The engine
A loop that remembers what it did.
Account data, your context, monitoring, and recommendations, on one record. Every finding points back to what changed. Every action waits for your approval.
How it runs
From account data to a decision.
Step through the six stages. Collection, analysis, and delivery each run on their own schedule; the record is what ties them together.
Bring the account data together.
Metrics and settings from every connected platform, collected on a schedule and written to one record. Each platform exposes a different depth; the record shows you what it has.
Read the account through your context.
Goals, baselines, and the business facts you save in Strategy give the numbers a frame. A $40 CPA is good or bad depending on what you told it.
Patterns and conditions worth investigating.
Trends run every weekday. Alerts fire on thresholds you set, or on movement against the account's own forecast.
Put the change next to the movement.
A performance shift and the account changes around it, side by side, so you can see what lines up and judge what it means.
Propose the next step.
A specific change with its target and expected effect, weighed against your objectives. You can read it without doing anything about it.
Approve, then read the outcome.
Nothing runs until you approve it. The approved change lands in History with its result, and the loop starts again from a better record.
Everything in the platform
Three pillars. Every feature.
Memory holds the record. Signal reads it. Action moves on it. Each card carries its current availability.
Memory
History and Strategy
History
liveTrace the changes behind the review.
Settings changes and platform events from every connected account in one timeline, with before-and-after values. Compare a snapshot, follow a thread, find the edit that lines up with the move. Coverage follows what each platform exposes.
Strategy
betaKeep your goals in the picture.
Business facts, targets, baselines, and corrections, saved as context the rest of the product reads against. Tell it once that a $40 CPA is fine on brand and unacceptable on prospecting. It remembers.
Signal
Alerts, Trends, and Reports
Alerts
liveMonitor the conditions that matter.
Set a threshold, or let the account's own forecast define normal. Each alert shows the affected account and the triggered period, and delivers to as many destinations as you want, on the frequency you choose.
Trends
livePatterns worth a closer look.
Opportunities, warnings, and insights drawn from the connected accounts every weekday, each one tied to the metrics and history that support it.
Reports
liveFrom a brief to a report you can send.
Write a brief, get a draft, edit it, and review its revisions. The Ad Spend suggests reports out of the box; you pick the scope, the recipients, and the schedule.
Action
Optimizer, Intent, and Attribution
Optimizer
private betaReview every account-changing action.
A recommendation with its target, the proposed change, and the expected effect. Approve it and it runs with the permissions you have granted; dismiss it and nothing moves.
Intent
roadmapA direction for audience decisions.
Audience targeting and response patterns, read against the same record as everything else. Coming after Optimizer.
Attribution
closed betaConnect the analysis to closed revenue.
HubSpot and Salesforce closed-won revenue alongside the ad spend that produced it, kept distinct from platform-reported conversions.
Why a record
The reasoning is the product.
A number tells you what happened. A record tells you why someone decided what to do about it. When the reasoning is written down, the account stops depending on whoever happened to be watching the dashboard that day.
See how memory worksIllustrative change record
- Recommendation
- Shift 18% of budget to Prospecting
- Reason
- Retargeting CPA up 31% over 9 days
- Expected
- Blended CPA down 8–12%
- Observed, 14 days later
- CPA down 11% over 14 days
Questions
Answers, not gestures.
- How is this different from automated rules?
- A rule evaluates a condition and fires. The Ad Spend remembers what changed, reads the movement against your own history and goals, explains the likely cause, and proposes what to do. A rule cannot tell you why.
- Does it change my campaigns on its own?
- No. Every action that touches an ad account is proposed first and runs only after you approve it, with the permissions you have granted.
- How confident is an explanation?
- It shows you the change and the movement together, and how far the movement sits from baseline. That is a strong lead, not a verdict. You make the call with the evidence in front of you.
- What history does the product start with?
- The initial sync pulls the recent history the platform makes available, then the record builds from there. How much of it you can browse depends on your plan; the pricing page lists what each includes.
- Which ad platforms can I connect?
- Google Ads, Meta, LinkedIn, TikTok, Reddit, Microsoft Ads and Pinterest Ads. OpenAI Ads aggregate campaign reporting is available to advertisers with API access. Each platform exposes a different depth, so what History captures and what Optimizer can change vary by platform.
- Can I use a compatible MCP client?
- Yes. An authenticated, hosted MCP endpoint lets compatible clients query the record; plan availability is on the pricing page.
See it on your account
Watch the loop run on your own spend.
Connect an account and the engine starts building its record from day one.