Frequently asked questions

Prime Signal is an AI-powered autonomous ad management platform. It watches your connected ad accounts around the clock. Meta (Facebook and Instagram), Google Ads and GoHighLevel are live today. Microsoft Ads, TikTok Ads, LinkedIn Ads, X Ads, Snapchat Ads and Pinterest Ads connect today with your own platform credentials, with one click sign in to follow. When an ad gets disabled, a budget overruns, or performance drops, Prime Signal diagnoses the issue, generates a fix, and acts on it. You set the autonomy level per client, from alerts only to fully autonomous, and you can change it in a second.

Meta Ads (Facebook and Instagram) and Google Ads are the live ad platforms today, and GoHighLevel is live for CRM and automation data. Microsoft Ads, TikTok Ads, LinkedIn Ads, X Ads, Snapchat Ads and Pinterest Ads are built but not yet released, so we do not count them as supported and they are not listed as options anywhere on this site. You connect by OAuth from the dashboard. No API keys, no developer, no code. Prime Signal syncs your account history, maps every campaign and ad set, and starts monitoring on the next scheduled pass, which is at most 15 minutes away.

Connecting is a single OAuth handshake from the dashboard. Set your operating mode (Monitor, Advise, Guard, or Autopilot) and Prime Signal takes over on the next scheduled pass. Baselines are built per day of week from up to 56 days of account history, and each day of week needs at least two samples, so the day-of-week baseline becomes usable after about two weeks and sharpens from there.

Media buyers, marketing agencies, ecommerce brands, and anyone running paid advertising who needs their campaigns watched around the clock. Solo operators use it to cover work they would otherwise hire a media buyer for. Agencies use the Amplify Agency plan, which adds client accounts at $29 per month each. Larger teams use it for unified reporting and governance across every connected account.

Four things, and you set all of them. First, per-client guardrails: maximum daily spend change, bid adjustment limits, and the threshold above which a change has to be approved. Second, every action is classified Low, Medium, High or Critical risk before it runs, and the top tiers wait for a person in every mode except Autopilot. Third, a budget increase consults the account’s delivery capacity first, and an account already at capacity gets a review recommendation instead of more spend, in every operating mode. Fourth, every change is recorded with its reasoning and the state of the account before and after, so a wrong one is visible and reversible rather than mysterious.

Prime Signal connects through Meta’s and Google’s own OAuth, the same door their approved partner tools use. Nothing is scraped, no password is shared, and no action is taken outside the platform’s own API. When a platform rejects an ad, Policy Guard classifies the violation type, writes a compliant version, resubmits it, and records whether the platform accepted it, so the next fix is based on what that platform actually approved rather than on a guess. Policy Guard can also pre-screen an ad before it is ever submitted. What we will not do is put a number on ban risk, because we have no dataset that would make one honest.

No, and the controls are structural rather than promised. Autonomy is a setting per client, changeable in a second. Approval is a step inside the workflow rather than a notification, so a change waits exactly where you told it to wait and nothing slides through while you are deciding. Disconnecting a platform revokes the OAuth token immediately. Every action carries its reasoning, the data behind it, and the before and after state of the account, in an audit trail you can search and export.

The product is built the other way around: the work comes to you. Everything that needs a decision lands in one ranked queue with the reason attached, and there is no second inbox anywhere in the product by design. Weekly and monthly reports are generated per account, ready to send. Notifications reach you in the app and by email, with SMS and voice escalation on the plans that include them. You can go a week without opening the dashboard and still know what happened.

Your campaign data is never used to train an AI model. Model calls run through one governed path on zero-retention APIs, and every call is recorded. Each organization’s data is isolated, and cross-organization access is architecturally impossible rather than policed. The reasoning layer never sees a raw platform token either: it works with opaque references, and the real credential is injected only at the platform API boundary inside a sealed context that wipes its memory afterwards. The learning stays inside your own organization.

Intent data identifies people who are actively researching or showing buying signals for products and services like yours. Instead of targeting broad demographics and hoping, you reach people who are already in-market. The result is higher conversion rates, lower cost per acquisition, and less wasted ad spend.

Deterministic resolution means every step in the identity chain is a verified, 1:1 match. A browser event ties to a specific hashed email, which ties to a verified offline record with name, address, and contact info. You can trace the full path from signal to identity. Probabilistic resolution guesses who someone is based on statistical patterns. An IP address maps to a zip code, a model picks the most likely person from that area, and confidence erodes at every step. There is no verified link, no way to trace the match, and no way to close the feedback loop. Our infrastructure uses deterministic resolution. That is why we can verify whether a signal actually led to a conversion and continuously improve accuracy, something probabilistic providers structurally cannot do.

