// FEATURES

Nine pieces. One platform.

Each of these works on its own. Together they replace the patchwork of CRM + sequencer + spreadsheet + directory most outbound teams stitch together.

QUICK ANSWER

What features does Castia include?

Castia combines discovery, a shared outbound board, multi-channel outreach, re-engagement, meeting prep, agent-mode updates, governance controls, scheduled jobs, and reusable blueprints in one platform.

Find and enrich accounts from public company sources, first-party lists, directories, and market signals.
Coordinate email, phone, LinkedIn, field notes, meetings, handoffs, and follow-up on one activity timeline.
Keep human approval, suppression rules, source records, privacy context, and audit trails inside the workflow.

weekly run·14 new·last sync 06:00

⌘R rerun

Northstar Cloud
SaaS · onboarding automation · 82 emp
owner: VP Growthsignal: hiring AEs
fresh
Atlas Underwriting
Insurance · commercial lines · 140 emp
owner: Head of Distributionsignal: new broker program
enriched
LedgerOps
Fintech SaaS · finance ops · 64 emp
owner: Revenue Leadsignal: funding
enriched
Harbor Mutual
Insurance · regional carrier · 310 emp
owner: Agency Directorsignal: territory expansion
enriched
IntegrateIQ
SaaS · integration platform · pending
owner: —
pending

01 — DISCOVERY

Stop paying for stale lists.

Business directories, geofenced web search, market signals, and first-party lists — pulled and cross-checked weekly. Each lead is enriched from public company sources so your team sees useful account context, not another generic address.

Source-aware
Business directories, company sites, review platforms, niche directories, and owned lists. Add your own.
Account enrichment
Decision-maker, domain, size, contact route, and trigger context from public company sources.
Suppression-aware
Customers, competitors, partners, and blocked accounts get filtered before they hit the board.
Scheduled refreshes
New leads land every Monday morning, automatically.

22 leads·Updated:Any24h7d

+ Add Lead

Discovered

Northstar Cloud

northstarcloud.com

coldSaaS

VP Growthin

6h ago

Atlas Underwriting

atlasunderwriting.com

coldInsurance

Distributionin

Mon

IntegrateIQ

integrateiq.com

coldSaaS

Mon

Reaching

Harbor Mutual

harbormutual.com

warmInsurance

Agency Directorin

Thu

LedgerOps

ledgerops.com

warmFintech SaaS

Revenue Leadin

Won

PolicyDesk

policydesk.com

warmInsurtech

Partnershipsin

02 — BOARD

Pipeline. Not paperwork.

Kanban with statuses your team actually uses. Activity timeline per lead. Territory views that group field work so every visit compounds.

Status pipeline
Discovered, researching, contacted, meeting, won, lost. Customisable per vertical.
Activity timeline
Email, phone, LinkedIn, in-person — one log, sorted by lead.
Route view
Filter by status and visit priority, group by territory cluster, drive less.
Changelog
Every change attributed and timestamped. Onboarding new team members takes minutes.

30d·4 variants live·top: saas-expansion-v3 · 17% reply

policy ON

saas-expansion-v3

best

Sent248

Accepted102

Replied42

Meeting14

Won5

accept 41%reply 17%paid 2%

linkedin-connect-v2

Sent112

Accepted37

Replied10

Meeting3

Won1

accept 33%reply 9%paid 1%

☏

phone-followup-v1

Sent64

Accepted18

Replied9

Meeting3

Won1

accept 28%reply 14%paid 2%

default

retiring

Sent84

Accepted8

Replied1

Meeting0

Won0

accept 9%reply 1%paid 0%

03 — OUTREACH

Four channels. Every variant tracked.

Email opens your client with a draft. Phone tracks the call without dialling for you. LinkedIn templates the note, you send. In-person logs the visit. Every variant runs as an A/B test by default — winners get more traffic next cycle, losers retire.

Per-vertical, per-channel variants
Formal or direct tone per market. Risk angle, value prop, and proof point baked into every versioned variant the platform tests against the others.
Full attribution funnel
Sent → accepted → replied → meeting → won — logged per variant, per channel, per market. You see which subject line pays, not just which one opens.
Engagement-gated escalation
Phone call only after the email moves when your policy requires it. Banners warn if you try to skip a step; nothing auto-sends in your name.
Auto-promote winners
When a variant beats the baseline by >2σ over the cohort, the agent shifts traffic toward it on the next run. Losers stay flagged for review, not silently killed.
In-person mode
Walk-in, business-card scan, voice note, on-the-go. The visit lands on the same activity timeline as the email and call.

