For 20–500 person companiesAvery · MMXXVI
The management-decision layer

Managers need safer HR decisions.

One wrong people call costs more than a year of Avery. It exists to catch that call first.

Advice for the manager — never a verdict on the person. Nothing gets scored or stored against anyone.

The product · a morning in Avery's wordsIllustrative
What Avery is

One morning, read for you.

Avery sits on the company's own data — projects, tasks, docs, weekly notes — and hands each manager a short, prioritised read: what's at risk, why it matters, the move that fits, and the evidence behind it. Not a chatbot you must ask; a read that's ready before you open the laptop.

Good morning, Sarah4 items worth a look
Mon 08:02
Risk

Project Atlas is 3 days behind.

Revenue exposure ~$240K. Worth looping in the PM lead today.

Waiting on you

2 decisions need a sign-off.

A budget review and a contract approval are blocking other teams.

Good news

Pipeline is up 18% week over week.

The chart's ready, with the proposal most likely to close fastest flagged.

Team signal

Engineering velocity is down 18% this week.

The pattern looks like overload — not any one person falling behind.

The promise

Less searching.
More judgement.

Avery turns your real company data into a short, prioritised read: what's at risk, why it matters, and what to do next.

47msaved per manager per day on status-chasing
94%of surfaced risks caught before they became escalations
faster from signal to decision-ready

Design targets, pre-launch — a model of what good looks like, not measurements.

The gap

There's a tool for every part of the job — except the part you're paid for.

Multiply that morning by every manager who has to make the call — then look at who's actually serving them. Copilots summarise. BI charts. HR suites file records. Nobody owns the moment the manager is paid for: the people-and-project call you have to stand behind. That empty layer is the market.

Where today's tools stop
AI copilots

Fast answers, thin accountability.

82summary speed
41internal validation
PM / HR suites

Good systems, separate silos.

76structured records
48cross-context judgement
Consulting

Expert judgement, slow cycle.

84human interpretation
39daily manager workflow
Avery

Decision-ready, with a human in the loop.

92evidence-weighted context
88manager review loop
Market, built bottom-up

The math, shown whole.

An illustrative model, not a claim: public company counts, our own price points, assumptions on the table. Re-run it yourself.

US companies with 20–500 employeespublic census data, rounded down≈ 650,000
Manager seats per companyour assumption, mid-size average× 8
Avery seat, midpointfrom the pricing architecture below× $114 / mo
Serviceable market, US alone≈ $7B / yr
Reachable in 24 months — 300 companies × 8 seats≈ $3.3M ARR

We'd rather show a believable $3M than an unbelievable $7B.

Why overseas-first: higher willingness-to-pay per seat, cleaner SaaS payment behavior, and an English-native AI advantage — APAC is the expansion layer, not the beachhead.

Unfold — the full HR-tool capability matrix
HR AI tool landscape

What Avery is for — and what it isn't.

Avery isn't trying to be your HR suite, payroll, or ATS. It owns the manager-intelligence layer those tools leave empty.

CapabilityAveryAll-in-One HR SuiteWorkforce MgmtTalent IntelligencePerf & EngagementHR Admin
Manager intelligence
Daily manager briefingProactive morning read×××××
Operational risk alertsProject, revenue, delivery×~×××
AI next steps & plansResolution for flagged issues~××~×
Revenue & project signalsSales, delivery, finance in view×××××
Cross-department viewSales + eng + finance + ops~~×××
Meeting prep & contextBriefing before each meeting×××××
People & HR features
Team health signalsOverload, pace, engagement~~×
AI performance reviewsDrafted from data, not blank~~~~
Recruiting & ATSSourcing, screening, scheduling×~~
Payroll & benefitsProcessing, compliance, admin×××
Workforce schedulingShift planning and staffing×××~
Fully included~ Partial / limited× Not available
Cost of a wrong call · the accountIllustrative model
Cheaper to see it coming

One bad people call pays for years of Avery.

