Fragmented context
Account state, campaign performance, budgets, delivery health, and automation history live in separate systems.
16 channels · one decision layer
Turn channel data into your next best move.
PulseDeck unifies every campaign, then turns observed performance into a guarded budget test, an explainable confidence score, and optional Claude-backed answers grounded in your numbers.
Free planNo cardDeterministic demoSecure ingestion API
app.pulsedeck.io/dashboard
Sample decision brief
A safe test is ready
82
Spend
$18.2kROAS
4.2xConf.
82/100+18.4% sample trend
Reallocate a capped test from the lowest-efficiency paid channel to the strongest evidence-backed performer.
Guardrail
Evidence passed
Action size
≤ 15% test
Illustrative sample workspace. Live decisions use your connected data.
16
normalized channel types
16
working import + ingest paths
0
keys needed to evaluate
100%
explicit data provenance
The operating problem
The reporting stack knows what happened. It rarely knows what deserves attention next.
Channel tabs fragment attribution, spreadsheets hide freshness, and generic AI summaries can sound certain without a measurable basis.
Slow investigation
Teams rebuild the same exports before they can compare a real movement with a normal range.
Unsafe handoffs
A recommendation becomes a chat message with no scope, idempotency, approval boundary, or audit trail.
Operational value
Estimate the reporting time the workflow could release
Adjust the assumptions for your team. The result is a planning estimate, not a performance claim.
Planning estimate
5.4h
potential weekly reporting time released
$1,052/mo
Illustrative operational value only. This is not a promised saving or performance result.
Automation examples
Route evidence into the system that owns the next step
Provider failure → support
A signed sync.failed event carries provider and non-technical error context into a help-desk or team channel.
Lead → CRM
A workspace-scoped lead event enters a CRM workflow with one event ID for deduplication.
Weekly performance → client report
A scheduled workflow reads normalized metrics, assembles a report payload, and sends it to the buyer's selected delivery system.
Pacing risk → account owner
Transparent spend projection crosses a configured threshold before a human decides whether to adjust anything.
Payment failure → recovery
A verified Stripe invoice event can trigger the buyer's recovery sequence without exposing billing credentials.
Data export → backup
A scoped key reads normalized data on schedule and transfers it to buyer-owned storage.
Trust boundary
No green connected badge without a real data path
Demo rows are labelled. OAuth cannot start without app credentials. Provider approval and buyer-system credentials remain visible external activation items.
AES-256-GCM
Provider tokens and webhook secrets at rest
Signed delivery
Timestamp-bound HMAC and replay controls
Read-only ads
No AI or workflow changes campaign state
16 integrations, one dashboard
Every channel your marketing team already runs
Evaluate every channel instantly with explicit demo data, then move to real metrics through CSV or a secure automation endpoint. Stripe Payments, Meta, and Google Ads also include direct server-side reporting adapters.
Google Analytics
Google Ads
Meta Ads
LinkedIn Ads
TikTok Ads
Microsoft Advertising
Pinterest Ads
Snapchat Ads
Reddit Ads
X Ads
Mailchimp
Klaviyo
Shopify
Stripe Payments
HubSpot
Search Console
Decision intelligence
One confident move. Not sixteen noisy guesses.
PulseDeck's deterministic engine sits between raw reporting and AI. It produces an auditable recommendation first; Claude can then explain or explore it without inventing the underlying numbers.
01
Normalize the signal
Compare paid channels on one clean efficiency and evidence model.
02
Apply the guardrails
Require a meaningful ROAS gap, minimum spend, conversions, and channel breadth.
03
Ship a bounded test
Cap the move, show confidence, and model upside from observed performance only.
Decision brief
Scale the strongest signal safely
Budget test
$450
Modelled upside
+$1,350
Health score
82 / 100
Evidence guardrail
PassedRecommendation withheld automatically when channel breadth, spend, conversions, or the observed efficiency gap is insufficient.
Illustrative demo values. Live decisions use the connected workspace's normalized metrics.
Features
Built for teams that report on marketing, not chase spreadsheets
01
Decision Engine
Convert observed channel efficiency into a capped budget test, projected upside, confidence level, and an explicit hold when the evidence is insufficient.
02
Unified analytics
Compare spend, impressions, clicks, conversions, ROAS, and CTR across all supported channels with visible source and attribution context.
03
Day-one data ingestion
Every channel accepts normalized CSV and scoped ingestion. Stripe Payments, Meta Ads, and Google Ads also include direct reporting paths.
04
Grounded AI insights
Ask plain-language questions, review anomaly flags, and generate ad-copy suggestions grounded in the workspace metric window.
05
Evidence-first recommendations
Every proposed move is deterministic and auditable, with minimum-sample guardrails and a maximum 15% source-budget test.
06
White-label delivery
Agency workspaces can apply their own brand name and accent colour throughout authenticated product surfaces.
How it works
Three steps, zero spreadsheets
1
Sign up free
Create a workspace in seconds - no credit card, no API keys required to start.
2
Connect your channels
Import a CSV, connect Make/Zapier/n8n to the scoped ingestion API, or activate an approved direct reporting adapter.
3
Act on one guarded decision
PulseDeck ranks the signal, proposes a capped test when evidence clears the guardrails, and lets Claude explain the data on demand.
Pricing
Simple plans that scale with your channels
Growth
Recommended$49/mo
- Up to 3 workspaces
- 5 members per workspace
- 6 connections
- 90-day history
- 100 AI queries / month
- Weekly AI digest
Agency
$99/mo
- Up to 10 client workspaces
- 25 members per workspace
- 50 connections
- 365-day history
- 1,000 AI queries / month
- Weekly AI digest
- White-label branding
FAQ
Frequently asked questions
Do I need provider credentials to get started?
No. Every workspace starts with deterministic evaluation data. Real metrics can then arrive through normalized CSV, a scoped ingestion API, or an activated direct provider adapter where supported.
Is the AI connected to a live model?
PulseDeck calls Claude through the Anthropic API when a key is configured. Without a key, responses remain clearly labelled as evaluation output so the product never implies a live model call.
How does the Decision Engine choose a budget move?
It compares paid-channel efficiency, checks minimum evidence, caps any proposed test at 15% of the source budget, and explicitly holds when evidence is insufficient. It is a test-sizing tool, not an automatic media buyer.
How are real channel metrics imported?
Every provider supports normalized CSV and a hashed, revocable ingestion token for n8n, Zapier, Make, or scheduled jobs. Activated direct adapters add provider-native reporting where supported.
Can I white-label the workspace?
Yes, on the Agency plan. The configured brand name and accent colour are applied to authenticated product surfaces.
How is billing handled?
Upgrades use Stripe Checkout and the Billing Portal when Stripe is configured. Otherwise the product reports the missing configuration explicitly.
Is workspace data isolated?
Authenticated reads and writes resolve a signed workspace membership, and tenant-isolation tests cover the core data paths.