01 · Analyst
Draft with permitted sources
An analyst drafts a proposal in a shared Session using the team's scoped sources.
Early use · founder-network pilots
Built and used in three free pilots. Now recruiting paid design partners.
LootMogul · SuperAthlete.ai · Ahi.vc
The team reports successful free pilots and initial partner usage. Paid deployments are the next step.
Collaborate → Review & approve → Reuse private knowledge → Reduce repeated AI work
Approved answers are reused only within their permitted scope, with freshness checks and human review.
First use case · consulting and advisory
From proposal review to reusable team knowledge
An AI draft still needs source checks, expert review, corrections and a handoff. The next engagement should start from what the team has already approved.
02 · Reviewer
Correct, review and accept
A subject-matter reviewer checks sources, corrects the work and accepts the result in the same Session.
03 · Next engagement
Reuse with human judgment
Approved scoped knowledge can inform a delivery handoff or the next proposal—after permission, freshness and human-review checks.
Illustrative workflow using shared Sessions and approved knowledge. Your team controls the handoff, reviews each reuse, and owns the resulting knowledge.
When Qua fits
- Repeated proposals and recurring client work
- Three or more contributors and reviewers
- A reusable, permitted source library
- Review and rework burden you can measure
Start with one team. Prove the workflow.
One 20–50-person consulting/advisory team. A six-week scoped pilot: agree the baseline, accepted quality and annual decision criteria with the practice lead or COO. Expand into recurring delivery work and adjacent teams only after retained use.
The three pillars
Three pillars. One collaborative workspace.
Multi-player
Shared work, accountable decisions
Contributors and reviewers share the task, check sources and correct or accept work in a Session. Decisions stay with the work.
Details →Multi-Layer
Approved knowledge before new generation
Consult scoped personal and approved company knowledge before new generation. Recheck permissions and freshness; human review remains essential.
Details →Multi-Model
Permitted models, one team workflow
Use permitted models while keeping team context and review together. Policy precedes routing; Open Models are an explicit choice, never automatic.
Details →Y Combinator’s Requests for Startups names multi-player AI; a16z’s Big Ideas names the enterprise orchestration layer. These are public thesis alignments, not evidence of buyer demand or endorsement.Read the thesis →
U · UNIFY — MULTI-PLAYER AI
One room. Your whole team, and the AI, on the same task.
A reviewer can check the analyst’s sources, correct the draft and accept work in the same Session. Shared task context supports a human-led handoff.
Q4 campaign read-out
qua.dev/s/a7k9d2- $0.0000
Draft the Q4 campaign read-out for the board — use our own performance data first.
- $0.0000
Answered from 4 approved sources in your workspace. No data left the perimeter.
- $0.0000
Use net revenue, not gross — finance restated October.
- $0.0031
Re-ran the summary on the restated figures and kept the citations attached.
- $0.0000
Accepted. Saved to team knowledge — the next person asking pays nothing.
One shared session. Three colleagues and the AI on the same task — a correction lands mid-run, a reviewer accepts the answer into team knowledge, and the running cost stays visible to everyone.
Shared
One link, one source of truth
Everyone sees the same task, the same sources, and the same running cost.
Supervised
Correct it while it runs
Step in mid-task, approve a step, or hand the work to a colleague without losing context.
Accountable
Who decided what, on the record
Shared Session history keeps review context together; admin analytics focus on outcomes, not prompt content.
What the team needs beyond a draft.
1A🛡
How do I control what data AI can see?
Define exactly what data and systems AI and agents can access—and what stays private.
Data access control
Public web
Internal docs
Customer data
Finance systems
2Q$
What does accepted work really cost?
Evaluate preparation, review, rework and implementation alongside AI costs.
Buyer-agreed baseline
Total
cost per accepted deliverable; rejected attempts included
3Q📖
How do I reuse what my company already knows?
Consult permitted, approved knowledge first; reuse still needs freshness and human review.
Private knowledge
Answer found in 2 sources
Est. cost$0.0021
4U👥
How do I let teams work together with AI?
Agent Sessions turn AI into a shared workspace—see it, edit it, approve it, together.
KKarthik edited2m ago
NNatasha added input3m ago
AAlex approved5m ago
5A⚙
How do I keep control no matter which models I use?
Use permitted models under policy. Managed cloud today; other deployment modes are scoped.
OpenAI
Anthropic
Gemini
Kimi
Your model
A · AMPLIFY — HOW QUA DECIDES
How Qua works: The Answer Waterfall
Qua starts with knowledge you already own, then climbs only as far as the answer needs. Open Models remain an explicit choice, never an automatic route.
01
Only if it is worth it.
The highest-capability models, for complex reasoning and premium tasks — reached only when the expected-value gate clears.
Pro Models · premium02
Can a cheaper model do it?