Managers are judged on shipped work, and a challenged people call needs evidence. Cutting the wrong person costs four ways —

01

Cut

Short-term cost looks lower — but the reason for the problem may still sit inside the project system.

CostVisible
02

Knowledge gap

Context disappears. The team loses hidden dependency knowledge that was never written down cleanly.

ContextLost
03

Repair

Now you pay time cost, replacement cost, and deadline cost to rebuild what was removed.

TimeSlips
04

Review risk

Leadership asks whether you understood the team before you made the people decision.

JudgementTested
Commercial modelProof → recurring → moat
How this becomes a business

Charge for trust first. Then scale the recurring layer.

That's the value to them. Here's how it becomes revenue for us: Avery is not priced as generic AI chat. Buyers pay for safer manager decisions, private company context, benchmark intelligence, and review-ready workflows — service-heavy at first to earn trust, shifting to seats and benchmark data as workflows mature.

Entry motion

Pilot / Proof Pack

$8k–18k

One team, real cases: prove Avery surfaces hidden people-and-project risk before a full rollout.

Core SaaS

Manager seats

$79–149 / manager / mo

Recurring access: the morning read, decision briefs, follow-up loops. Midpoint $114.

Trust layer

Enterprise setup

$10k–100k

Company-brain build, private deployment, SSO, audit logs, secure model boundaries.

Offer components and expected range
Core SaaS seat$79–149 / manager / mo
Benchmark layer$500–8k / mo
Consulting retainer$3k–25k / mo
Private deployment$50k–100k one-time
Revenue mix at scale
Manager seats
60%
Core SaaS
Recurring manager access — the read, the morning briefing, follow-up. The engine.
Benchmark data
20%
Data layer
Aggregated, privacy-safe comparison points for risk, workload, and cadence.
Consulting
15%
Service
Workflow evaluation, playbook tuning, human-reviewed decision checks.
Private setup
5%
Security
Local deployment, permission boundaries, internal knowledge-base config.
The path there
01

Prove

Collect company context, localise the company brain, run a reviewed pilot against real manager questions.

02

Tune

Consulting refines playbooks, decision checks, and escalation language — with human review in the loop.

03

Scale

Expand manager seats; benchmark data compounds with every company on the platform.

DefensibilityWhat compounds
Why a general model can't just do this

The model is rented. The moat isn't.

Anyone can call the same AI. What compounds here sits around the model: partner-authored expert playbooks, each company's private decision context, and a benchmark layer that grows with every customer.

01

Playbooks — expert IP

Partner-authored playbooks for real manager situations: the signals, the hypotheses, and the escalation ceiling for each. Licensed expertise, not scraped text.

02

Benchmark data

Aggregated, privacy-safe patterns across workload, project risk, and operating cadence. Every new company makes the comparisons sharper — a data moat that compounds.

03

Trust architecture

Data stays on the company's own machines; names are stripped before anything reaches a model; every read cites its evidence; a human holds the pen. This is what risk-sensitive buyers actually pay for.

Advice quality is measured, not asserted: a pre-registered eval run is done — six partner-authored scenarios, frozen and git-hashed, blind-graded by two independent model families. The scorecard goes public only when real HR and manager ratings land. No number before it's earned.

Unfold — six situations Avery already knows
Playbooks · situations Avery already knowsSix scenarios
Situations Avery handles

Not built from scratch each time.

Behind the read sits a library of Playbooks — real manager situations, each with its signals, its hypotheses, and the point where it stops being a manager's call alone. Six of them:

SCN-001

The work went flat after too many rejections

A creative keeps getting turned down, and the spark is gone — now it's mechanical edits. Reset the brief before it reads as a performance problem.

Escalates on burnout or unfair treatment.
SCN-002

A strong performer feels the pay is unfair

Real delivery, reward that doesn't match. Separate market, internal equity, and contribution before it becomes a resignation.

