Every traditional way of buying software delivery is broken in a different way. Time-and-materials makes the vendor profit when work takes longer. Fixed-bid contracts collapse the moment scope shifts. Per-FTE subscriptions punish customers for the redundancy that AiDOOS's operating model requires. Milestone billing works in theory but turns every scope conversation into a renegotiation. Buyers know all this — and yet vendors keep selling these same broken models because the alternative requires actually pricing the output, not the input.
AiDOOS prices delivery in Delivery Units (DUs) — a universal output-based currency that decouples what the customer pays from how long talent takes, how senior they are, or what stack they use. This piece explains what a DU is, why traditional models fail, how the four AiDOOS tiers work, and why the trust mechanisms wrapped around DU pricing remove buyer risk in ways that hourly contracts structurally cannot.
Why traditional pricing models keep breaking
Software delivery has tried four major pricing patterns. Each fails in a predictable way:
Time and Materials (T&M)
The customer pays for hours worked, regardless of what gets shipped. The vendor's incentive is to expand hours; the customer's job is to police timesheets. The economic model rewards delay. T&M is the dominant model in staff augmentation and Big-IT consulting precisely because it puts all delivery risk on the customer while preserving the vendor's margin.
Fixed-Bid SOWs
The customer pays a single price for a defined scope. Sounds clean — until specifications shift, which they always do, and the change-order procedure turns into a contracting cycle that costs more than the change itself. Fixed-bid only works when scope is genuinely stable, which in modern software means rarely.
Per-FTE Subscriptions
The customer rents a named engineer for a monthly fee. Common in modern remote-work platforms (Toptal monthly contracts, Deel EOR, Turing). The unit economics break the moment the operator wants to overstaff for redundancy — buyers refuse to pay for "6 people when we need 4," even though running pods understaffed is the most common cause of delivery failure.
Milestone Billing
The customer pays per shipped milestone. Closer to outcome alignment — but the unit (a milestone) is engagement-specific, so every new project requires re-negotiating what counts as a milestone, what acceptance looks like, and what a milestone is worth. Comparison across vendors becomes impossible. Procurement teams revert to T&M because at least hours are comparable.
Each model has a different broken assumption. T&M prices effort instead of output. Fixed-bid assumes scope stability that doesn't exist. Per-FTE prices people instead of work. Milestone billing prices work but the unit isn't standardized.
The fix isn't a fifth pricing model. It's a different primitive entirely.
What a Delivery Unit (DU) is
A Delivery Unit is a standardized measure of cognitive output, calibrated against a reference catalog of typical software work. Roughly: 1 DU equals about 1/2 day to 1 day for a mid-level engineering output, but the hours are a benchmark, not a billing unit. A senior architect who delivers 1 DU of work in 1 hour and a junior engineer who delivers 1 DU of work in 8 hours both produce 1 DU. The customer pays the same.
The DU primitive does three things at once:
- Decouples price from time. Customer doesn't care whether a senior or a junior shipped the feature. They paid for the feature.
- Decouples price from technology. A 5-DU story is a 5-DU story whether it's React, SAP ABAP, or a SQL migration. AiDOOS takes care of skill, seniority, scarcity, and complexity multipliers in the DU calculation. Customer sees one universal rate.
- Future-proofs the unit for AI delivery. When an human + agent delivers 1 DU, it gets paid the same as a human delivering 1 DU. The verification framework adjudicates quality regardless of who or what produced the output.
This is the same pattern that AWS uses (you pay for compute, not for which physical server runs it), Uber uses (you pay for the trip, not the driver's salary), and airlines use (you pay for the seat, not the airframe age). The customer-facing primitive stays simple while the internal economics stay sophisticated.