Permitted models for routine work, with visible cost. Accepted-work quality and total workflow economics still need buyer evaluation.
Fast Models · metered03
Do you want to run it on an open model?
Open-weight and open-source models, available whenever you ask for one. Qua never routes here on its own.
By request only · never auto-selected04
Does your company already know?
Searches connected company data and approved workplace sources — including org-verified answers — then a cited answer.
In-perimeter · ~$005
Do you already know this?
Retrieves what you saved or connected — your verified answers, scoped knowledge, and instant deterministic resolution.
$0 · instantPolicy first. Personal Knowledge → Enterprise Search → Fast Models → Pro Models. Open Models by explicit choice.
A · AMPLIFY — MULTI-LAYER IN MOTION
Knowledge first. New generation when needed.
Your knowledge first, then approved company sources, then a fast model. A frontier model is reached only when the value gate clears — and you see the price either way.
The question“What did we commit to on data residency in the Meridian contract?”
- Personal KnowledgeNothing saved covers this yet$0
- Enterprise SearchFour approved sources match — cited answer composed$0
- Fast ModelsUsed to tighten the wording on retrieved facts$0.0031
- Pro ModelsValue gate did not clear — not usednot billed
Resolved for $0.0031 · frontier-only would have cost $0.0129 · sources stayed in your perimeter.
Model costs are supporting economics
Below is a fully modeled 30-day scenario for a 380-seat enterprise — our Meridian Group demo dataset. The same 286,400 queries are priced three ways: what they cost through Qua's Answer Waterfall, what an all-open-source setup would cost, and what routing everything to a premium model would cost. Every figure on this page reconciles — totals, tiers, and daily series all add up.
89.4%
below all-premium routing
84%
of queries resolved without reaching Pro Models
$1.65M
modeled 30-day savings vs. frontier
$0.96
blended cost per 1M tokens
The daily cost gap
Qua modeled — $194.8K
All open-source — $561.8K
All frontier — $1.84M
Where modeled queries resolve
Pro Models · 16% · 45,824 queries
Fast Models · 22% · 63,008 queries
Open Models · by request — 0 in this scenario
Enterprise Search · 30% · 85,920 queries
Personal Knowledge · 32% · 91,648 queries
84% of work never reaches the most expensive tier.
Model mix in this scenario
Frontier models process 6% of the tokens but half the bill — routing is one input, not the full accepted-work cost.
Illustrative modeled scenario (Meridian Group demo dataset), not customer results. The open-source and all-premium figures are modeled counterfactuals at list rates; the Qua figure is also modeled, not an actual customer invoice. Labor-value figures, where shown, use a blended $85/hr rate.
Why choose Qua · your pilot scorecard
Measure the value of accepted work.
The buying test is lower total time and cost to an accepted deliverable, with quality maintained. Compare Qua against the current stack before deciding.
Agree what good meansUse the same permitted source set and accepted-quality standard. Include preparation, expert review, rework and rejected attempts.
Count the whole workflowInclude implementation, retrieval, infrastructure and model costs. Saved hours are capacity value—not automatically cash savings.
Earn recurring useAgree adoption and annual decision criteria upfront. Expand only after retained use; there is no guaranteed savings result.
Six-week scoped comparison
Start with one team. Prove the workflow.
Baseline → weekly review → readout
Practice lead/COO, named reviewers and permitted data. Agree evaluation volume, quality, adoption and the annual decision before kickoff. Compare completed work with your existing workflow.
Reuse that earns its place
Test whether approved proposal knowledge reduces review and rework in the next proposal or delivery handoff. Your corpus stays yours and exportable; no lock-in or exclusive-feature claim.
For investors and portfolio teams
Connect the right team with a paid pilot.
Introduce a portfolio company with a recurring team workflow and approved knowledge sources. Scope a paid pilot with a named owner, clear data permissions and agreed outcomes.
Q · QUANTIFY — YOUR MODELS, YOUR WALLS
Managed cloud today. Other boundaries scoped with you.
Qua runs in managed cloud today. Customer VPC, on-premises and air-gapped deployments require scoped feasibility, security and implementation review; they are not generally available deployment options.
Private knowledge first. External model access only under enterprise policy. Scoped sources and approved knowledge govern reuse; policy can permit data to reach external models for inference. Retrieval and infrastructure still have costs.
Neutrality is the business model: Qua never trains shared models on your knowledge, and admins see outcomes — never prompt content.
Qua cloud
Your VPC · scoped
On-prem / air-gapped · scoped
Availability and assurance requirements are agreed per deployment.
Don’t rent your answers. Own them.
Quantify · Unify · Amplify
Qua turns your enterprise’s private knowledge into shared AI intelligence—so teams can work together, reuse approved answers, and spend less on AI.