Three layers of verification run continuously. First, the consumer database is refreshed through national change of address records every 30 days. Second, emails are verified daily through multi-layer validation (online, offline skip-trace, and public data sources), which the provider reports holds verification rates in the 70 to 90 percent range. Third, a closed feedback loop validates every intent signal against actual downstream conversions. If a signal does not lead to real buying behavior, the model downgrades it. Every figure in this answer is published by the third-party identity data provider that supplies the graph. Prime Signal does not measure them independently and does not publish a competitor comparison.

Most intent platforms follow a one-way pipeline: capture a signal, score it, sell it. There is no verification of whether the score was right. Our model closes that loop. Intent signals are validated against real conversions, engagement, and buying events. After enough confirmed outcomes, the model begins fingerprint matching against proven converters. Every classification strengthens the graph. Every outcome improves precision. The identity data provider reports 3x higher conversion lift for verified intent against generic intent data. That figure is theirs, published by them, and Prime Signal does not measure it independently.

When someone visits your site from a coffee shop, airport, or coworking space, IP-based tools match every person on that network (50 to 100 people) to a single visit event. That inflates match rates on paper while delivering near-zero precision in practice. Our infrastructure uses geo frame resolution: each IP is cross-referenced against the most recently verified household coordinates, and geographically implausible matches are discarded. The provider reports a true usable match rate of 60 percent or better on that basis. That figure is theirs, and Prime Signal does not measure it independently.

Every identity is scored across 28 features in 6 groups: recency and velocity, intent domain signals, email engagement, web behavior, centroid similarity, and cross-feed composites. Signals are graded A through D. Grade A requires corroboration from 3+ independent sources and high active-research behavior. Grade D signals (generic news consumption, low search referral, no cross-feed corroboration) are filtered out before the audience is ever delivered to you. You never pay for noise.

The provider reports 14 million domains monitored for intent signals, against roughly 750,000 for the largest comparable network. That works out to about 18.7 times the coverage (derived: 14,000,000 divided by 750,000). The same provider reports over 250 billion URLs processed weekly, 100 billion or more weekly intent signals, and 400 million or more email engagement signals per day. These are the provider’s published figures, not Prime Signal measurements. Scale matters because it means signals come from specialist, high-density sources rather than the same handful of generic news and publishing sites that dominate smaller networks.

Prime Signal watches every campaign, ad set, and creative on your connected accounts around the clock. The schedules are fixed and published: ad status (disapprovals) and spend checks run every 15 minutes, funnel and tracking checks every 15 minutes, campaign metric syncs every 30 minutes, and creative fatigue scoring once a day. When a check finds an issue, Prime Signal either alerts you, recommends a fix, or acts on it, depending on your operating mode.

Four modes, configurable per client or per account. MONITOR sends alerts only. ADVISE sends alerts with AI-generated recommendations. GUARD automatically fixes safe issues (pausing bleeders, flagging policy violations) while escalating risky changes. AUTOPILOT handles everything autonomously with full audit trail. You can change modes anytime.

When an ad is flagged or disabled, Prime Signal classifies the violation type, explains exactly what triggered it, generates a compliant alternative, and resubmits. It tracks appeal success rates per platform and per violation type so the fixes get better over time. Policy Guard can also pre-screen ads before submission to catch violations before they cost you downtime.

Prime Signal learns what performs for your business, your audience segments, and your seasonal patterns. Budget continuously shifts toward winning campaigns and away from underperformers. If a campaign starts bleeding spend with declining returns, Prime Signal pauses it and reallocates. Your ROAS improves even when you are not watching.

Ad performance decays over time as audiences see the same creatives repeatedly. Prime Signal tracks CTR trends, frequency saturation, and engagement decay per creative. When a creative starts losing effectiveness, Prime Signal flags it for rotation with specific performance data before it wastes budget.

Yes, for the platforms that connect today: Meta (Facebook and Instagram), Google Ads, and GoHighLevel for CRM and automation data. One dashboard unifies metrics from every account you connect, with cross-platform performance comparison and unified spend tracking, so you are not jumping between platform dashboards. Microsoft Ads, TikTok Ads, LinkedIn Ads, X Ads, Snapchat Ads and Pinterest Ads are built but not released. The connect screen reads the deployed credentials and shows exactly which platforms are ready in your account, so it can never drift from this answer.

Explicit rules first, then statistics measured against each account’s own history, then AI reasoning for the judgment calls. Detection is deterministic: thresholds, policy rules, and day-of-week baselines built from your own account data. The reasoning layer, which writes diagnoses, recommendations and plain-language reports, runs on Anthropic’s Claude models through a single governed path that applies your organization’s AI policy and records every call. Prime Signal does not ship proprietary machine-learning models, and we will not claim one until it has real outcome labels and a published governance record behind it.

When you connect your accounts, Prime Signal syncs your full campaign history and begins building baselines specific to your business. It learns your normal CPA ranges, your seasonal patterns, your audience behavior, and your platform-specific quirks. The AI gets sharper with every campaign it monitors and every outcome it observes. There is no generic model applied equally to everyone.

