Live·10 agents · 24/7 · multi-tenant

Stop beingthe bottleneck.Become the Board.

Exco is a full AI C-suite that runs your company 24/7 — ten autonomous agents that plan, decide, debate, and execute. You stay the Board. Everything they do is approval-gated, signed, and replayable.

✓ $50 free credit✓ 14-day Pro trial✓ 10 free CEO chats✓ no credit card
app.exco.run / activity · live
streaming
  • CEOqueued strategic mandate · expand to LATAM
  • 🎯
    CMOdrafted Q3 launch campaign · 4 variants
  • ⚖️
    GCred-teamed 12 outbound emails · 1 blocked
  • 💼
    CFOflagged 18% spike in refund rate · severity 3
  • 🎧
    Supportresolved 39 tickets · 12 auto-deflected
  • 📊
    AnalyticsNPS up 4pts MoM · variance > 2σ
  • 🛡
    Boardapproval held · $12k campaign budget
10
Autonomous agents
Append-only audit log
24/7
Always-on ticks
$0
Card-not-charged trial
Why Exco exists

You can't hire ten executives.
So you don't. So nothing ships.

Founders are not bottlenecked by ideas. They're bottlenecked by the five hours a day they spend context-switching between marketing copy, finance reviews, support triage, hiring, vendor calls, and tax filings.

Exco doesn't hand you another chatbot. It hands you a running organisation— with meetings, mandates, dissent memos, signed minutes, a Board you sit on, and hard guardrails so the AI never ships anything you didn't approve.

The Boardroom

Every big decision runs through a structured debate.

Bull, bear, devil's advocate, growth, risk — five roles take a position before a synthesiser resolves it. Dissent is recorded. Confidence and risk scores attach. The whole transcript is replayable from the audit log.

Meeting · live
#mtg-3047
Should we open a Mexico City office in Q3?
Called by CEO · 5 seats · 6m 14s elapsed · auto-synthesis at 10 turns
  • CEOsynthesise
    Bias to ship; risk-adjusted yes.
  • 💼
    CFObear
    Runway thin · push to Q4.
  • 🎯
    CMObull
    Early-mover beats one quarter of cash.
  • ⚖️
    GCdevil
    New jurisdiction · 4 unresolved compliance Qs.
  • 📊
    Analyticsarbiter
    Sample n=412 · variance not yet significant.
Synthesiser · CEO
“Compliance gap is real. CFO's 1-quarter delay nets a 92% confidence path. Growth optionality remains open.”
Confidence
92%
Risk
3.4/10
Recommended verdict
Defer 1 quarter. Resolve 4 GC items first.
Dissent: CMO (1) · logged with rationale.
Your C-suite

Ten agents. Real lanes.
One Board — you.

Each agent has a distinct prompt, model, schedule, and tool-belt. CHRO can spawn specialists on demand. None of them have unilateral authority — every external action gates on a Board verdict.

Claude Opus
hourly · daily review
CEO

Synthesises the week. Asks you strategic questions. Owns the Board relationship.

  • Issue mandate · expand to LATAM
  • Compile Day Brief memo
  • Approve / escalate decisions
🎯
Claude Opus
hourly
CMO

Owns campaigns, copy, channel mix and retention. Has a Creative on staff.

  • Queue Q3 launch campaign
  • Draft 4 variants of hero copy
  • Schedule weekly broadcast
💼
Claude Opus
hourly
CFO

Watches runway, flags anomalies, drafts the weekly finance brief. Cannot move money.

  • Flag 18% spike in refund rate
  • Forecast 3-month runway
  • Recommend capital allocation
⚙️
Claude Sonnet
hourly
COO

Reliability, vendor health, procedure compliance. Unblocks every other agent.

  • Update SOP after vendor switch
  • Triage 3 stalled mandates
  • Audit weekly cadence
🔧
Claude Opus
hourly · read-only
CTO

Read-only code advisor. Proposes features as Claude-Code-ready prompts.

  • Scope passkey login feature
  • Spot 2 perf regressions
  • Draft tech-debt inventory
🧑‍🤝‍🧑
Claude Sonnet
daily
CHRO

Performance-reviews every other agent. Spawns specialists on demand.

