Delivery Intelligence: Human, AI-assisted and agent attribution

When an agent opens the PR, who delivered it?

An engineer scopes the task. An agent writes the code. A bot reviews it. It merges. Every delivery metric you report just became unreadable. Pensero credits orchestrated delivery to the person who invoked the agent, splits it into human, AI-assisted and agentic, and weights every unit by magnitude and complexity.

Book a demo

Measuring the agentic workforce at fast-growing companies around the world

Know what's actually shipping

Magnitude & complexity scoring

The work item "Add workspace setup to self-serve onboarding" with its activities and delivery points. Its Pensero score is 10.0: magnitude Large • 5.0, complexity Medium • 2.0.

Filter noise, measure value

Volume is not delivery. Pensero scores what shipped by magnitude and complexity, with boilerplate, repetitive changes and auto-generated noise excluded. A high-volume agent doing trivial work doesn't outrank the engineer who shipped the hard thing.

Magnitude and complexity scoring
Business value measurement
Boilerplate and auto-generated noise filtering

Scope of work breakdown

Pensero roadmap alignment chart with example data: 78% of delivery aligned to roadmap, +6pp vs the previous period, against a weekly average of 73.0% and an industry median of 57%. Weekly bars show the aligned and not aligned share of delivery over 9 months. Top work aligned, in delivery points: Self-serve onboarding 842.6, Payments platform v2 731.2, Mobile offline mode 518.9, Usage-based billing 402.4, SSO for enterprise 287.5.

See where effort actually goes

Break down work across features, tech debt, bugs, maintenance and AI initiatives, and see how much of each was delivered by humans, assistants or agents. Know whether capacity is going where you said it would.

Work categorization
Human and agent capacity allocation
Strategic initiative tracking

Native attribution

Weekly delivery points per person: 12.1 pts this period, 9.6 pts on average each week.

Credit the person who directed the agent

Work an engineer directs through an agent counts as their delivery. Defects introduced by that agent count too. Credit and accountability stay attached to the same person, and the work that enables delivery is visible, not just the code that gets written.

Orchestrated delivery credit
Individual contribution visibility
Human, AI-assisted and agent attribution

Strategic alignment tracking

Delivery

Where the team shipped this quarter

5 epics · 2,782.6 pts shipped

Epic	Contributors	Delivery
Self-serve onboarding	ZNLC+4	842.6
Payments platform v2	ATNK+3	731.2
Mobile offline mode	SM+1	518.9
Usage-based billing	EBMP	402.4
SSO for enterprise	NKLS	287.5

Delivery points by epic: Self-serve onboarding 842.6, Payments platform v2 731.2, Mobile offline mode 518.9, Usage-based billing 402.4, SSO for enterprise 287.5.

Build what matters most

Connect work to company objectives, whoever or whatever delivered it. See whether human and agent effort is focused on the projects that drive real business impact.

Initiative alignment
Delivery by objective
Strategic progress visibility

"Efficiency multiplier for my engineering organization."

Nick TraylorHead of Engineering, Stord

Read story →

"Pensero is my compass in order to know where the team is heading to."

Jordi MiróCTO, Zynap

Read story →

Questions engineering leaders ask

How is this different from DORA or flow metrics?

Pensero includes workflow visibility and DORA-relevant context, but extends beyond flow by scoring the complexity and magnitude of what was actually delivered, not just how fast work moved through the pipeline. It measures the value of output, not only its pace.

Does this require my team to change how they work?

No. Pensero connects your existing repositories, tickets, and communication sources and normalizes them into comparable views, with no surveys and no manual reporting overhead.

Can Pensero measure contribution that isn't code?

Yes. Pensero pulls in tickets, messages, documents, and conversations alongside code, so you can see contribution across roles and collaboration patterns, not just PRs or commits.

How quickly do we see useful signal?

Most teams connect core tools in about an hour, see early insights within a day, and reach board-ready reporting in about a week, with historical data available from day one.