A person answers for every step.
Who can do each step. Who does it. How well it went.
HRPulsar started as a skills platform for people. The direction is work planned and measured across people and agents, with a person who answers for every step an agent takes. The first question, who can do each step, is where the product starts: Coverage. The rest is direction.
This page describes where HRPulsar is going. Every block says what already ships and what does not.
The vision grows from a platform that already ships.
Not a blank slate. The skills-first core it builds on, the workforce model, the capability data and the assessments behind it, is live in production now.
Talent platform v2.5.3
Competencies, assessments, development, recruiting, grading, exams, talent market and analytics, in the product and the demo.
Coverage
Describe a process, see which steps an agent can take, which stay with people, and where nobody covers it yet. The gap becomes a vacancy; the agent step comes with a skill file that runs in your own tools.
AGPLv3 core
Self-host on your infrastructure or use the cloud. The core platform is open and inspectable.
From AI readiness to execution.
Most organizations can assess readiness. Running day-to-day work across people and agents is a different problem:
Unknown agent ownership
Agents do real work, but nobody owns them the way managers own teams.
No shared skill model
Human skills and agent capabilities live in different worlds, so nobody sees the full capacity of a team. This is the gap Coverage is built to close.
Thin governance
Quality controls and policies that exist for people rarely extend to the agents working next to them.
Weak link to outcomes
Readiness scores are easy. Connecting capability to delivery speed, cost and quality is the hard part.
Three questions. Six pillars.
Who can do each step, who does it, how well it went. Each pillar carries its status: shipped, with the version it shipped in; in development; next; or later.
Who can do each step
The question Coverage answers, per step, for your own process.
Capability Mapping
One skills model already covers people: frameworks, assessments, grades. Coverage describes what an agent type can do in the same seventeen capabilities, so a step reads the same whoever takes it.
Workforce Graph
Employee profiles and competencies ship today. The registry of the agents a company runs is an API today: each active access an employee has to an agent carries a responsible person, and a change of that person goes to the audit log. Its screen is next, and so is an owner for the agent itself.
Who does it
Execution stays in your own tools. What we add is the control around it.
Quality & Risk Controls
Coverage marks the steps where an agent drafts and a person must check, and asks who signs. Enforcing that at run time, with escalation paths, is next.
Work Orchestration
Route a step to the right executor, human or agent, by capability, policy and load. Coverage says who could take a step; it does not route it.
How well it went
Measured, not estimated. This needs execution data we do not have yet.
Outcome Analytics
Speed, quality and unit economics of work done by people and agents together, connected to what was actually delivered. Coverage estimates hours; this measures them.
Development Loops
Development plans for people ship today. The same loop for agents, evaluation, tuning and re-deployment, comes with execution.
People and agents are described in one vocabulary: seventeen capabilities derived from 39 published processes, each with a dated verdict on what a model can do today.
Governance by design.
Controls belong in from the start. Some hold at the planning stage, before any agent runs; the rest arrive with execution.
Access per process
A process is seen by the whole company or only by the roles, positions and people you name. The owner edits it; admins and HR keep the keys.
History
Changes to the owner, the visibility and the rules stay in the process's history.
Evidence behind every verdict
Each capability chip shows the model's confidence and the fragment of your description it relied on. A doubtful chip is a suggestion until you confirm it.
Registered agents on record
Each agent you register carries its vendor, data classification, security review and EU AI Act risk level.
Policy engine
Routing rules, escalation paths and approval gates defined per process and enforced when the work runs.
Runtime trace
A full trace of human and agent actions: who did what, with which model, under which policy, reviewed by whom.
The numbers, in the order they arrive.
No number on this page says what share of your work an agent can take. The product answers that per step, for your process, with your own hours.
Per process
SHIPPED · v2.5.3An estimate you correct, not a measurement.
- Hours a year behind each step, and what they cost at your own rate
- The share of the work that moves to an agent, moves to review, or stays with people
- The steps nobody covers yet
Across companies
LATERA direction, not a commitment. Only capability codes are compared, never a company's text.
- How much of a comparable process others already hand to agents
- Which skills work for them
With execution
LATERNo invented benchmarks. We publish real numbers as teams produce them.
- Time-to-delivery
- Cost-per-task
- Quality pass rate
- Work actually handed to agents
A level that can be checked, and that travels.
The scale and the five-question self-assessment are public today. Everything below is a plan. The benchmarks and the verification API will be Cloud only: no single self-hosted instance can build a network on its own.
AI Fluency level, in the product
Built on the evidence the assessment modules already collect. The evidence behind a level stays visible to the person it describes.
Industry benchmarks
Your team's Basic / Middle / Advanced distribution next to your industry, and a public page of industries by AI maturity. A number is published only once 20 companies in that industry take part.
Public verification API
A third party checks a stated level against a signed API, so a level can be verified rather than trusted.
A level that travels with the person
A level earned at one employer goes with the person to the next: signed, and owned by the person it describes.
See what the vision grows from.
The best way to judge a direction is to look at what already ships. Open the demo, or talk to the founder about where this is going.