full capability picture.
• Comparison. Up to five vendors side by side, with a written verdict rather than a feature grid. The same 96 capabilities, the same scored dimensions, plus the full research record behind each one. It tells you where they genuinely differ and where the difference is only marketing.
• DPG Labs Advisor. Describe the problem in plain English, get a shortlist with the reasoning shown: why these vendors, what they share, what you would be trading off. Ask follow-ups and argue with it.
• Natural language search. Find vendors by describing what you need, not by guessing which category someone filed them under. Categories are a vendor’s own choice. Capabilities are ours.

We use many frontier AI models, from across the leading labs, each doing the job it does best: research, extraction, assessment, writing.

Ask a chatbot about this market and it has to guess the question before it can answer. It decides for itself what matters about a vendor, then decides differently next time. It reaches for whatever it can find, which is mostly the page the vendor wrote about itself. Every vendor here claims every capability, so the claim comes straight back to you, fluently. It surfaces the loudest names, which measures marketing rather than the market, and it cannot show you what it never found.

It also invents its vocabulary on the spot. It will describe the same capability three different ways across three vendors, so the difference you think you are reading may be nothing more than the words that came out that day. You cannot lay two answers like that side by side, which was the only thing you wanted to do.

We settled our questions before we asked any model anything. Ninety-six capabilities, out of fifteen years of buying, deploying and occasionally getting this wrong inside some of the largest companies in the world. They are the shared language we could never find, written down. We ask every vendor the same questions, in the same words, every time.

Knowing what a vendor does is only half of it. You cannot answer whether they are any good at it one vendor at a time. Enterprise maturity, technology signals, market signals, analyst recognition: each one places a vendor against every other vendor in the index, on a rubric we publish. A chatbot has no population to measure against. It can only tell you a vendor sounds impressive.

We tie every finding to its source, date it, and cross-check it across independent domains. Where we have not looked, or could not, we score nothing at all: we never record an unchecked vendor as a zero. We keep what we find, including 42,000 archived versions of vendor pages, so we can show you what changed rather than what a model recalls.

A chatbot gives you half an answer, at a point in time, with no way to check it and nothing to hold it against. What we have built is the thing you were trying to do with the question.

We build everything we publish from public information. No vendor supplies it. No data broker sells it to us.

Across the index we have mapped 2.6 million published pages and hold more than 300,000 sourced claims. We read four kinds of public evidence, for every vendor.

The vendor’s own words. We map the full published footprint of every vendor site and read deeply from it. We also hold 42,000 archived versions of pages that have changed since we started watching, so what a vendor said last quarter stays as visible as what it says today.

The vendor’s public activity. Over 108,000 posts, 11,000 open-role descriptions and 11,000 videos from the last twelve months. What a company hires for tells you more than what it says it does.

Deep independent research. We run a structured research pass on every vendor across news, filings, funding records, customer references, analyst commentary and regulatory sources. That gives us over 300,000 claims from more than 200,000 sources across 29,000 domains. We hold each one with its source and the date the source published it.

The record of change. Over 24,000 material events: funding, acquisitions, launches, leadership moves, customer wins. We date and attribute every one.

Every claim keeps its source. We cross-reference across independent domains rather than trusting any single one, and where sources disagree, we keep the disagreement rather than smoothing it over. No vendor describes itself into the index, which is the only reason a comparison between vendors holds at all.

Figures as at September 2026, refreshed monthly.

No — both as a matter of policy and because the architecture prevents it. There is no field to type a score into. Every refresh recomputes every score from the underlying evidence, so a score is not a stored opinion; it is the current reading of what anyone can observe.

Vendors do not supply the evidence behind their own scores. The same method reads every vendor, from the same kind of evidence, so the comparison holds — which only works if nobody is describing themselves.

Vendors can subscribe. What they buy is the ability to read the same analysis everyone else reads, framed for their own position. No subscription, at any price, changes a number.

There is one way a score moves, open to every vendor equally and at no cost: put the evidence into the open. When it is there, the next refresh finds it.

Dr. Elouise Epstein, an independent procurement thought leader, runs her own personal endorsement of the most exciting early-stage vendors in procurement technology on the platform. She gives it independently, without payment, and on her own judgement alone. Read more about the endorsement.