Blog · Page 4
Field notes.
Page 4 of 31. Browse the archive of RFP workflows, grounded-AI architecture, and proposal operations notes.
New models, quarterly eval: Sonnet 4.6, GPT-5.2, Gemini 3.1 Pro
An internal eval across three current-generation models for our specific workloads — drafting, claim verification, extraction. What moved, where we switched defaults, and why one workload still sits on a year-old model.
SME collaboration, the six-month update
Revisiting the September 2025 SME-collaboration series. Which patterns persisted in real teams, which ones had to be rewritten, and the one I was wrong about.
Compliance extraction, revisited
The grammar we moved to for requirements extraction, why we stopped treating 'shall' as a single class, and the evaluation showing a 38% drop in false-positive requirements.
The Q2 FY kickoff Monday-morning triage
Eight federal RFPs dropped into the queue over the first weekend of Q2 FY. The 25-minute drill that sorts them into respond-today, respond-this-week, and wait-for-the-amendment.
Vendor risk management, patterns we see on the procurement side
A cross-cut of roughly 200 DDQs from the last six months — the fields that repeat, the fields that vary, and what the repetition tells us about how vendor risk teams actually operate.
In preview: per-answer quality score with breakdown
A four-dimensional quality score — clarity, grounding, compliance, brevity — is rolling out in preview on drafted answers. How the score is computed, where it lives in the UI, and what it changes about the review pass.
Pricing at one year: what we're changing in Q2
The pricing experiment continues. Two new tier lines, one that's retiring, and the reasoning behind each move after a year of published pricing and real customer usage patterns.
March reliability incidents, documented
Two incidents on the platform this month — one degradation, one full outage. What triggered each, how long they ran, what the user impact was, and the specific changes we made after.
Mapping every response paragraph to the scoring rubric
The discipline that turns a 60-page response into an evaluator's checklist. Why every paragraph needs a rubric citation, and how to make the mapping visible without cluttering the document.
The SLA on draft generation: 45 seconds, 95th percentile
The operational target we hold draft generation to, why it's 45 seconds and not 30 or 90, and the specific things we do to hold the number under peak federal-FY-Q2 load.
Announcing versus educating, a product-comms pivot
Why we moved from launch-day announcements to continuous smaller posts about how the product works. The shape of the shift, what we gave up, and what we gained as a team.
A 2026 map of enterprise procurement platforms
Coupa, Ariba, Workday, Ivalua, GEP, and the newer entrants. Where AI shows up in each platform's RFP workflow, what's real, what's marketing, and what it means for vendors who have to respond through these systems.
Prefer to see the product?
Take the 5-minute tour, or start a trial workspace and see PursuitAgent draft answers with citations.