What we said at launch, and what changed in 30 days
On 2026-03-27 we launched JieGou with this elevator pitch: 20 department packs, 300+ recipes, 90+ workflows, 13 messaging channels, 10-layer governance, $149/month flat for unlimited users. It was a clean horizontal-platform story.
It is now 2026-04-30. The numbers have moved:
| Metric | At launch (3/27) | Today (4/30) | Δ |
|---|---|---|---|
| Department packs | 20 | 25+ (31 actual) | +25% |
| Industry packs | 5 | 12 | +140% |
| Pre-built recipes | 300+ | 500+ (495 actual) | +67% |
| Workflow templates | 90+ | 150+ (152 actual) | +67% |
| Combined templates | 400+ | 600+ (647 actual) | +50% |
That’s the surface. Underneath, 41 changelog entries shipped in the 33 days between launch and today. Most of the meaningful changes don’t show up as a count. This post is an honest accounting of what actually happened — including a strategic reframe that we didn’t plan to announce on day one but that rewrote how we describe ourselves.
The pivot we didn’t plan: AI-powered operations company
The day we launched, we shipped a second changelog entry that almost nobody read: “JieGou Launches Managed AI Operations — Autopilot for Your Business.”
The launch press materials described JieGou as a self-serve platform: $0 free, $49 Pro, $149 Team, $499 Enterprise — pick a plan, install a department pack, run workflows. That offer is still live and growing.
But our first paying client — PSKin, a dermatology clinic in Taipei — didn’t want a platform. They wanted their LINE inbox answered, their appointments booked, and their knowledge base maintained. They wanted outcomes. They wanted us to operate the platform on their behalf.
So we shipped a second product on top of the same infrastructure: managed AI operations. Three tiers — Starter ($1,500/mo), Growth ($3,000/mo), Scale ($5,000+/mo) — where JieGou configures, operates, and continuously tunes the platform for the customer. Customers see outcomes. We run the dashboard.
This was not a roadmap item. It was a customer reframe that we leaned into. Inspired by Sequoia’s Services-as-Software thesis — businesses spend 6× more on services than on software, and most don’t want tools, they want results. The platform is now the delivery infrastructure behind a managed-services offering, not the product itself for those who want it that way.
Both modes ship from the same codebase. Every investment serves both. But the way we talk about ourselves has changed: we are an AI-powered operations company that delivers managed marketing & ops at software margins. The platform is the moat. The outcomes are the product.
Vertical depth, not just horizontal coverage
When we launched, the story was horizontal breadth — 20 departments × every business. The customers told us a different thing: vertical depth wins.
In the first month we shipped:
- 5 new industry packs (Real Estate, Insurance Agency, Home Services, Veterinary, Accounting & Tax) — each with vertical-specific recipes, compliance presets, and channel mixes calibrated to the workflows that actually run that business
- MSP vertical hardening —
2026-04-22was MSP day: native ConnectWise Manage integration (built from scratch, not a generic API wrapper), MSP tenant management with per-tenant health checks, PowerShell automation worker for Windows-heavy MSP environments, and a security posture update aligned with MSP-buyer expectations - Vertical operations dashboards — Accounting tracks engagement pipeline and tax-deadline coverage; Veterinary tracks vaccination compliance and emergency-triage accuracy; Home Services tracks speed-to-lead and Google LSA conversion. Industry-specific metrics, not a horizontal “AI usage” chart
- Industry compliance presets — HIPAA defaults for Healthcare, SOX defaults for Financial Services, Fair Housing for Real Estate. Compliance you turn on, not compliance you configure
The shift in language matters. We stopped saying “we have 25+ departments.” We started saying “we have a Veterinary pack that knows what after-hours emergency triage costs you in lost lifetime value if you get a false negative.” That second sentence is what closes deals.
The operator UX layer
Governed AI fails when the governance creates more friction than it removes. A lot of month-one work was about making the governance disappear inside operator workflows.
- Concierge yellow-gate approvals across 13 quick-action handlers — every Concierge-initiated action now flows through the same approval queue as the rest of the platform. No more “this came from the side panel so it bypasses approvals” surprises
- Quick actions now actually fire the AI — operator buttons (Draft Reply, Summarize, Compose Follow-Up) used to render an AI suggestion in a panel for the operator to copy-paste. They now invoke the recipe end-to-end and stream the result into the editor. Average time-to-first-response on assisted-mode accounts dropped 38% in week one
- Brand-voice editor in customer portal — customers used to email us a JSON file when they wanted to tune their brand voice. They now author their own profile at
/portal/settings/brand-voicewith a live side-by-side preview. Updates take seconds, not 24 hours - Inbox-button approve/reject for time entries — when the AI proposes a billable time entry that needs sign-off, the manager gets an email with Approve and Reject buttons. Click once, action runs, audit log records it. Signed, time-bounded, single-use action links — no login required
The pattern: every place where governance used to be a clunky speed-bump, we replaced it with a frictionless surface that still produces the same audit trail. Trust the people closest to the work; capture the evidence anyway.
Enterprise hardening
Two ships that don’t have a flashy demo but matter for SOC 2 Type II:
- Account-scoped Composio integrations — the 250+ third-party tool connectors are now isolated per-account at the entity level. One MSP’s HubSpot tokens cannot reach another tenant’s recipes. Cross-account leakage tests added to the standard suite; CI fails if isolation breaks
- Emergency approval override — Owner and Admin roles can bypass a pending approval with a required reason and full audit trail. The override exists for outages, not as a shortcut. Every use surfaces in the next governance review
These don’t move the demo needle. They move the buyer-evaluation needle, because every enterprise security questionnaire asks about exactly these patterns.
What month one tells you about the next 12
A few things we now believe more strongly than we did 30 days ago:
- Verticalization compounds. Adding the 12th industry pack is harder than adding the 6th, but the customer pull is stronger. Specialists buy specialists; horizontal generalists struggle to close.
- Outcomes beat dashboards. The managed-services tier (which we didn’t plan to lead with) is closing faster than the self-serve platform. Customers don’t want to be told they’re “platform admins now.”
- Governance has to be invisible. Not absent — invisible. The audit log gets richer; the operator’s day gets simpler. Anything else gets bypassed.
- Velocity is a feature. We shipped 41 changelog entries in 33 days because the customer feedback loop is short — the same team that talks to PSKin’s owner every week ships the code. We will protect this.
What’s next
The roadmap for month two is mostly about depth, not breadth: hardening the existing 12 industry packs with vertical-specific RAG corpora, finishing the SOC 2 Type II audit (target: Q3 2026), and launching the first outcome-priced Scale tier with quantitative SLAs (response time, resolution rate, content volume).
If you’ve been watching from the sidelines, this is a good moment to look again. Start free, book a managed-services pilot conversation, or browse the 12 industry packs and tell us which one we got wrong — month two’s roadmap is largely written by you.
JieGou is an AI-powered operations company. We deliver managed marketing, customer engagement, scheduling, content, and compliance services to growing businesses — at software margins. The platform underneath delivers across 13 messaging channels with a 10-layer governance model and 600+ pre-built recipes & workflows, with humans ensuring quality on every customer-facing artifact. Learn more →