- The talent scout becomes a relentless, source-aware headhunter. Instead of stopping early, the scout keeps digging until it has gathered real depth — a floor of at least twenty strong candidates — searching each talent source through its own dedicated door and ranking who comes back across all of them. It now reads far more about each person before deciding — their bio, their prices, how long they've been working, their credentials and the real text of their offerings — and carries your own stated preferences into the final order. Its one clarifying question is now asked mid-hunt and picked up right where it left off.
- Budgets can be a range, and the low end becomes a gentle floor. When you give a budget as a range, the bottom number is remembered and used to rank people who sit below it a little lower — never dropping them — and an upside-down range (bottom above top) is politely refused instead of quietly breaking the search.
- The human "why they fit" pitch gets sharper and reaches everyone. PORTIA's short pitch for each match is now two tight, fit-focused sentences with an optional real review quote, falls back gracefully when a profile is thin, and shows even when you're just trying the app without signing in — threaded through the whole search so it never gets lost.
- The assistants get more careful in conversation. Outreach leads with a talent's interest before money and keeps an estimate separate from a real limit, the brief now shows your budget and timeline the moment they're settled, and a live offer is no longer dropped when the negotiator escalates in the same turn — with a safety net that falls back to a wider search if a run collects nothing.
- The brief you download now looks exactly like the app. Brief exports are rendered straight from the real app components, so the PDF is pixel-identical to what you see on screen, with the print styling fetched fresh from the web rather than frozen into the backend.
- The talent cards get a design-exact redesign. The match cards were rebuilt one-for-one to the new design — a full-width call-to-action footer, cleaner spacing, and no awkward gaps on short titles.
- The app comes back cleanly after a refresh or sign-out. A broad hardening pass means the buyer app restores its state reliably, shows a proper loading state instead of a false "no matches" while results arrive, keeps the chat from bouncing, caps the thinking strip to a few tidy rows, and smooths a long list of small copy and layout rough edges.
- A new admin tool can replay the scout. The team can now re-run the talent search on past projects and compare the fresh finalists against the old ones, run quietly in the background so nothing times out.
- The foundations of a data & analytics platform land. Behind the scenes, Mira gets a private analytics warehouse and a self-serve dashboard, with strict guards so sensitive columns and raw sign-in identifiers never leak, and key events timed faithfully — plus scale-readiness groundwork (sturdier AI retries, longer jobs, more database headroom, autoscaling alarms, a first production environment, and stronger account-deletion cleanup).
The biggest day of the week. The talent scout turns into a relentless, source-aware headhunter — one door per source, a floor of twenty-plus, and deep reading of each person before it ranks them — budgets can now be a range with a soft floor, PORTIA's pitch gets tighter and reaches guests, the brief exports pixel-identical to the app, and the crew lands a design-exact card redesign, a broad refresh-and-restore hardening pass, an admin scout-replay tool, and the first foundations of a data & analytics platform.