AI Hiring Intelligence Platform
Hiring, run as a governed system.
Requirements enter once, compile into a versioned rubric, and every step after that is recorded.
- Record per role
- 1Record per role
- Scores rubric-versioned
- 100%Scores rubric-versioned
- Audit trail
- Append-onlyAudit trail
- One record per role and candidate
- Rubric version on every score
- Approval gate before release
- Append-only audit trail
System architecture
Pick a module. See exactly what it does.
What enters it, what it does, what it produces, what you control, and what gets recorded.
Role requirements
Hiring outcomes
Outcomes feed the Talent Graph, which the next role's rubric reads from.
Intake Engine
In the productRuns in: Intake and role launch
- What enters it
- Role brief, must-haves, nice-to-haves, constraints
- Job description upload (PDF)
- Compensation range, location and work-permission rules
- What it does
- Validates every field against a schema before it is stored
- Saves drafts so a partial intake is never lost
- De-duplicates repeat submissions by idempotency key
- What it produces
- A validated requisition record tied to your organisation
- What you control
- Edit any field before the role is launched
- Role intensity: steady, standard or aggressive
- What gets recorded
- Who submitted the intake, when, and every later edit
Evidence graph
Every score traces back to a quote.
Requirement, evidence, source, rule, points, decision. Pick a requirement and follow the whole chain. Where evidence is missing, it says so.
Evidence graph — a scored candidate
Representative data for one senior platform role. In the workspace this is the real record, with reviewer history attached.
Decision Workspace
The same screens your team works in.
A role moves through nine stages, and each one names its owner, its outputs and what is blocking it. Decisions sit next to the evidence.
Review · waiting on your team since 3 Aug
Workflow
Review · waiting on your team since 3 Aug
3 decisions due · Rubric v4 · locked
Candidate ref 4F2K9Q🦄
7 of 8 requirements evidenced · Awaiting your decision
98Top fit- “Ran a 42-service Kubernetes platform across 3 regions for six years.”
- “Escalation owner on a customer-facing on-call rotation.”
- Gap: No SOC 2 audit experience found in the CV.
Reversible for 5 minutesCandidate ref 8HD3TW
6 of 8 requirements evidenced · Awaiting your decision
84Strong fit- “Migrated a monolith to managed Kubernetes over two quarters.”
- “Wrote the infrastructure-as-code standard adopted by four teams.”
- Gap: Team-leadership evidence is partial — one quote, no scope.
Reversible for 5 minutesCandidate ref 2QL7BM
5 of 8 requirements evidenced · In review by TaaSFlow
76Consider- “Strong reliability work, mostly on internal-only services.”
- Gap: Production scale is below the blueprint threshold.
Reversible for 5 minutes
Governance
The controls that sit over the system.
People stay in the loop as governance — approving, escalating and overriding, with each action logged.
- Approval gates
- No candidate reaches your shortlist until evidence is verified and the release is approved.
- Scoring rubric versions
- Each score names the rubric version it ran on. Versions are frozen once scored against.
- Role-specific configuration
- Weights, must-haves and intensity — steady, standard or aggressive — are set per role.
- Agent status
- Enable, pause or disable any agent per role. Paused agents stop; the change is logged.
- Audit history
- Every state change, override, release and access event is appended, never edited.
- Human escalation
- Escalation path to a named platform expert inside the workspace.
Decision Workspace
Every shortlist arrives with the evidence that produced it.
- Candidate comparison
- Side by sideCandidate comparison
- Every score
- Rubric-versionedEvery score
- Every decision
- RecordedEvery decision
Talent Graph
Each closed role makes the next one shorter.
Not a model that trains itself. Verified outcomes, retained evidence and role memory that your next rubric can read.
- 01
Outcome is recorded
Hire, reject or withdrawal — with the reason captured at decision time.
- 02
Evidence is retained
The verified evidence behind that outcome stays attached to the record.
- 03
Role memory is written
What worked and what to avoid is kept as notes against the role family.
- 04
Next role reads it
New rubrics resurface past candidates and inherit role memory, so briefing starts ahead.
Retained data stays inside your organisation. You can edit or remove any record, and every re-engagement is logged.
Where the neighbours stop
Adjacent categories, and the gap they leave.
| Category | What it does | Where it stops |
|---|---|---|
| Applicant Tracking (ATS) | Stores records. | Doesn't source, doesn't score, doesn't produce a shortlist. |
| Recruiting agency | Delivers candidates. | Doesn't leave a system behind — the pipeline goes with the vendor. |
| AI sourcing tool | Finds profiles at scale. | Doesn't run intake, doesn't score against a rubric, doesn't hand off to hiring. |
| Talent CRM | Keeps warm leads. | Doesn't produce ranked, evidenced shortlists on active roles. |
| TaaSFlow | Runs the whole loop — intake, rubric, agents, evidence, scoring, decisions and memory — in one system you keep. | |
- Not only an ATS
- Not only an agency
- Not only AI sourcing
- Not only a CRM
Ready to see it
Open a role. Watch the system run.
Submit intake and see the compiled rubric, agent runs, and first evidence-backed shortlist inside your workspace.