The latest features, improvements, and fixes shipped to Sentrum.
On a phone the session rail filled the whole screen, and the thread and the message box were pushed out of view with no scroll that could reach them back. The rail is now a drawer: the conversation and composer own the screen, a menu button opens the session list, and tapping outside it or picking a session closes it again. Pin, rename and archive no longer need a hover to appear.
On value-optimised campaigns ROAS is a return, not a cost — but eleven verdicts across campaign detail, Daily Pulse, demographics, placements and creative detail still read it as a cost. A 5.00x ad set was told it was above campaign average and offered for pausing; a 1.00x one was offered as a scaling opportunity. Winner, loser, trend and fatigue are now judged in the right direction on every one of those surfaces, and each prints the objective’s own noun and unit.
Two different campaigns could show word-for-word the same next step, because the advice was a fixed sentence rather than something written from the diagnosis. Across Daily Pulse, the Diagnostics tab, campaign cards and the audience-overlap check, every action line is now composed from that campaign’s own diagnosed driver and its real numbers. Where nothing crosses a significance threshold the card ships the diagnosis with no action row at all — saying less beats filler.
The Ad Library import now survives the links buyers actually paste — share redirects, quoted or truncated URLs, a missing https — and every rejection says which part was wrong instead of one generic failure. An empty answer states whether nothing was found or nothing was searched, always with the public Ad Library link, and a partial import reports how many ads landed rather than quietly showing fewer. Each imported reference is read against your own measured creatives, so you see how that hook has really performed in this account before you borrow it.
A creative that clears the winner threshold now carries a Remix this winner chip on its card in Creative Intelligence. It opens Creative Studio already seeded with that ad’s copy and format, and a Why it wins disclosure shows the numbers that earned the label — its share of spend and how its efficiency compares to the campaign average — so you are remixing something the data actually called a winner. A one-time note on the creatives page points the chip out, and stays dismissed once you close it.
Twenty-nine responsive defects across the dashboard, campaign detail, Daily Pulse and Creative Intelligence. The date picker’s second calendar and its Update button sat off-screen, so a date range could not be set on a phone at all. The Reporting, Creative and Profitability menus opened past the right edge and could not be scrolled to. The dashboard tab bar slid behind the top bar and vanished on scroll. And explanations that lived only in hover tooltips — why an ad wins, why a creative is fatigued — are now reachable with a tap.
Creative Studio was folding your headline and call to action into the video prompt and asking the model to render that copy legibly in frame. Video models cannot spell: takes came back with garbled burned-in type, and once with an invented disclaimer nobody wrote. The video lanes now ask for clean space and no text in the frame, your copy goes over the finished clip instead, and the Creative Direction panel says which of the two is happening on the model you picked.
On value-optimised campaigns, higher is better — but the deck builders were reading ROAS as if it were a cost, so the best ad set could be labelled the loser. One wrap recommended shifting budget away from a 4.65x ad set toward the weakest one. Winner and loser are now decided per metric, in the right direction, across the cadence deck and report, the client PDFs, the scheduled report email, and the CSV export.
The reach report drew a frequency distribution chart that was generated from a statistical model, not from your data. It is gone. Reach now states the basis it is actually measured on — summed daily reach — instead of implying a period-unique number Meta never returned.
A sweep of every interactive control in the product found several that looked live and were not — including the Slack webhook "Test" button, which could never have worked from the browser and now posts through the server. Failed saves and failed exports now show you an error instead of failing silently.
The onboarding screens no longer promise a finish time we cannot keep, or a depth of history that was never being fetched. The first sync is not reported as complete until the data behind it has actually landed. And the progress bar now reflects the sync phases as they finish, instead of sitting at zero for the whole run.
Our connection and legal pages described the Meta permissions we request more narrowly than what we actually ask for, and named the wrong infrastructure for where your data is stored. Both are corrected: the connect screen, privacy policy, terms, and security page now state the real approved scope and the real database provider.
A pass over the paths between screens fixed eight of them. The "syncing your data" screen after connecting now resolves instead of hanging. A report run no longer spins forever after one failed status check — it tells you it lost contact, and offers a retry for runs you started in that session. Switching tabs stops quietly resetting your platform and date filters. And on agency workspaces, drill-down links from the strategy view now carry the workspace with them, so you land in the client you were reading, not a different one.