41 dormant·6 signals today

9 re-engaged · 30d

Harbor Mutual
quiet 73d
harbormutual.com
website changelaunched broker growth program
signal_detecteddraft ready
draft assistant
you →
Win-back — reference the site relaunch, attach the integration shot.
agent →
"Hi Dana, saw the new broker program. We help teams keep every partner touch visible without slowing the agents down..."
integration-demo.pngAccept → step 1Push to Gmail
Northstar Cloud
quiet 41d
northstarcloud.com
linkedin hiringhiring 2 account executives
ready_for_warm_outreachre-engage
Atlas Underwriting
quiet 120d
atlasunderwriting.com
fundingnew commercial program announced
to_contactre-engage
LedgerOps
re-engaged 2d ago
ledgerops.com
website changeopened the integration link twice
respondedresponded

04 — RE-ENGAGEMENT

Quiet leads aren’t dead leads.

Every prospect who ghosted or said “not now” drops into a warm pipeline instead of a graveyard. Castia watches each one for a buying signal — a site relaunch, a new hire, fresh funding — and surfaces the moment to walk back in. Then a draft assistant writes the win-back with you.

Signal-triggered surfacing
Site changes, LinkedIn hiring, funding announcements — the moment a dormant lead moves, it resurfaces with the reason attached.
One-click stage moves
Push a lead through the win-back lane — imported → signal detected → contacted → re-engaged — straight from the card, no drawer.
A draft assistant that takes direction
Brief it like a colleague: “reference their relaunch, attach the integration shot.” It writes the email and pulls in your screenshots and mockups.
Accept, then push to Gmail
Approving a draft overwrites the step-1 sequence and drops it into your Gmail drafts, images inlined. Nothing sends in your name without you.

synced from calendar·matched · harbormutual.com

prepped ✓

Harbor Mutual · Working session

Tomorrow · 14:00 · 30 min · Dana R. + you

auto-prepped

Prep · assembled by the agent

Talking points

Broker handoff risk, partner coverage, renewal timing

Objection prep

'Too much admin' -> show approved workflow and audit trail

Competitor context

Today: CRM notes plus spreadsheet handoffs

Pitch angle

From scattered follow-up to one governed board in 8 weeks

After the call

notetaker joinedtranscript captured · 28 Min

summary

Interested in broker-growth workflow; budget available, decision in Q3.

decisions

Pilot with one regional team, start after summer.

action items

Send pilot scope by Friday
Attach broker-growth workflow example

✓ follow-up draft readyReview and send

05 — MEETINGS

Walk in briefed. Leave with the follow-up written.

Calendar invites sync in and match themselves to the right company. Before the call, Castia assembles the brief from the lead’s own history; a notetaker sits in and turns the conversation into notes and a follow-up draft — so the work after the call is review, not reconstruction.

Calendar-synced, auto-matched
Each invite becomes a meeting row, matched to its company by domain. No manual linking, no copy-paste.
Prep assembles itself
Talking points, objection prep, competitor context, and a pitch angle — drawn from the lead’s activity, not pieced together across five tabs at 13:55.
A notetaker takes the call
It joins automatically and turns the conversation into a summary, the decisions made, and a clean action-item list.
Follow-up, pre-written
The same transcript drafts your follow-up email. Review, edit, and send from the meeting page minutes after you hang up.

Chat·Local Gemma · read-only pipeline data

⌘Z undo

you →

Harbor Mutual briefed today. Renewal motion in Q3.

agent →

Status set to 'meeting booked'. Follow-up created for 14 Sep 2026. Want a prep email one week before?

you →

Yes, include the distribution-risk checklist.

agent →

Prep email saved. Risk anchors: partner coverage, handoff leakage, renewal timing. Send date: 7 Sep 2026.

Ask about the pipeline, campaigns, or replies… (Shift+Enter for newline)

06 — AGENT-MODE

Type. The board responds.

No forms, no field-by-field updates. The agent reads what you wrote, updates the right records, schedules the right follow-up, and surfaces the next move.

Plain-language input
DE, EN, mixed — speak the way you’d brief a colleague.
Multi-step actions
One sentence can update a lead, schedule a follow-up, and queue a mail.
Playbook-aware
The agent knows the sequence for your vertical and won’t skip steps.
Reversible
“Undo last action” works. Nothing gets shipped without your sign-off.

defaults that stay enforcedaudit log · CSV export

Channel eligibility rules
ON
Outreach follows the approved channel policy for the market, audience, and account type.
0 violations · 90 days
Privacy basis logging
ON
Source, purpose, owner, and retention status logged at lead creation.
1,284 events logged
Suppression pre-filter
ON
14,237 entries. Refresh weekly. New leads checked before they hit 'to research'.
last sync · Mon 06:14
Cold-call banner
ON
Warning shown if a B2B call is logged before any engagement signal exists.
3 warnings this month

07 — COMPLIANCE

The defaults that keep teams in control.

Channel eligibility, privacy logging, audit trails, and approval steps for regulated teams. The platform enforces them; the user does not have to remember them.