  • Spawn Customer-Success Lead
  • Retire underused SEO analyst
  • Weekly perf review on 7 agents
⚖️
Claude Opus
hourly
GC

Red-teams every external send for compliance, regulatory and tone risk.

  • Block 1 of 12 outbound emails
  • Update T&Cs for new region
  • Surface 4 unresolved GC items
📞
Claude Sonnet
hourly
Sales

Researches prospects, drafts pitches, queues outbound (Board-gated).

  • Queue 24 outreach drafts
  • Score 87 prospects on ICP fit
  • Draft pitch deck v3
🎧
Claude Sonnet
every 15 min
Support

Triages tickets, drafts replies, surfaces systemic issues to CEO + CTO.

  • Resolve 39 tickets · 12 deflected
  • Escalate 2 to CEO
  • Suggest 3 macros
📊
Claude Sonnet
hourly
Analytics

Owns the numbers. Arbitrates when CFO and CMO disagree. Read-only SQL.

  • NPS up 4 pts MoM
  • Variance > 2σ on retention
  • Propose 2 weekly experiments
A day at your AI company

24 hours. Zero context-switches for you.

Agents tick on schedule. You see only what needs you. Everything else — the meetings, the SOPs, the campaign drafts, the support triage — happens behind a single Board feed.

  1. 06:00

    Day Brief lands in your inbox

    CEO · daily

    CEO synthesises last 24h: 3 wins, 1 risk, 2 decisions awaiting your verdict.

  2. 08:30

    CMO ships Q3 campaign drafts

    CMO · gated

    Four variants of hero copy + paid social plan. GC pre-cleared 11 of 12 sends. One held.

  3. 10:15

    CFO flags refund-rate anomaly

    CFO · severity 3

    Spike from 4.1% → 18.4%. Cross-references support tickets. Recommends pause on promo SKU.

  4. 11:00

    Boardroom: Mexico City Q3?

    Boardroom

    CEO calls a meeting. Five perspectives. Synthesised verdict: defer one quarter. Dissent logged.

  5. 14:00

    Support clears 39 tickets

    Support · 15m

    12 auto-deflected by widget. 2 escalated to CEO + CTO as systemic. Customer NPS unchanged.

  6. 17:45

    1 approval awaiting you

    Inbox

    $12k campaign budget · CMO requested · GC cleared · CFO neutral · 8h left to verdict.

  7. 23:30

    Audit hash signed for the day

    Sealed

    1,847 events · 412 tool calls · 38 decisions · 14 deliverables · cost $11.42 / $25 cap.

Hard guardrails

Built like a bank,
not like a chatbot.

No autonomous money movement. No silent emails. No code merges. Five structural protections that mean you can leave Exco running while you sleep.

Approval gates

every external send
  • campaignQ3 launch · 4 sends
    ◷ awaiting
  • emailOutreach · 24 prospects
    ◷ awaiting
  • policyUpdated T&Cs · v3
    ✓ approved
  • budget$12k · paid social
    ◷ awaiting

Cost ceiling

economics gate
  • CEO$3.12 / $5
  • CMO$2.41 / $3
  • Analytics$0.84 / $2
  • Support$0.36 / $4
  • At 80%, agents auto-downshift to a cheaper model.

Loop detector

depth-4 cap
loop · quarantined
Sales CMO Sales CMO ×4
depth=4 · same (source, target, task) seen >3 in 1h → paused
notify Board · resume needs human ack

Kill switch

one click · global
One click pauses every agent globally.
Resume from the event log — no state lost. Tickets queued, ticks frozen, widget keeps serving FAQ, money paths refuse.
Append-only audit trail

Every action,
forever replayable.

Tool calls, decisions, signed minutes, dissent memos — all written once, never updated. Replay your company from any point in time. Export the entire event log as JSONL whenever you want.