Yes. The ReWire workflow engine lets you build visual automations with a drag-and-drop editor. Pre-built templates cover common scenarios like budget reallocation, creative rotation, performance alerting, and audience refresh. You can also build custom workflows with AI analysis nodes, conditional logic, parallel execution, and human approval gates. Prime Signal recommends workflows based on your account activity.

No. Your campaign data is never shared with AI providers for model training. LLM calls use zero-retention APIs. Your proprietary performance data, audiences, and creative strategies stay yours. The platform learns patterns specific to your accounts, but that learning stays within your organization.

Every AI action includes a full reasoning chain, the data it analyzed, the action it took, and the outcome. Approval gates let you require human sign-off before any high-impact change. If an action causes a regression, automatic safeguards trigger. Every decision is logged in an immutable audit trail you can review, search, and export.

Prime Signal acts as an always-available strategist for your ad operations. It analyzes your campaigns holistically and surfaces insights you would miss looking at individual platforms: which audiences are saturating, which creatives are approaching fatigue, where your budget allocation is suboptimal, and what your competitors are doing differently. It provides actionable recommendations, not just dashboards.

Prime Signal monitors competitor websites, ad copy, landing pages, and pricing changes. It detects when competitors launch new campaigns, change their positioning, or adjust pricing. You get alerts and strategic recommendations so you can react before being outmaneuvered.

Yes. Automated weekly performance reports and monthly summaries are generated for every account. If you manage clients, each client gets their own portal with real-time ad health, spend tracking, and ROI metrics. Reports are ready to share with no manual assembly required.

Prime Signal monitors every step of your conversion funnel: landing pages, checkout flows, form submissions, and redirect chains. The funnel scan runs every 15 minutes, so a page that goes down, a form that breaks, or a redirect that stops working shows up on the next pass. It also tracks conversion drop-off patterns and recommends where to focus optimization efforts.

Audiences are built from deterministic identity data across 308 million+ verified US consumer profiles with 65+ data points per person. You can build retargeting audiences from pixel data, lookalike audiences from your highest value customers, custom segments from intent signals, and suppression lists to avoid wasting spend on existing customers. All audiences are graded for signal quality before delivery.

The identity data provider reports 80 percent or better match rates for UID2 hashed email activation. Meta is the only ad platform Prime Signal activates against today. The reason the figure is high is that deterministic identity resolution produces clean, verified records that platforms can match against their own user databases, rather than cookie or IP inference.

Programmatic advertising is the automated buying and selling of digital ad space in real time. Instead of negotiating placements manually with individual websites, programmatic uses software to bid on ad inventory across millions of sites, apps, and connected TV platforms in milliseconds. When someone loads a webpage, an auction happens in the background. Your ad competes against others, and the winner's ad appears instantly. The advantage is precision: you define who you want to reach (by behavior, intent signals, demographics, geography) and the system finds them wherever they are online, across display, video, native, and CTV formats.

Programmatic is in development and is not part of a plan you can buy today. The design is to apply the same deterministic identity data and intent signals that power social campaigns, build audiences from verified identity records, and let the closed feedback loop validate which placements convert. We will say it is available when you can connect it yourself from the dashboard.

The underlying identity database is refreshed continuously. NCOA verification runs every 30 days on the full database. Email verification processes 10 million+ records daily. Intent signals are rescored on a three-layer cadence: daily signal rescoring, weekly grade recalculation, and monthly full model retrain. Your audiences always reflect current behavior, not stale snapshots.

Three plans, billed monthly or annually, with no per-platform or per-campaign fees. Amplify is $29 per month ($23 per month billed annually) and includes 20 AI actions a month, 2 ad platforms, 1 website, and the in-app assistant. Amplify Pro is $79 per month ($63 per month billed annually) and includes 100 AI actions, 5 ad platforms, 5 websites, and assistant support by app and email. Amplify Agency is $199 per month ($159 per month billed annually) and includes 300 AI actions, unlimited platforms and websites, and priority support; extra client accounts are $29 per month each and add 50 AI actions each. Annual billing is about 20 percent off the monthly rate. These are the same figures checkout charges.

Every plan includes around-the-clock ad monitoring, alerts, AI-powered diagnosis, budget overspend protection, and performance reporting. Usage-based charges apply only once you go past your plan’s included allotment, and they cover AI actions, SMS notifications, voice escalation, and report generation.

Every new organization starts with 100 evaluation credits, so you can connect a real ad account and watch Prime Signal work on your own campaigns before you pick a plan. If you would rather be walked through it, email hello@primesignalsoftware.com or book a demo from the website.

It takes over the work that usually sits with a media buyer watching dashboards, a separate monitoring or alerting tool, and the hours spent assembling client reports by hand. What that is worth depends on what you pay for those today, so we will not put a number on it. What the plans actually include is listed on the pricing page.