The AiDOOS pricing tiers
Four pre-purchased credit packs plus a Project-flow entry point and an Enterprise band. Same dollar-per-DU rate across both flows at any given DU count — there is no "on-demand premium."
| Tier | DUs | Price | $/DU | Validity | Best fit |
|---|---|---|---|---|---|
| Project | Variable (sized by Instant Proposal) | Tier-rate × DU count | Custom | Per engagement | Funnel entry — instant proposal sizes the work, customer pays the tier rate that band falls into |
| Starter | 10 | $2,000 | $200 | 90 days | The $2K bet — bypasses procurement, credit-card checkout |
| Small ★ Most Popular | 60 | $10,000 | $167 | 6 months | Activation tier — most customers land here |
| Scale | 250 | $40,000 | $160 | 12 months | Multi-team expansion — best per-DU rate in self-serve flow |
| Enterprise | Custom | Custom | Custom | Custom | Strategic relationship with MSA, DPA, dedicated success |
The Project flow uses tier bands to determine the rate: 1–30 DUs at the Starter rate, 31–150 DUs at the Small rate, 151–499 DUs at the Scale rate, 500+ DUs at Enterprise custom. A 200-DU Project pays the Scale rate ($160) for every DU. A 50-DU Project pays the Small rate ($167) for every DU. This creates natural soft up-sell — if your work is sized at 28 DUs, expanding scope by 3 DUs to hit the Small band saves you $99 on the existing 28 DUs.
Why credit packs, not subscription
Per-seat or per-VDC subscription contradicts AiDOOS's operating model. The platform deliberately overstaffs for redundancy — better to have a senior architect on call who only intermittently delivers DUs than to undersize a pod and miss commitments. Buyers cannot stomach paying for that redundancy under a per-FTE model. They reasonably ask: "Why am I paying for 6 people when only 4 are billing?"
Under credit-pack pricing, that question disappears. The customer pays for DUs delivered. AiDOOS's internal pod composition is invisible — the platform decides whether to deploy 4 engineers or 6 engineers to ship 60 DUs in 6 months, and the customer's invoice doesn't change either way. This is the unit economics fix that lets a delivery platform actually overstaff without punishing buyers for it.
Credit packs also remove subscription friction. There is no monthly charge while you're not using the platform. Unused DUs roll forward when you top up. You can pause the engagement for a quarter and resume without renegotiating.
The trust mechanisms
DU pricing only works if the customer trusts the DU calculation. AiDOOS wraps the pricing in three guarantees that hourly contracts and per-FTE subscriptions structurally cannot match:
Pre-flight DU estimation, transparently shown
Before you spend a DU, the platform tells you exactly how many DUs the work will consume. Estimates come from the DU Dictionary — a reference catalog of past work, with AI-assisted matching against your story description, validated by the calibration board for unusual work. You see the breakdown: which factors drove the estimate, which alternatives were considered, what the historical accuracy of similar estimates has been.
Re-delivery on acceptance miss
If delivered work fails your acceptance criteria, AiDOOS re-delivers at no additional DU cost. The talent is paid by AiDOOS for the re-do; the customer's wallet is not debited a second time. This bounds the customer's risk: you cannot pay twice for the same outcome.
Refundable unused DUs, no questions asked
This is the mechanism nobody else can offer. DUs sitting in your wallet (not yet allocated to active work) are 100% refundable, pro-rata at the rate paid, within 24–48 hours to the original payment method. AiDOOS pays $0 of talent cost against unconsumed DUs, so refunding them costs the platform nothing. Toptal's monthly contracts cannot do this. Deel's EOR cannot do this. A Big-IT MSA cannot do this. The combination of refundable DUs + re-delivery guarantee + pre-flight estimation creates a buyer trust posture that competitors structurally cannot replicate — and it costs AiDOOS nothing on the unconsumed side.
How $/DU is calculated internally
Customers see a flat $/DU rate per tier. Behind the scenes, AiDOOS converts the work to DUs using a multi-factor calculation that is invisible at the invoice line. The factors include:
- Base unit — calibrated against the DU Dictionary reference catalog
- Technology multiplier — a Salesforce DU may consume more than a Python DU because Salesforce talent is scarcer
- Seniority multiplier — work that genuinely requires architect-level judgment consumes more DUs than work a senior engineer can ship
- Scarcity multiplier — niche stacks (rare ABAP variants, regulated-industry credentials) consume more DUs
- Complexity multiplier — a story that touches three systems consumes more DUs than the same code change in one system
Under the hood, AiDOOS multiplies a base unit by factors capturing technology, seniority, scarcity, and complexity. The customer never sees those factors as line items — they see one number: 8 DUs. The factor framework is rational and consistent; the specific values are calibrated continuously and aren't published, so the customer's mental model stays simple while the engine economics stay accurate.