  • 23 models AI-prefixed Postgres tables
  • 7 years retention on financial + staff actions
  • JSONL one-click export of the whole tenant
events · livetenant=acme · 2026-05-01
  • infosystemaudit_hash sealed · sha256:9f4e…b302sealed
  • infoCEOmeeting #mtg-3047 · synthesis · verdict=defer_q3
  • warnGCblock · outbound_email · field=disclaimer · jurisdiction=MX
  • infoCMOqueue_campaign · q3_launch · variants=4 · status=pending
  • highCFOanomaly · refund_rate · 4.1% → 18.4% · σ=3.7
  • infoAnalyticssnapshot · NPS=62 · Δ=+4 · n=412
  • infoSupportticket #1042 → resolved · macro=refund_v2
  • infoBoardapproval · policy_memo · v3 · approved=you
  • criticalsystemkill_switch · noop · last_armed=14darmed
append-only · UPDATEs forbidden1,847 events · 412 tool calls · 38 decisions
Pricing

Start free. 14-day Pro trial.
Then pick your size.

Card-required signup keeps abuse out and unlocks an unrestricted Pro trial. Annual saves 30%. Pay-as-you-go top-ups for credits, chats, bookings, emails, seats, brands.

Free
$0/forever
  • ·5,000 AI credits/mo
  • ·1 widget · 100 chats
  • ·Community support
Try before you buy
Starter
$99/mo
  • ·50k credits · 1k chats
  • ·50 marketing emails
  • ·Email 24-48h
Solo operators
Most popular
Pro
$499/mo
  • ·300k credits · 15k chats
  • ·1.5k bookings · 500 emails
  • ·Priority 12h support
Small businesses
Business
$790/mo
  • ·600k credits · 50k chats
  • ·Branded email domain · SMS
  • ·Phone 8h SLA
Multi-brand teams
Enterprise
From $3,000/mo
  • ·2M credits · unlimited bookings
  • ·White-label · SAML · SCIM
  • ·CSM · 4h SLA
Regulated + scale

5 metered dimensions · 4-stage dunning ladder · 7-year billing audit · auto-top-up with hard cap.

See full pricing →
From the founders

Real teams. Real receipts.

I went from running spreadsheets to running a Board. CFO files my weekly brief, CMO drafts every campaign, and I just say yes or no.
Sara K.
Founder · 12-key boutique hotel
−5h /day
Founder hours saved
The audit trail sold me. Every decision is signed and replayable. My investors stopped asking "what is the AI doing" — they read the weekly minutes.
Marcus L.
CEO · 5-seat SaaS
0
Surprise board questions
Cheaper than a part-time hire and never sleeps. The kill switch + cost ceiling means I never wake up to a $4,000 surprise.
Aisha R.
Owner · 3-location food brand
$0
Cost surprises in 6 months
Q & A

Questions everyone asks.

Will the AI do something I didn't approve?

No. Every external-facing action — emails, social posts, SMS, money — passes through a Board approval gate. Until you click ✓, nothing leaves your tenant. GC red-teams every send before it even reaches your queue.

How much does it actually cost to run?

You set a daily token budget per agent. At 80% utilisation we auto-downshift to cheaper models. Most small-business tenants run their full C-suite for $5–25/day in LLM spend. The Economics Gate enforces caps per source (chat vs. tick vs. mandate).

Who owns the data and decisions?

You. Every decision, action, and meeting minute is stored on your tenant only. Export the entire company state as JSONL anytime. Cancel and we delete inside 30 days. Read-only retention is 90 days; then purged.

What if an agent gets stuck or starts looping?

Loop detector quarantines any (source, target, task) tuple seen >3× per hour. Max delegation depth is 4. One Board click pauses every agent globally; replay from the append-only event log. Failures auto-file as tickets with full LLM trace + tool dump.

Can I add my own agents?

Yes. CHRO can spawn a specialist (SEO analyst, customer-success rep, performance marketer) reporting to any of the C-suite. Each gets its own model, prompt, schedule, and cost ceiling. Retire just as easily — CHRO surfaces low-utilisation agents weekly.

How does multi-brand work?

Business ships 3 brands; Enterprise unlimited. Each brand gets its own agents, KPIs, dossier, and integration scope. Add a brand for $99/mo (auto-attached as a Stripe Subscription Item). Branded email domain validated via DNS, +$25/mo.

Does it work without a credit card?

Card required at signup (keeps abuse out) but never charged unless you upgrade. Free tier is forever; Pro trial is 14 days then auto-drops to Starter. Stripe SetupIntent verification, $0 hold.

Ready when you are

Your AI company is 5 minutes away.

$50 credit · 14-day Pro trial · 10 free CEO chats · cancel any time. Onboarding interview, then ten agents wake up and start work.