About Tanssor Research
Every Tanssor Note goes through a two-tier review pipeline: multi-model AI cross-check against the underlying tender documentation, followed by mandatory analyst sign-off on source verification before delivery. Nothing reaches a customer without both tiers — that accountability is what we think should earn your trust.
Why this exists
Public procurement records — TED, national contract registers, trade statistics — are genuinely public and genuinely unusable at bid-team speed. A single tender can require cross-referencing a dozen award notices, a national company registry, and a framework's own contract history just to answer "is this lane actually open to us." That reading is mechanical and repetitive in a way software should be doing, and the synthesis at the end still needs a human decision about what actually matters for one specific bid. Tanssor exists to do the mechanical part at scale and hand back something a person can act on.
How the work actually gets done
An automated pipeline reads official award notices and structures them into evidence — entities, dates, values, sourced facts — before any narrative gets written. A first AI pass drafts the Note from that structured evidence, citing back to it rather than writing from memory. At least two independent AI models, from different providers, then try to find what's wrong with the draft — unsupported claims, weak inferences, missing context — neither of them the model that wrote it, before it's shown to anyone. Every Note that reaches that point goes through sign-off before it's sent. Nothing ships without that check. Full detail on how the evidence tiers work is on the Methodology page, and the honest version of what AI does and doesn't do here is on Transparency.
Every Note follows the same documented process:
- TED award-notice extraction and normalization
- Buyer and winner identity resolution against national registries
- Historical procurement pattern analysis for that buyer
- Evidence grading — OBSERVED / DERIVED / INFERENCE, each with a confidence level
- Independent AI review, then mandatory analyst sign-off, before delivery
- A public correction, in place, if a factual error is found afterward
What this isn't
It isn't a platform, and it doesn't claim decades of sector experience it doesn't have. It's a working research pipeline that reads more official procurement records than manual review could keep up with, and it gives an honest account of what those records do and don't support.