Individual stories are capped at roughly 20 DUs — anything bigger must be decomposed. This caps DU inflation, keeps invoices predictable, and forces a discipline of breaking work into shippable units rather than letting "the integration project" become an opaque 200-DU blob.
What "outcome-based" actually means here
Many vendors claim outcome-based pricing. Most mean T&M with a milestone wrapper. AiDOOS DU pricing is outcome-based in the strictest sense:
- You pay only for delivered, accepted work. DUs in your wallet are your money until they consume against shipped output.
- You see the price before you authorize. Pre-flight estimation tells you what the work will cost in DUs before the platform commits a single hour of talent time.
- You get a re-delivery if delivered work fails acceptance. You don't pay twice for the same outcome.
- You can refund unused DUs at any time. The platform's only revenue against you comes from work that actually shipped and was accepted.
The economic shape is the inverse of T&M. Under T&M, the vendor profits when work takes longer. Under DU pricing, the vendor profits when work ships faster — because faster shipping means more DUs delivered against the same talent capacity.
When DU pricing is the right fit
DU pricing fits engagements where:
- The work can be decomposed into shippable units (most software engineering)
- Acceptance criteria can be agreed up front (or refined within the first sprint)
- The customer values predictable cost per delivered outcome over fixed monthly spend
- Scope is expected to evolve — DU pricing handles scope changes by consuming DUs differently, without contracting cycles
DU pricing is less natural for pure capacity-on-demand engagements where the customer wants a pod available regardless of how much it ships. For those, an Enterprise tier with a minimum-DU-floor commitment can approximate the retainer model — but if you're really buying capacity rather than output, you may be the rare buyer who genuinely wants per-FTE pricing.
Frequently asked questions
How do I know how many DUs a feature is worth?
The Instant Proposal flow estimates DU count for any work you describe. You can also browse the DU value catalog — vertical-specific examples to make the abstraction concrete before you commit budget. After you purchase the DUs, whenever you create a task, the DUs are automatically calculated.
Can I see the breakdown of how a DU estimate was calculated?
Yes. Every estimate shows the reference work it matched against, the factors that drove the size, and where in the historical accuracy distribution the estimate sits. Disputes go to the calibration board. The DU Dictionary is the platform's core IP — its calibration improves continuously from actual-vs-estimated tracking.
What happens if I buy a Small pack and only consume 30 DUs in 6 months?
Two options. You can refund the unused 30 DUs at the rate paid, no questions, no fees. Or you can top up before the 6-month window expires; the unused DUs roll forward into the new pack.
Does AiDOOS publish per-technology DU rates?
No — and that's deliberate. Publishing "ML DU = $250, React DU = $180" would force every conversation into multiplier negotiation. Publishing one universal $/DU per tier and letting the work consume more or fewer DUs based on its complexity keeps the customer's mental model simple while preserving rational unit economics underneath.
Can I pre-pay for DUs to lock the rate?
Yes. Pre-purchased packs lock the per-DU rate for the validity window (90 days for Starter, 6 months for Small, 12 months for Scale). Enterprise packs typically run multi-year with rate guarantees baked into the MSA.
How does this compare to a Big-IT services SOW or Toptal contract?
Big-IT SOWs are T&M (often dressed up as fixed-bid with painful change-order processes); Toptal contracts are per-FTE monthly subscriptions. Both put delivery risk on the customer. DU pricing puts the risk on AiDOOS — the platform only earns when work actually ships and is accepted.
Where to start
If you have a specific scope of work in mind, the fastest path is the pricing page — pick the tier that fits your DU range, or run an Instant Proposal that sizes the work in DUs and lands you on the right tier automatically. Starter ($2K, 10 DUs) is engineered specifically to be buyable on a manager's discretionary spend without procurement engagement.
If you want to compare DU pricing against your current vendor's TCD (total cost of delivery) before committing, the TCD framework walks through the conversion. For broader context on how AiDOOS Virtual Delivery Centers operate, see what is a Virtual Delivery Center. For the apples-to-apples comparison against staff augmentation, see VDC vs staff augmentation.
For an enterprise engagement with custom DU rates, MSA, DPA, and dedicated success management, schedule a 30-minute call and we'll size the engagement and quote a custom rate based on commit volume.