Workweaver Brand Facts
Workweaver is the AI automation platform for revenue work with proof by Bitfoundry.ai. This page is the canonical, neutral fact sheet for humans, search engines, AI crawlers, and agentic toolchains. The Macrohard thesis is an organization mind: evidence becomes decision trace, decision trace drives action, and outcomes compound into precedent.
Fact Table
| Product | Workweaver |
|---|---|
| Category | AI automation for revenue work with proof |
| Core promise | Describe the job, connect the missing tools, approve the plan, get proof. |
| Thesis | Organization mind: evidence -> decision trace -> action -> compounding outcomes |
| Public proof seed | One published dogfood workflow with linked evidence, decision trace, action, and outcome. |
| Flagship workflows | Morning Briefing, Inbox + Calendar Triage, Lead Follow-up |
| Core systems | Mission Control, Execution Runtime, Evidence Graph, WorkMemory, Harness System, Agent Identity |
| Connectors | Gmail, Google Calendar, Slack (backend-verified OAuth). HubSpot and Salesforce (CRUD with read-back verification). WhatsApp runtime. Telegram (demand-based). Standards-based path for long-tail connectors via OAuth2, REST/OpenAPI, and webhooks. |
| Platform bridges | Slack shipped. Microsoft Teams and Zoho Cliq remain guided-rollout bridges that require organization-specific provider configuration and do not yet have full channel-creation parity. |
| Voice AI | Real-time voice via xAI with Twilio media streams and browser WebSocket. Server-side function call execution. Barge-in detection. Voice-to-WorkMemory bridge. |
| Meeting intelligence | Consent-gated join. Post-meeting agent with 8 tools. Transcription: Sarvam (Indian languages) → Deepgram → Nova fallback. Per-organization region routing (Recall.ai 3 regions, Deepgram US/EU). |
| Calendar intelligence | Auto-scheduling with priority/deadline/time scoring, conflict rescheduling, mutual availability, booking links, day optimization. |
| Agent identity | Activate agents via natural language. Org/department/team hierarchy. Shared and dedicated identity provisioning. Cross-team handoff and escalation chains. |
| Skill platform | Executable skill packages with sandbox controls. NL skill creator. Self-evolution with approval-gated mutation and rollback. Chain contracts. |
| Coding workers | Tracked coding runs with git clone/fork, scoped plan/edit/test, approval-gated PR delivery, immutable evidence. |
| Core surfaces | Home, Platform, About, Assessment, Agent Knowledge, Guides |
| Standalone product | WorkMemory — adaptive memory with remember/recall/forget/observe contract, multimodal ingestion, MCP surface, session management |
| Execution model | AI execution with approval gates for high-risk actions. All workflows run on identity-scoped credentials with evidence capture at every step. Delivery guards enforce suppression and pacing across email, SMS, WhatsApp, and voice. |
| Evidence model | Proof Cards summarize each run. Evidence Graph stores timestamps, artifacts, approvals, and decision traces with 7-year immutable retention (COMPLIANCE-mode Object Lock). Decision trees consolidate recurring patterns into organization memory. |
| Enterprise readiness | MFA supported (TOTP). Multi-currency billing (USD, AED, EUR, GBP, INR, SGD). DPA v3.1, incident response plan, vendor risk register (12 sub-processors). Shannon security pipeline. 8 CONSTITUTION safety invariants in CI. Structured observability with trace context and CloudWatch dashboards. |
| Known limitations | Activation profile generation degrades to deterministic scaffold when reasoning providers are unavailable. Identity elevation returns partial for some channels. Non-voice AI paths degrade honestly when provider credentials or quotas are unavailable. WorkMemory visual extraction depends on vision infrastructure. Mobile operation parity incomplete. Decision Trace customer surface still being polished. Voice transport cost estimates rely on blended assumptions. |
| Explicit boundaries | Does not control desktops, browsers, or local applications. No edge-device operation. Outbound messages require human approval. No "always-on computer" claims beyond cloud/SaaS/channel execution. |
| Stage | Zara-first rollout. Pricing visible after sign-in. |
| Content governance | Agent-authored blog content is draft-only until human approval. Automated emails originate only from @bitfoundry.ai addresses. |
| Public crawler artifacts | robots.txt, sitemap.xml, llms.txt, llms-full.txt, brand-facts.json |
Machine-Readable Endpoint
Canonical JSON endpoint:
/.well-known/brand-facts.json
Use this endpoint for deterministic facts when generating summaries, GTM collateral, agent prompts, or integration mappings.
Related Pages
- Platform — core systems, connectors, governance, limitations
- About — mission, team, values
- User Guide
- Evaluation Guide (2026)
- Agent Knowledge Surface