Your account’s edit history — budget changes, status flips, new ad sets — used to populate only when somebody opened Campaign Deep Dive. It is now pulled automatically on every sync, so the Change Log is already there when you go looking for what happened before a metric moved.
Demographic breakdowns stopped at the first page of results from Meta, so accounts with many age, gender, placement, or platform combinations were quietly reporting partial totals. All breakdown pages are now read through to the end.
Subscription checkout now accepts promotion codes, so a discount or founding-member code can be entered on the Stripe page instead of being applied by hand afterwards.
Connect Meta, run your first diagnostics, and explore the dashboard. Current plan and billing terms are on the pricing page.
Use Sentrum directly inside Claude, Cursor, ChatGPT, and other AI assistants. Ask your ad account questions and get diagnoses without leaving the tools you already work in.
Pro, Developer, and Agency workspaces now refresh their Meta data every hour, so the accounts you care about most stay current without manual syncing. Starter refreshes daily.
For Advantage+ Shopping campaigns, Sentrum now breaks out new vs. returning customers — share, cost per acquisition, and ROAS for each — so you can see where real growth is coming from.
A plan for media buyers who run their workflow inside an AI agent — high-volume AI usage, custom tooling, webhook notifications, and priority support, on top of everything in Pro.
Connecting a new account is dramatically faster — your first diagnosis is ready in about a minute and a half instead of several minutes.
Export polished .xlsx reports from Campaign Deep Dive and the dashboard — a branded cover sheet, KPI tiles with period-over-period change, daily trends, and color-coded breakdowns. Ready to hand to a client.
New accounts now land in the dashboard with an AI diagnosis already generated for your highest-spend campaign — so you see Sentrum working before you click anything.
A guided walkthrough that takes your account apart in order: spend health, results, efficiency, supporting metrics, mix shifts, recent changes, and recommended actions — a complete read in seconds.
The Attribution view now opens with a plain-English diagnosis — over-credit risk, retargeting inflation, and diminishing-returns campaigns — instead of just charts you have to interpret yourself.
Diagnoses now speak the current Meta language — Andromeda, Advantage+ Shopping, ODAX objectives, CAPI and event quality — and account for how Advantage+ campaigns behave differently from manual ones.
Reach and frequency now reflect true unique reach per campaign, so the unique-user numbers mean exactly what they say.
Your Settings page now shows your current AI usage against your allowance.
A complete visual refresh of the marketing site and key pages — cleaner typography, a dark-first palette, and sharper presentation. Same product, better first impression.
Our diagnostic engine now reasons in stages — detect what changed, find the root cause, recommend the fix, and explain it — producing materially deeper analysis across Daily Pulse, Campaign Deep Dive, Blueprint, and Attribution.
Rebuilt creative analysis with hook-rate scoring, hold-rate tracking, fatigue curves, and concept clustering — so you can see which creative elements actually drive performance.
Set product margins so Sentrum can frame performance against profit, not just revenue.
A unified feed of AI-detected insights, anomalies, and recommendations across your whole account — the most important signals surface automatically, no digging through tabs.
Invite teammates with role-based permissions — admins, editors, and viewers — to control who can see data, change settings, and manage integrations.
Send anomaly and performance alerts to Slack or your own endpoints, with per-alert routing, severity thresholds, and schedules.
Press ⌘K to instantly search campaigns, jump between pages, or trigger exports — built for a keyboard-first workflow.
Automatic detection of CPM spikes, CPA drift, CTR drops, and spend anomalies — so small issues get flagged before they get expensive.
See which campaigns are over- or under-spending, with AI-suggested reallocations and daily pacing tracking.
Export any table — campaigns, ad sets, ads, or creatives — as a clean, properly formatted CSV.
Generate a presentation-ready post-campaign recap in one click — top-line performance vs. benchmarks, audience learnings, creative deep dives, and next steps, at the campaign, ad set, or ad level.
Ask plain-English questions about your account — "why did CPA spike last week?" — and get answers grounded in your real data, referencing auction dynamics, creative fatigue, placement shifts, and learning-phase effects.
The core of Sentrum: a daily automated sweep for CPM spikes, CPA drift, CTR drops, creative fatigue, and budget inefficiencies — with severity, root-cause context, and recommended actions, so you catch them before they cost you.
Get notified when we ship new features and updates.