Set the course
Define what AI should do, who owns it, what context it needs, and what safe success looks like.
Build safely
Use OpenData’s factory, your team, or the AI tools already in motion without losing the build trail.
Run with control
Keep every workflow visible through ownership, support, risk, value, and change history.
One course, many confident views
The same workflow answers finance, security, operations, and support.
Forge records the intent once. Factory preserves the build trail. Pulse turns the result into the view each role needs: value for finance, provenance for technology, ownership for operations, and evidence for security.
forge.opendata.tools/runs/r-8f3a91 Live
Forge run · executed 12 seconds ago
pr-automation Engineering
CFO view
$95
- Time recovered
- 38 min
- Tier
- Quality
- MTD recovered
- $22.4k
CTO view
Pass
- Audit hash
- 8f3a91…ec2
- Approver
- K. Stein
- Verification
- Tests 42/42
Forge
Where the business sets the course.
Forge captures what AI should do, what context it needs, which policies apply, who can approve it, and what safe success looks like. The workflow starts with direction instead of becoming unmanaged work after the fact.
forge.opendata.tools/factory/wf-payment-fraud-review Live
Security Verify · running
A real workflow, narrated
Watch Forge define, build, and hand off a workflow.
One workflow, six steps, the same shape across every factory. The cycler rotates examples; the lifecycle doesn't change. The counter is dollars saved on this single run, computed from time saved × your hourly rate, minus build and token cost.
Ticket arrives Saved this run $0
- 1Ticket arrives
PROJ-1248 · Ship password-reset feature.
- 2Forge drafts the plan
Pulling ticket → drafting spec → identifying affected services → generating test plan.
- 3Branch + commit + tests
$ git push origin auth/password-reset · 47/47 tests passing
- 4PR opens
pr #1247 opened · 2 files · 184 lines · auto-summary attached
- 5Security review attached
Auth review: clean. Snyk triage: no new CVEs.
- 6Handoff + value logged
#eng-shipped: $187 saved, workflow record written to Pulse.
- 1Merge request hits the gate
svc-billing · merge to main · touches /admin scope.
- 2Forge runs the review
Pulling diff → OWASP scan → secret-scan → dependency CVEs → IAM impact.
- 3Findings + tier assigned
0 high · 2 medium · 4 low · tier: Risk avoidance · fully-loaded value $340
- 4Policy gate fires
admin-scope change → named approver required · paused awaiting M. Levine
- 5Approver signs
M. Levine → approved with note · evidence hash c4d2f0…811
- 6Compliance evidence written
SOC 2 ledger updated · 6 gates re-verified · $340 saved, board-ready.
- 1Draft submitted for review
pulse-launch.md · 1,420 words · marketing draft.
- 2Forge reads the brand voice
Pulling style guide → banned-word grep → tone analysis → length check.
- 3Edits proposed
12 corrections · 3 banned-word swaps · 2 sentence-fragment fixes
- 4Human approves with one tweak
K. Stein accepts 11/12 · keeps "ramp" instead of "scale"
- 5Document published
pulse-launch.md → docs site · brand-voice gate: pass
- 6Logged to Pulse
Tier: Routine · $48 saved · workflow record written.
- 1Forecast cycle triggered
Q3 capacity forecast · 12 active workflows · 30 day horizon.
- 2Forge pulls the signals
Workflow health → utilization → pipeline weighted → hiring lag → known leaves.
- 3Drift detected
Engineering capacity gap ↑ 6% week-over-week · likely Q3 over-capacity by 2.3 FTE.
- 4Recommendations drafted
Defer Eng-2 hire · reallocate to Pulse capability ramp · attribute $42k saved.
- 5COO reviews + signs
M. Levine reviews · accepts with adjustment · evidence hash 5e2b77…
- 6Forecast posted to leadership
Pulse dashboard updated · $220 saved on this run · tier: Risk avoidance.
or scroll Auto-cycling workflows every ~12s
Factory
Build with OpenData’s software factory, your team, or the AI tools you already use.
