Applying science
to the art of sales.
A shared, universal standard for what good, bad and great commercial craft looks like: an ontology, a rubric, a methodology and an evidence base that every Peer product, and every AI system that judges commercial skill, can be built on.
Every serious profession has a science underneath it.
Medicine has it. Engineering has it. Even sport has it.
There is no shared, universal standard for what good, bad, or great sales craft looks like.
As a result, we cannot do some of the most fundamental things the profession needs: coach effectively, train consistently, provide clear career development trajectories, or reliably improve revenue performance. The data bears that out across every market and every industry.
And now that same gap has a new consequence. We cannot train AI to sell well, to judge sales craft accurately, or to provide actual, useful, consistent coaching and feedback. The standard does not exist for humans. It does not exist for machines either.
The sales industry is full of good intentions and genuine effort. Coaches coach. Managers mentor. Consultancy firms build frameworks. Enablement platforms train on activity. Conversational intelligence tools analyse what was said on the call. Methodologies and scorecards provide structure. All of it helps. None of it solves the actual problem.
Because every one of these people and every one of these tools operates without a shared, universal definition of the foundational craft skills that determine whether a seller is bad, good, or great. They are all improving the surface without ever agreeing on what sits underneath it.
To apply science to the art of sales.
To use data-led, methodical, and academically rigorous approaches to build the first shared, universal standard and toolset for sales craft, giving every organisation and every AI system a common definition of what good, bad, and great actually looks like.
The standard is four pillars, built together.
A standard isn\'t a manifesto; it\'s a working system. Four pillars, each meaningful on its own, each checked against the others. Remove any one and the standard collapses into opinion.
The map of the craft.
A taxonomy of the skills that decide whether someone is bad, good or great at commercial craft: five functions, hundreds of named competencies, organised into families, sub-skills and dependencies. A working map of the craft itself.
- 5 functionsSales · Customer Success · Support · Management · Culture
- 424 competenciesNamed, defined, dependency-aware
- Function- and role-awareThe same skill behaves differently in different seats
What good looks like, observed.
For every competency, an observable definition of what each level of mastery looks like. Four tiers, expert-authored, with worked criteria a trained human, or a machine, can apply consistently.
- Four-tier ladderFoundation → Proficient → Advanced → Expert
- Observable criteriaBehavioural, not abstract
- Worked examplesFor every tier of every competency
From conversation to defensible score.
The system that turns a real conversation into a defensible score. Specialist assessors, each owning a sub-domain, each grounding observations in cited evidence from the transcript. Every score traces back to the words that produced it.
- Specialist assessorsOne per commercial function, trained separately
- Evidence-groundedEvery observation cites the quote it came from
- 95.5%Within one rung of a trained assessor, on the same moment
Grounded in real human work.
Grounded in real human-graded conversations, not synthetic test sets. Years of expert work sit under every tier definition and rubric line. New competencies enter only after evidence and review.
- Real conversationsSourced from real commercial work, not LLM hypotheticals
- Expert reviewTier definitions authored and reviewed by domain experts
- Living standardVersioned, reviewed, governed: Peer v2 today
Started with sales. Extends across the commercial function.
We started with sales because the gap was widest and the cost of getting it wrong highest. The same standard now extends to every adjacent function where commercial outcomes depend on craft: each built on its own, tied back to one shared ontology.
See the standard at work.
The pillars are the summary. The Atlas takes you inside: the assessment pipeline, the progression model, the dependency graph, the evidence base and the certification pathway. No client data. No rubric content. The system, end to end.
The Sales Mastery System
A guided tour of how Peer measures, scores, and progresses sales skill mastery: a roster of specialist agents, four assessment tiers, a dependency graph, the certification ladder, and the evidence base behind it.
Customer Success · Support · Management · Culture
Each function's Atlas follows the same architecture, tied back to the shared ontology. Published as each reaches review.
Every Peer product is built on the same standard.
The standard is the moat. Every product PeerLab ships is a different consumer of the same underlying ontology, rubric, methodology, and evidence base. That is why they compose with one another, and why a score in one place means the same thing in another.
Applies the full standard to a customer's own commercial conversations. Ontology to map competencies, rubric to grade them, methodology to produce defensible scores, provenance to back the calibration. Mastery grounded in the same definition across every seat.
Anchors a person's career-long portfolio in the standard, not in role titles or employer-specific systems. Mastery you can carry between jobs, signed by the same calibration every assessor uses.
Sanitised, licensable corpus of behavioural primitives, rubric-graded against the standard's tier criteria, sanitised and versioned. Training data and evaluation rubric for the next generation of commercial AI.
A living standard.
- Peer v2 today. 424 competencies across 5 functions, four-tier mastery ladder, expert-authored rubric content, calibrated against human raters.
- Evidence before publication. New competencies enter the standard only after the supporting work is graded, reviewed, and held to the same calibration as the rest of the ontology.
- Versioned and dated. Every score in every Peer product is anchored to the standard version it was assessed against. Changes to the rubric don't silently revise the history.
- Open about the boundary. The methodology is public. The rubric content, the IP that distinguishes the standard from any other framework, sits inside the engine. See where the line is drawn →