Helm is not a bundle of separate tools. Every product feeds the same organisational graph. The more you use, the smarter the platform becomes — without anyone filling out a profile.
The integration architecture
Crew
The organisational graph. People, teams, skills, relationships — built automatically from signals.
Atlas
The intelligence layer. Queries the graph to answer questions no other tool can.
Product integration map
Every product is both a standalone tool and a signal source for the organisational graph. No product is an island.
Sends to Crew
Receives from Crew
Sends to Crew
Receives from Crew
Sends to Crew
Receives from Crew
Sends to Crew
Receives from Crew
Sends to Crew
Receives from Crew
Sends to Crew
Receives from Crew
Sends to Crew
Receives from Crew
Sends to Crew
Receives from Crew
Critical integration points
If any of these break, the graph degrades and Atlas becomes unreliable. These are the integration points we validate continuously.
01
Moonboard → Crew
Relationship seeding
When employee A creates a board for employee B, Crew creates a relationship edge between them. Every message, reaction and contribution reinforces the graph. Cross-team interactions receive higher weight.
Real-time02
Docs → Crew
Structure confirmation
When a manager approves a document request, Crew receives one of the strongest structural signals available — a real-world workflow confirming the manager-report relationship. Not imported from a CSV. Observed from behaviour.
High weight03
Launchpad → Crew
Org structure foundation
New employee onboarding creates the highest-weight structural signal in the system. Department, team, manager, location and start date — all confirmed through a live workflow, not a static import.
Highest weight04
Spotlight → Crew
Skill and influence graph
Every recognition contributes a relationship edge. Skill tags build the expertise graph. Cross-team recognitions receive a 1.5× weight multiplier. Manager recognitions receive 1.3×.
Weighted05
Learning → Crew
Certified skill evidence
Completed certifications carry the highest skill signal weight. Expiry dates enable automatic decay — so the graph stays current without anyone maintaining it.
DecayingTwo paths to Helm
Two fundamentally different entry points with different goals, timelines and success metrics. Both lead to the same platform.
The transition between paths
When a Moonboard-first user sets up their organisation, everything carries forward — boards, messages, contributions. Historical interactions are retroactively matched to employee profiles and converted to relationship signals.
The transition is voluntary, clearly motivated and low friction. The graph gets its first real data the moment the import completes.
Signal architecture
Every action in Helm emits one of seven signal types. Each carries a weight, a source and a timestamp. Together they build the most complete picture of how your organisation actually works.
Relationship
Board created, message posted
Moonboard, Launchpad
Recognition
Peer shoutout, skill tagged
Spotlight
Lifecycle
Document requested, policy signed
Docs, Launchpad
Skill
Certification complete, tag added
Learning, Spotlight
Contribution
Demo presented, project shipped
Showcase
Structure
Manager confirmed, team assigned
Launchpad, Docs
Sentiment
Check-in submitted
Pulse
Confidence-weighted
Every edge in the graph carries a confidence score derived from signal weight, recency and source diversity. Atlas never asserts without evidence.
Time-decaying
Signals lose weight over time. A certification from five years ago matters less than one renewed last month. The graph stays current automatically.
Cross-team amplified
When signals cross team boundaries, they receive higher weight. Cross-team collaboration is harder to observe and more valuable to surface.
The graph promise
Month 1
Month 3
When Atlas ships
Atlas readiness
We will not launch Atlas to any organisation until their graph has sufficient coverage to produce reliable results. Insufficient data returns an honest “not enough information” rather than a guess.
Per-organisation gate
Built on trust
Three levels of inference
Level 1: public knowledge. Level 2: team-level insights, visible to managers. Level 3: sensitive HR-only queries with audit logging and 90-day retention.
Evidence, not assertions
Every Atlas insight shows a confidence score and a "Why?" drawer. Every recommendation links back to the source signals that informed it.
Employee transparency
Every employee can see what Helm knows about them and why. No hidden scoring. No black-box profiles. Disputed signals surface to HR admin.
Database-level access control
Row-level security on every table. Salary data in restricted tables. Document URLs signed and expiring. Audit log append-only and immutable.
GDPR by design
Built to EU standards from day one. Data export, right to erasure and consent management built into the architecture, not bolted on.
Humans decide
Atlas recommends. People decide. The system surfaces confidence-weighted options — accountability stays with your team, not with the algorithm.
Every Moonboard message is a relationship signal. Every Docs request is a lifecycle event. By the time Atlas ships, your organisational graph is already populated.
No card required. Free to start.