Active ICPs
2
Primary + secondarySettings / Control Plane
System configuration workspace
Review the current outbound operating system configuration before live operator editing and external provider adapters are added. This workspace reads from typed config, the shared repository layer, and selector-backed control-plane view models.
Active offers
1
Front door + laterScoring tiers
3
High / medium / lowConfigured channels
2
Email + researchMock integrations ready / pending
2 / 3
Ready / pendingFeature flags
Runtime toggles for experimental or high-risk automation features. Enable these only when you are actively monitoring the pipeline.
Auto-enroll leads after enrichment
When enabled, leads that complete enrichment with a valid email contact are automatically enrolled into their recommended campaign without manual intervention.
Env var: FEATURE_AUTO_ENROLL · Current: false (inactive)
Bounce circuit breaker
Monitors the 24-hour rolling bounce rate via Mailgun webhooks. Automatically pauses all outbound sending if bounces exceed 5% of sends (minimum 10 sends). Manual pause and resume controls are available below.
Global outbound sending
When paused, the sequence cron exits immediately on every tick — no emails are sent until sending is manually resumed or you call resumeSending().
How the circuit breaker works
The bounce rate is evaluated after every Mailgun permanent.fail or temporary.fail event. If bounces ÷ sends exceeds 5% over a 24-hour window with at least 10 sends, sending is automatically paused and a Slack alert is fired.
Webhook endpoint: /api/webhooks/mailgun/bounce
Required env var: SLACK_ALERT_WEBHOOK_URL — Slack Incoming Webhook for alert notifications.
Warmup schedule
Automated daily send cap ramp for domain mg.driveinsight.ca. The active row is determined by the most recent starts_at ≤ today. No manual env var changes needed.
Domain
mg.driveinsight.ca
Active week
Week 2
Daily cap
40 / day
| Week | Starts | Daily cap | Status |
|---|---|---|---|
| Week 1 | Apr 14, 2026 | 20 / day | Past |
| Week 1 | Apr 14, 2026 | 20 / day | Past |
| Week 1 | Apr 14, 2026 | 20 / day | Past |
| Week 2 | Apr 21, 2026 | 40 / day | Active |
| Week 2 | Apr 21, 2026 | 40 / day | Active |
| Week 2 | Apr 21, 2026 | 40 / day | Active |
| Week 3 | Apr 28, 2026 | 60 / day | Upcoming |
| Week 3 | Apr 28, 2026 | 60 / day | Upcoming |
| Week 3 | Apr 28, 2026 | 60 / day | Upcoming |
| Week 4 | May 5, 2026 | 80 / day | Upcoming |
| Week 4 | May 5, 2026 | 80 / day | Upcoming |
| Week 4 | May 5, 2026 | 80 / day | Upcoming |
| Week 5 | May 12, 2026 | 150 / day | Upcoming |
| Week 5 | May 12, 2026 | 150 / day | Upcoming |
| Week 5 | May 12, 2026 | 150 / day | Upcoming |
ICP configuration
Current ICP profiles, tier guidance, channel preference, pains, proof angles, and front-door offer recommendations.
Primary ICPU.S. independent used car dealerships
Primary Dealer ICP
Dream-fit independent dealers selling roughly 11-25 cars per month with a website, a claimed Google Business Profile, visible review signals, and enough digital maturity to respond to structured outbound.
Target industries
Automotive retail • Independent auto retail
Target subindustries
Independent used car dealerships • Owner-led local dealer groups
Preferred channels
First-offer recommendation
Reviews / Reputation • Reviews / Reputation
Dream fit summary
Dealers moving roughly 11-25 cars per month with visible review-response gaps, claimed GBP coverage, and enough digital maturity to act on outbound quickly.
Tier 2 summary
Still attractive when the same dealer shape exists but urgency is softer, review pain is less acute, or contact confidence needs another human pass.
Avoid summary
Avoid dealers below the minimum sales floor, accounts without website or GBP foundations, and operators already deeply locked into Birdeye or Podium. All car dealerships — independent, franchise, and auto group — are valid ICP targets.