Channel eligibility
Outreach follows the rules you set for each market, channel, and audience.
Privacy log
Each lead has a per-event source, purpose, owner, and retention record.
Suppression enforcement
Customers, competitors, partners, and blocked accounts never enter the funnel.
Export-ready records
Privacy, audit, and activity records can be exported in minutes, not days.

12 jobs·4 shown

Refresh+ New Job

discover-saas-accounts

Weekly pull from directories, review sites, and hiring signals. Filter: mid-market SaaS, 50-500 employees.

0 6 * * 1

skill→discovery-search

✓ success8 May, 06:00(2m 14s)

Trigger now

enrich-company-profile

Decision-maker, domain, size, and trigger context from public company sources. Skip if already enriched.

*/5 * * * *

api-call→/api/leads/enrich

✓ success8 May, 22:55(20ms)

Trigger now

weekly-digest

Collects weekly activity and sends a summary to revenue leadership. Pauses when backlog exceeds 50.

0 17 * * 5

api-call→→ revenue lead · digest

⊘ failed3 May, 17:00(5m 1s)

Trigger now

stale-followup

Finds leads with no touch in 14 days and moves them to 'Follow-up due'. Policy banner active.

0 8 * * *

skill→stale-checker

✓ success8 May, 08:00(6m 15s)

Trigger now

08 — SCHEDULER

Things happen on time. Without you.

Discovery refreshes, enrichment, weekly digests, follow-up reminders, LinkedIn pending checks. Cron-driven, retried on failure, observable in one place.

Cron-driven jobs
Skill-as-a-job: define a prompt, set a schedule, watch it run.
Run history
Every run logged with output, timing, and replay button.
Retries & circuit breakers
Failed jobs retry; persistent failures pause the queue, not the platform.
Manual triggers
Run any job on demand from the UI when a discovery batch can’t wait til Monday.

4 blueprints·last optimisation 14:32·+18% vs last cycle

+ New Blueprint

SaaS · Mid-market expansion

Account discovery · signal outreach · AE handoff · 8-week cycle

v3live

Composes

discovery-searchprofile-enrichseq-saas-expansionweekly-digest

+18%reply rate vs v2 · 14d

last · 2h ago · 14 leadsnext · Mon 06:00

Insurance · Broker growth

Broker and agency teams · human approval · meeting-oriented

v2live

Composes

lookalike-discoverylinkedin-outreachseq-brokerstale-followup

+9%accept rate vs v1 · 21d

last · yesterday · 8 leadsnext · Wed 09:00

Recruiting · Tech

Senior engineers · referral-first · multi-touch

v1drafting

Composes

talent-discoverylinkedin-outreach

last · warm-up · pendingnext · manual review

D2C · Founder Reach

Brands €1–10M · Founder · Partnership offer

v4paused

Composes

storefront-scanfounder-enrichseq-partner-en

+4%reply vs v3 · paused for review

last · 11d agonext · —

09 — BLUEPRINTS

Skills are atoms. Blueprints are the playbook.

A skill is one thing the agent can do — discover, enrich, sequence, digest. A blueprint is what your team actually runs: a vertical-shaped composition of skills, templates, compliance defaults, and scheduler triggers, versioned and self-learning. You don’t buy skills. You buy the blueprint that runs them.

Composes skills into a playbook
One blueprint = a discovery query + an enrichment shape + a sequence + a compliance preset + scheduler triggers. Edit the blueprint once, every run inherits.
Self-learning between runs
Each cycle feeds reply, accept, and meeting metrics back into the templates. Phrases that converted on Tuesday are reused next Tuesday. Sources that bounced get downranked before the next discovery run.
Versioned and reversible
Blueprints are tracked v1 → v2 → v3 with deltas like “+18% reply vs v2 · 14d”. If a new version regresses, the agent rolls back automatically.
Per-vertical defaults built in
Tone, channel rules, approval steps, and regulated-market guardrails are inherited the moment you stamp out a new blueprint.

// VS THE ALTERNATIVES

What a platform does that a stack of tools can’t.

Capability	CRM	Sequencer	Spreadsheet	Castia
Auto-discovers prospects	—	—	—	Yes
Multi-channel activity log	Partial	Email only	—	Yes
Consent-aware sequencing built-in	—	—	—	Yes
Plain-language steering	—	—	—	Yes
Route and territory view	—	—	Manual	Yes
Per-vertical playbooks	—	Templates only	—	Yes
Re-engages quiet leads on a signal	—	Time-based only	—	Yes
Auto-preps meetings + writes follow-up	Manual notes	—	—	Yes
Stores deals after-the-fact	Yes	—	Sort of	Sends to your CRM

// PILOT SHAPE

Eight weeks. Your territory. Real numbers.

We co-build the playbook for your vertical, run it on a real territory, sit with the data at week 4 and week 8. If it works, we scale. If it doesn’t, you’ve learned what doesn’t work and we both walk away.

Book a working session See the use cases