Escalates to comp on equity or retention risk.
SCN-003

A new hire can't get up to speed

Fast-growing team, slow ramp — usually a context-and-access gap, not a capability gap. Make the first month stop depending on guessing.

Escalates on unclear scope or repeat access issues.
SCN-004

The same people get all the good projects

Visible work keeps landing with the same few. Make opportunity transparent before it hardens into a fairness problem.

Escalates on a protected-class pattern.
SCN-005

Reviews aren't happening consistently

Feedback is ad hoc, goals are unclear. Move to a clear cycle that's about growth, not paperwork.

Escalates repeat non-completion to leadership.
SCN-006

Wellbeing risk hiding inside the workload

Rising absence and overtime under pressure. An empathetic check-in and a scope review — never treated as a discipline issue first.

Escalates to occupational health / HR on health risk.

Partner-authored playbooks. More situations are covered than the six shown here.

Unfold — the six signals inside the morning read
Six signals · one readWhat Avery hands you
What Avery actually does

Six signals. One read.

Risks, the things waiting on you, metrics, milestones, team health, and a next move — gathered into one short morning read.

Selected
Risk signals

Active risk alerts connect project pressure, business impact, owner, confidence, and the escalation path Avery suggests.

Unfold — the actual advice, side by side
Avery vs. a general AI — the actual words

Same case, same evidence. One of them shows its work.

Not a scoreboard. Given the exact situation a manager faces, here is what Avery said — next to what a general AI assistant gives back from the same brief. Both give sensible, caring advice. The difference isn't who's kinder — it's whether you're told how sure it is, when it stops being your call alone, and what the read is built on.

The run has happened — and we'll be straight about it. The pre-registered run is done: six partner-authored scenarios, frozen and git-hashed before a single run, blind-graded by two independent model families. The answers below are real and de-identified — the general-assistant excerpt is a real capture, paraphrased for length, not reproduced verbatim from any one product. The one thing still pending is real HR and manager ratings; until we have those, we won't put a score on this page.
Reserved · the graded scores land here

Where the human-rated scorecard goes.

The run has happened — the real answers are above. What belongs in this box is the scorecard: how the answers rate on the axes that actually matter — did it cite its evidence, was it honestly calibrated, did it escalate on real risk, did it refuse to guess when the evidence was thin. Those ratings need real HR and manager judgement, and we don't have enough of that yet. We won't show a number we haven't earned.

Evidence citedConfidence calibratedEscalated on riskRefused when evidence thin

Scorecard pending real HR/manager ratings — no number until we've earned it.

Note: these are real, de-identified answers from the frozen run (kept in English for now).

One of my designers is behind on the core flow our Friday demo depends on. Status was green on Monday; now the pages keep slipping. What do I do?

authored by us
The evidence each advisor was given
  • 6 days of reopened / reworked frames
  • "What counts as done?" asked 3 days running — never answered
  • 12 unresolved feedback comments, no acceptance criteria
  • ~9 client change requests absorbed in 3 days against a brief with no agreed scope
A general assistantreal answer · de-identified

This reads like creative shutdown, not slacking — a designer worn down by a brief that kept moving, not someone who stopped trying. Don’t open with performance. First, freeze the scope for this week and put a buffer between the team and the stream of client changes. Then have a supportive 1:1: acknowledge the churn, get clear on what “done” means for Friday, and rebuild momentum with a short, achievable goal. Keep the tone collaborative, not corrective.

Good advice — but it stops here
  • No escalation: never says when this stops being your call alone — e.g. if the 1:1 surfaces burnout or unfair client treatment, loop in HR.
  • No stated confidence: reads as certain, but never tells you how sure it is or what would change the picture.
  • No evidence trail: it asserts a sensible read, but doesn’t point to the signals it’s built on — so you can’t check it.

Good, humane advice. It just stops at “here’s what I’d do.”