Primary pains
Inconsistent review generation
Weak review response discipline
Underperforming Google Business Profile presence
Limited time for structured marketing operations
Proof angles
Visible public review and response gaps
Operational clarity instead of marketing fluff
Evidence tied to appointments, not just sends
Secondary ICPU.S. independent used car dealerships
Secondary Dealer ICP
Still-fit dealers with similar operational signals, but usually slower-moving, more pragmatic, and less digitally active.
Target industries
Automotive retail • Independent dealer operations
Target subindustries
Smaller independent used car dealerships • Pragmatic local dealer teams
Preferred channels
First-offer recommendation
Reviews / Reputation • Reviews / Reputation
Dream fit summary
Best secondary-fit dealers show real post-sale follow-up or review friction, light tool adoption, and enough trust to engage with a simpler operating conversation.
Tier 2 summary
Pragmatic dealers with visible pain but slower urgency, smaller teams, or less appetite for aggressive personalization should stay in the active working set.
Avoid summary
Avoid groups that want heavyweight systems immediately, dealers with no visible digital footprint, or operators already deeply committed to another reputation stack.
Primary pains
Inconsistent customer follow-up
Lower urgency around reputation gaps
Less appetite for complex systems
Proof angles
Low-friction setup
Simple reputation lift story
Operational fit for smaller teams
Offer configuration
Current seeded offers, their operating role, usage footprint, fit signals, pricing notes, and CTA guidance.
ActiveReviews / Reputation
Reviews / Reputation
A dealer-first reputation offer focused on review response discipline, trust signals, and measurable credibility lift.
Offer role
First-offer / front-door wedge
Availability
2 active campaigns • 200 configured sequences
Pricing notes
No structured pricing is configured in seed data yet.
Problem solved
Weak or inconsistent review management that erodes shopper trust and hides visible reputation gaps.
Fit signals
Google rating around 3.2-4.0
15-150 reviews
Low or inconsistent review response
Visible trust gap against nearby competitors
CTA guidance
Invite the dealer to a short working session to review the visible reputation gaps and where appointments may be leaking.
Lead with the visible review-response gap and how it affects trust without adding operational burden.
Scoring / priority configuration
Priority tiers, score buckets, recommended handling, visual treatment, and routing notes for queue management.
Dream / Tier 1High fit bucket
Tier 1
Dream-fit dealer accounts with clear pain, reachable decision-makers, and strong front-door offer alignment.
Score range
80-100
Recommended offer
Reviews / Reputation
Recommended channels
Email • Manual Research Follow-Up
Bucket meaning
Ready for priority review and likely campaign enrollment.
Visual status treatment
Dream / Tier 1
This tier uses the same priority signal treatment seen across queue and review surfaces.
Readiness / routing notes
Route into fast operator review, validate the decision-maker, and prioritize campaign-ready enrollment work.
Tier 2Medium fit bucket
Tier 2
Still valuable dealer accounts, but with lower urgency or slightly weaker fit signals.
Score range
60-79
Recommended offer
Reviews / Reputation
Recommended channels
Bucket meaning
Worth keeping in rotation with lighter personalization.
Visual status treatment
Tier 2
This tier uses the same priority signal treatment seen across queue and review surfaces.
Readiness / routing notes
Keep active in the working queue with lighter personalization and a slower review cadence behind Tier 1.
Avoid / Tier 3Low fit bucket
Tier 3
Low-priority or lower-confidence accounts that should not consume early workflow attention.
Score range
0-59
Recommended offer
Reviews / Reputation
Recommended channels
Bucket meaning
Hold or suppress until stronger signals appear.
Visual status treatment
Avoid / Tier 3
This tier uses the same priority signal treatment seen across queue and review surfaces.
Readiness / routing notes
Suppress from early workflow attention until stronger fit, timing, or contact signals appear.
Workflow / channel configuration
Current outbound lanes, approval behavior, activation notes, and future placeholders for operator workflows.
ConfiguredPrimary operating lane
Primary outbound lane for campaigns, sequences, replies, and appointment-driving CTA tests.
Objective
Run personalized first-touch and follow-up motion while keeping provider complexity outside the product for now.
Approval behavior
Operator approval required
Current coverage
3 active campaigns • 200 active sequences
Provider binding
email contract available
Activation notes
Email-first is straightforward to model in campaigns and sequences.
It supports personalization without heavyweight provider complexity.
Replies become structured learning data for future classification and booking flows.
