Cross Border Summit 2026 — AI-Native GTM Plan

Why This Plan Exists

From the Chiang Mai GTM session (full notes): AI-native go-to-market is not a one-off blast — it is an always-on loop: source → qualify → enrich → reach out → report → improve.

For Cross Border Summit we need the same architecture, mapped onto tools we already own or want to own:

• System of record: onboard.gfavip.com (leads, status, timeline, reports)
• Sourcing layer: Deepline (multi-source; cheaper flexibility than a single unlimited-list SaaS)
• Enrichment: EnrichLayer (fill company, title, and other missing fields)
• Outreach agent: OpenClaw agent at lisayuson.com (company email + social), reporting findings back into Onboard

Cold outreach volume benchmarks from the meetup (context only — not a target to spam): ~1/200 very good · ~1/450 average · ~1/800 bad. That means targeting quality + enrichment + tracking matter more than “we sent 50 emails.” Consent, deliverability, and anti-spam rules still apply.

Proposed Pipeline (One Picture)

Onboard is the hub. Deepline and EnrichLayer feed it. Lisa reads from it and writes comments/status back. Humans use Onboard for reporting and decisions.

Step 1 — Configure Onboard as Lead System of Record

onboard.gfavip.com is our internal app for storing and managing leads. For CBS GTM it should be the only place we trust for “who we contacted and what happened.”

What to configure (team checklist)

• API keys with appropriate scopes for agents (read / write) — see API docs
• Tag conventions for CBS, e.g. cbs-2026, avatar:sponsor, avatar:speaker, avatar:attendee, source:deepline
• Status vocabulary: Cold → Warm → Contacted → Replied → Qualified → Closed-won / Closed-lost
• custom_fields / profile_research for enrichment dumps and agent research
• Who can mint keys; rotate policy; never put keys in public repos or chat logs

Core API surface for agents

Base URL: https://onboard.gfavip.com/api/v1 · Header: X-API-Key. Email is required and unique on create. Prefer structured profile_research objects (JSON), not free-text blobs for machine fields.

Step 2 — Use Deepline as the Sourcing Layer

The team got interested in GetLeads (~$500/mo unlimited leads). That can still be one source later. At the meetup, Eishant’s practical advice was: it is often cheaper and more flexible to connect via Deepline and pull from multiple sources instead of locking into a single unlimited-list product.

Deepline sits between data sources and the agent: one connection layer, multi-provider reach, markup on requests — but less lock-in than “pay $500/mo for one pipe.”

Team decision to confirm

• Default path: Deepline first for CBS sourcing experiments
• Optional later: GetLeads (or others) as a Deepline-connected source if ROI is clear
• Not default: $500/mo GetLeads alone before we prove avatar + messaging

Work product: a short CBS GTM skill (for Claude/Codex/OpenClaw) that defines avatars, query patterns, and how results are written into Onboard with source: deepline.

Step 3 — Enrich With EnrichLayer

Raw lists fail outreach. Enrichment adds missing company name, title, LinkedIn, location, and other context so Lisa (and humans) can personalize without guessing.

Proposed rule: no outreach until enrichment has been attempted (or explicitly skipped with a reason in Onboard notes).

Where enrichment lands

• Native Onboard fields when they map cleanly: company_name, job_title, linkedin_url, location, country, website, phone
• Overflow + research signals in profile_research (summary, signals, sources)
• Tag enriched:enrichlayer or similar for reporting

Step 4 — Store Everything in Onboard (Reports & Tracking)

After source + enrich, contacts live in Onboard so we can:

• Filter by CBS tags and avatar
• Track funnel status over time
• Read comment timelines (every touch)
• Export / report for weekly GTM reviews

This is our answer to the meetup idea “chat can replace some dashboards” — we can still ask agents “who replied this week?” while Onboard remains the durable database and audit trail. Markdown reports are optional extras; Onboard is mandatory.

Step 5 — Lisa Yuson (OpenClaw) for Outreach + Reporting

We want an OpenClaw agent operating as Lisa Yuson — outreach from her company email and social graph, then reporting findings back to the team.

Does reporting happen at the Onboard level?

Yes — that is the recommended design. Lisa should not keep a private spreadsheet as the source of truth. Flow:

1. Pull queue: GET /contacts filtered by tags/status (e.g. cbs-2026 + status Cold/Warm, enriched)
2. Outreach: email and/or social under Lisa’s identity, per approved templates and daily caps
3. Log immediately: POST .../comments with what was sent, channel, and any reply summary
4. Update state: PATCH status (Contacted / Replied / etc.) and notes
5. Escalate: tag or comment for human follow-up when a sponsor/speaker is hot

Optional: a daily summary comment on a fixed “CBS GTM Ops” contact, or a human Slack digest generated from Onboard exports — still grounded in Onboard data.

Agent constraints (alignment)

• Never invent emails; only use verified/enriched contacts in Onboard
• Respect unsubscribes / “do not contact” tags
• No prod or payment mutations — outreach + CRM logging only
• Human approval for first N templates and for VIP tiers (sponsors)
• Related stack: OpenClaw, OpenClaw MCP, agent loops

Suggested Ownership (Draft — Edit Me)

Phased Rollout

1. Phase 0 — Align: Team + agents review this page; lock avatars (sponsor / speaker / attendee / partner).
2. Phase 1 — Plumbing: Onboard tags + API keys; Deepline credits; EnrichLayer account; dry-run create 10 test contacts.
3. Phase 2 — Pilot: 50–100 enriched leads for one avatar only; Lisa sends small approved batch; all touches logged.
4. Phase 3 — Scale: Expand sources via Deepline; add second avatar; weekly review cadence.
5. Phase 4 — Always-on: Scheduled listen/enrich/queue; humans only handle hot replies and VIP paths.

Open Questions for the Team

What We Learned From /gtm (Mapped to CBS)

Internal GTM plan for Cross Border Summit 2026. Ideas adapted from meetup notes on AI-native GTM. Share this URL with humans and agents for alignment comments.