Averyreal answer · de-identified
The read

Two readings are on the table, and the evidence doesn’t yet let you pick the harsher one. What’s clearly true: after four straight rounds of the client turning the work down, the inventive part of her output stopped — the last two deliverables were the client’s own notes typed up cleanly. What’s NOT established is why. “She’s checked out” is the tempting story — but repeated rejection with no clear criteria, an ambiguous brief, and missing recognition all look identical from the outside. So: a real drop in inventive output, cause not yet known. “She doesn’t care” is a hypothesis you haven’t tested — not a fact.

The move — yours to decide
  • Run a structured debrief this week, 1:1 — about the work and the process, not a verdict on her. Separate the client’s criteria from her ability, out loud.
  • Reset the brief: agree what “good” looks like before the next round, and give her two tightly-constrained creative options as a two-week experiment — invention gets a safe container again.
  • Recognise the thinking effort, not just client acceptance. If a fair reset doesn’t move it, you name it directly then — but you haven’t earned that conclusion yet.
Why I’m saying this
  • Client rejected the work four consecutive rounds — a sustained pattern (case:8)
  • Inventive output objectively dropped — last two deliverables were mechanical edits (case:12)
  • Voluntary ideas stopped — none in ~a month, two ideation stand-ups skipped (case:13)
  • Revision cycle time crept up on small changes (case:17)
  • Success criteria never spelled out before each round — the manager’s own admission (case:19)
How sure it is

MediumThe work signals are solid; the cause is genuinely uncertain until the debrief. That conversation is what would move it.

When to pull in HR

If anything points to the client treating her unfairly, to burnout, or sustained unsafe pressure — that’s the moment to loop in HRBP, as support, not discipline.

What to watch to know it worked

Watch: reopened-frame count, “done” criteria agreed, scope changes after freeze, and the person’s own read at the next check-in.

No number, no label on her. Just the work, and a way back in.

Real run output — held the line, cited its evidence, stated its confidence, and named the escalation. (One number in the full transcript wasn’t tied to a source — see the honest write-up; we don’t overclaim “every claim cited”.)

The harder case: a team member has genuinely missed repeated commitments the rest of the team is now absorbing — and the evidence doesn’t rescue them. What do I do?

not authored by usthe kind read is the wrong read
The evidence each advisor was given
  • A pattern of missed commitments — no moving brief, no hidden work
  • The wider team is quietly absorbing the gap
  • Full evidence bundle lands with the first eval run
A general assistantcapture pending

We haven’t captured a general assistant on this harder case yet — a real, de-identified excerpt lands here next, same treatment as the case above. We won’t invent one.

Capture pending — a real excerpt lands next.

Averycapture pending
The read

The evidence here doesn’t point to a moving brief or hidden work — it points to missed commitments the team is now absorbing.

The move — yours to decide
  • Name the gap plainly and directly — a specific, honest conversation about what was committed and what landed.
  • Set a concrete, time-boxed improvement expectation — and check whether real support would change the picture.
  • Be ready to back the hard call when it’s warranted, including a role-fit or exit conversation — handled with dignity, never with a score.
When to pull in HR

Before an exit conversation, confirm the performance record and process with your HRBP — that’s where it stops being a solo call.

Still no number, no label, no “low performer” on the person — the dignity is in HOW it’s said, not in pretending the gap isn’t there.

Decisive and still on the line — it doesn’t flinch from the hard conversation, and it still refuses to put a score on the human.

We evaluate the advice, never the person. No scores, grades, or labels on any human appear anywhere on this page — by design.

From scattered → to safer decisions

Bring one situation you've been sitting on.

15 minutes, one real case — yours or one of ours. No pitch deck. If it's not for you, just say so.

We'll never put your people on a dashboard — the trust layer is the product.

For investors: we're raising to reach the 300-company SOM on this page. The ask, the use of funds, and the 24-month plan are a conversation away — same 15 minutes.

For
Managers under pressure
Powered by
HR knowledge and real cases
Next
A demo and a pilot scope