Launch approvals and ambiguous reply handling stay operator-controlled in v1.
Operator assistedSupporting operator lane
Manual research follow-up
Human-led research pass used to sharpen proof points, qualification notes, and timing-sensitive follow-up.
Objective
Support Tier 1 and timing-sensitive accounts with better context before or after the core email motion.
Approval behavior
Operator approval required
Current coverage
1 scoring tier recommends this assist lane
Provider binding
Workflow placeholder only
Activation notes
Used as a supporting motion where public signals justify extra human attention.
No autonomous sending, scraping, or enrichment is attached to this lane yet.
PlannedFuture workflow lane
LinkedIn workflow
Reserved for a future research-assisted social workflow once channel guardrails and provider choices are settled.
Objective
Add governed social context collection and optional manual follow-up after the email backbone is stable.
Approval behavior
Recommendation only
Current coverage
0 live workflows • policy and provider still undecided
Provider binding
Workflow placeholder only
Activation notes
Provider selection and safe-usage rules are not defined yet.
Any future motion must remain recommendation-first until approval boundaries are proven.
Memory / learning configuration
A first-pass view of tracked outcomes, optimization targets, approval boundaries, and memory categories using mock/config-only settings.
Learning policy
The learning layer should absorb reply, appointment, and operator-review outcomes while leaving final send, restart, and escalation decisions with humans.
Mock readyappointment
Appointments booked
Track scheduled and completed appointments as the north-star outcome for campaign quality.
Current mock signal
0 scheduled or completed appointments
Mock readyreply
Positive replies
Capture interest and booking-intent replies so offer and sequence quality can be compared.
Current mock signal
0 positive replys
Mock readyreply
Objection and timing patterns
Track objection, not-now, and wrong-person signals to improve routing and future recommendation quality.
Current mock signal
0 objection or timing signals
Mock readymemory entry
Operator review gates
Store human hold/approve decisions for sensitive re-entry, scope, and workflow constraints.
Current mock signal
3 constraint entrys • 0 flagged replys
Optimize for
Booked appointments over raw send or reply volume
Cleaner first-offer selection for each ICP profile
Fewer false-positive re-entry recommendations after timing replies
Higher-confidence routing into operator review when signals are mixed
Recommendation behavior
Approval required
Launching or restarting campaigns and sequences
Promoting a later-offer motion like NAPS into active follow-up
Acting on low-confidence or ambiguous reply classifications
Automatic recommendations
Draft ICP and offer-fit suggestions from visible signals
Suggested routing notes for scoring tiers and review queues
Proposed memory tags and insight summaries for operator approval
Memory entry categories
0 entrys
System learning
Durable patterns the system should eventually promote into reusable knowledge.
0 entrys
Operator note
Human notes captured during review, qualification, or follow-up decision-making.
1 entry
Playbook
Stable execution guidance such as email-first or offer-positioning rules.
3 entrys
Constraint
Guardrails that block automation, restarts, or unsafe recommendations until review.
Environment / integration readiness
Practical readiness state for future providers and infrastructure, represented as mock control-plane status only.
Mock readyOutbound engine
Email provider
Campaign, sequence, enrollment, reply, and channel contracts are stable enough for adapter work, but no live provider is wired.
Next step
Attach a provider adapter and secret management without changing the selector or page surfaces.
Blocked by
Credential management and send safeguards
Mock readyLearning layer
AI provider
Insight and memory entities can already absorb recommendations, but model/provider selection is still open.
Next step
Choose the provider contract, evaluation rubric, and approval boundaries for recommendations.
Blocked by
Model policy and evaluation harness
PendingCore platform
Database
Typed repositories and selectors are in place, but persistence is still entirely in-memory and mock-backed.
Next step
Introduce schema-backed repository implementations behind the existing data-access contracts.
Blocked by
Schema design and migration plan
PlannedScheduling
Calendly
Appointment entities exist, but booking-provider sync and inbound event handling have not started yet.
Next step
Define a booking event contract and map it into typed appointment lifecycle states.
Blocked by
Provider contract and event mapping
PlannedGrowth ops
LinkedIn workflow
Only a placeholder workflow exists today, with no provider contract or automation layer behind it.
Next step
Decide whether this remains manual research assist or becomes a governed provider-backed lane.
Blocked by
Policy and provider decision