You do not need an AI Factory to start. OpenData brings one. If your teams are already trying Cursor, Claude, ChatGPT, n8n, internal scripts, or a custom agent platform, Factory gives that work a build record, verification path, and handoff into Pulse.
forge.opendata.tools/library/capability Live
84 capabilities · 10 domains · ramp independently
Starter Standard Senior
Engineering Lifecycle
Docs & Brand
Ops & Media
Platform / Meta
Workflows
Security
Thinking
Design
Research
DevOps
Each meter is its own dial. Add factory capabilities in a domain → ramp that domain. Tier marks at 33% / 66% / 100% of headroom.
Operationalized value
Know what AI is actually returning.
Every managed workflow can carry a value model. We log the time saved, multiply by the rate that role would normally bill at, subtract the build and token cost, and keep the result attached to the workflow in Pulse.
| Tier | Time saved | Internal rate | Examples |
|---|---|---|---|
| Routine | 15–45 min | $150 | Status reports, grant packets, intake routing |
| Quality | 30–90 min | $150 | Test generation, quality checks, API review |
| Risk avoidance | 60–180 min | $150 | Security review, change control, compliance evidence |
200 managed workflow runs per month × 45 minutes saved each = 150 hours of capacity recovered = $22,500/mo at $150/hour.
What adding 5 factory capabilities looks like.
Add 5 factory capabilities at Quality tier. 200 runs/month each × 30 minutes saved × $150/hour = $7,500/mo per capability. Five capabilities = $37,500/mo recovered, tied to the workflows Pulse manages.
"But doesn't AI make math and code errors?"
Real concern, real answer. Forge does not blindly trust model output. Runs can require verification, tests, type checks, deterministic math, policy gates, and named human approval before Pulse treats the workflow as ready to run.
Pulse
Where AI adoption stays on course.
Pulse is the run layer for operational AI: ownership, support, change management, security posture, evidence, value, and lifecycle health. It keeps AI workflows visible after Factory hands them off.
pulse.opendata.tools/governance/signals Live
Watching · 312 runs/hr · 47 policy gates · 12 connectors
47/47
Gates green
312
Runs / hr
2
Open tickets
8.4k
Replayable
Recent governance signals
Policy halt · agent attempted deprecated
/admin/users at 14:32 UTC, change ticket #C-1142 opened, security on-call pagedCourse ownership
Every workflow has a named owner and support path.
Change control
Routed, reviewed, approved, logged.
Evidence & replay
Every run remains inspectable when questions come up.
Real-time security signals
Suspicious calls halted, ticketed, surfaced.
Support & escalations
Escalation surface for the things the gates catch.
Value tracking
Capacity, cost, and risk reduction tied back to the workflow.
Integrations
Your tools can come with you.
OAuth where the vendor supports it, API keys where they don't, Model Context Protocol for AI-native systems, and adapters for the AI tools your teams already use. Factory keeps the build path flexible without losing control.
pulse.opendata.tools/connectors/health Live
30+ integrations · last sync per source
HubSpot
just now
Salesforce
just now
Pipedrive
just now
Close
just now
Productive
just now
Monday
just now
Asana
just now
Basecamp
just now
Jira
just now
Linear
just now
GitHub
just now
GitLab
just now
Shortcut
just now
Slack
just now
MS Teams
just now
Discord
just now
Notion
just now
Confluence
just now
Google Docs
just now
QuickBooks
just now
Xero
just now
Stripe
just now
BambooHR
just now
Gusto
just now
Rippling
just now
Harvest
just now
Toggl
just now
Clockify
just now
Claude MCP
just now
OpenAI
just now
Onboarding 20 design partners through Q2.
Tell us where your teams are trying tools and where AI is already showing up. Within three business days you get a 90-day course plan:
Set course
Workflows, owners, context, policies, and approval paths.
Build in Factory
OpenData’s software factory first, or the AI tools you already use.
Launch
Release paths, support handoff, rollback, and change records.
Run with control
Health, risk, value, support, and accountability in Pulse.
Pricing on the deployment call, not a quote engine.
or