Most partnerships between a prime and a specialist engineering firm do not end in a dispute. They end because the prime asked for something reasonable-sounding that quietly changed the arrangement from a partnership into a supply relationship, and the partner stopped bringing its best engineers, stopped investing ahead of pursuits, and eventually stopped answering the first call. Six requests do most of that damage. Each has a version that gets the prime what it actually wanted, at no more cost.
This is written for the program manager or subcontracts manager making these requests, usually with good intentions and under real pressure. None of these is misconduct. They are ordinary asks that carry a cost the asker does not see, because the cost lands on the next pursuit rather than this one.
Unpaid proposal work with no commitment attached
The request: help write the technical volume, supply engineers for color team reviews, build a demonstration for the customer briefing, and if we win, we will talk about workshare.
Why it backfires. Proposal support is the specialist's most expensive labor, drawn from the same senior engineers who are billing on delivery work. A firm asked for that with nothing committed will supply its available people rather than its best, which is exactly the opposite of what the prime needs, because the technical volume is where the pursuit is won or lost. Repeated across several pursuits, the firm concludes that this prime's opportunities are lottery tickets and starts prioritizing teams that commit.
What works instead. Commit something specific and small. Exclusivity on this pursuit for a defined period costs the prime nothing when it has already chosen the partner, and it changes who the partner assigns. A named workshare range in the teaming agreement, contingent on award and on the scope surviving into the final solicitation, is normal and sufficient. For substantial pre-award engineering, such as a demonstration that will be shown to the customer, fund it as a small task order and own the result. A prime that pays for a demonstration owns an asset it can carry into the next pursuit; a prime that asks for it free owns nothing.
Requests ranked by how much damage they do to a specialist partnership
Editorial weighting, illustrative rather than measured. The last row is deliberately low: negotiating price on defined scope is normal and healthy.
Staff augmentation disguised as workshare
The request: the proposal described the partner owning the analytics subsystem, but after award the program needs three engineers on the integration team, reporting to the prime's technical lead.
Why it backfires. It converts an accountable scope into supervised hours, and those are different products with different prices, different management costs and different outcomes. The specialist firm can no longer be held to a result because it no longer controls the boundary. The prime gains a management burden it did not have and loses the accountability it bought. It also creates a problem in the file: the proposal described a scope, the execution does not match it, and the next performance conversation is awkward for both parties.
There is a second cost that is easy to miss. Specialist firms retain senior engineers by giving them ownership of technical outcomes. Placed as supervised staff on someone else's integration team, those engineers leave for a firm that gives them scope. The prime experiences this as turnover on the subcontract and blames the partner.
What works instead. If the program genuinely needs staff augmentation, buy it as staff augmentation from a supplier that sells it, and price it that way. If the program needs the specialist's capability, preserve a boundary the partner can own, even a smaller one: an interface, a service, a data layer, with acceptance criteria and its own milestones. When priorities change mid-program, renegotiate the scope explicitly rather than letting it dissolve into an hours pool by accretion.
Intellectual property assignment without compensation
The request: standard flow-down, everything the partner produces assigns to the prime, including anything it brings.
Why it backfires. Every specialist worth engaging arrives with existing tooling: pipelines, evaluation frameworks, deployment patterns, model architectures refined across projects. That is what makes them fast. A blanket assignment clause asks them to hand over the thing that lets them deliver, in exchange for the price of one engagement. A firm that reads the clause carefully will decline, negotiate, or price the loss into the bid. A firm that signs it without reading has told you something about its diligence.
There is a technical cost as well. A partner unwilling to assign its background will build around it, writing a bespoke version of tooling it already has. The prime pays for that reconstruction in schedule and quality, and receives a weaker system than the one that would have used the mature components.
What works instead. Separate three categories in writing. Background, meaning what the partner brings, named specifically, excluded from assignment and licensed to the government and the prime for use, maintenance and modification of the delivered system, on terms that will not block a future maintainer. Foreground, meaning what is created under this subcontract, assigned or licensed per the prime contract's data rights requirements. And jointly funded work, where ownership follows the money and is settled before anyone writes code. On copyright specifically, remember that a work-made-for-hire recital does not by itself transfer ownership of software from an independent contractor; under 17 U.S.C. § 101 a commissioned work qualifies only if it is written and falls within one of nine enumerated categories, and software is not among them. A written present assignment is what actually transfers copyright.
Open-ended key personnel commitments
The request: commit these three named engineers full time for the five-year ordering period, with no substitution.
Why it backfires. Nobody can honestly make that commitment, so one of two things happens. The partner declines, and the prime loses a strong technical volume. Or the partner agrees knowing it cannot hold, and the prime has a compliance problem the first time somebody changes jobs. Meanwhile the clause has told the partner that the prime prefers a promise it cannot keep to an honest constraint, which shapes everything else the partner says.
What works instead. Commit the people whose commitment the evaluation actually needs, for the period the evaluation actually covers, usually the base period. Write a substitution path with a real standard: qualifications equal or better, notice period, prime approval not unreasonably withheld. Ask for allocation percentages rather than assuming full time, because a fifty percent commitment that is real is worth more than a full-time commitment that is aspirational. And ask about depth: how many people at the partner can do this work, which is the question that actually predicts continuity.
| The request | What it costs the prime | The version that works |
|---|---|---|
| Unpaid proposal support, no commitment | The partner's available people instead of its best, on the volume that decides the bid | Pursuit exclusivity, a workshare range, and funded pre-award engineering the prime then owns |
| Workshare converted to staffing | Accountability lost, management cost gained, senior engineers leave | Preserve a boundary the partner owns, with acceptance criteria; renegotiate scope openly |
| Blanket IP assignment | Bespoke reconstruction of mature tooling; slower, weaker delivery | Background named and licensed, foreground assigned per the prime contract, joint work settled first |
| Open-ended key personnel | An unkeepable promise or a lost technical volume | Base period commitments, stated allocations, a real substitution standard |
| Pass-through pricing pressure | Junior staffing, thin coverage, and quality problems in month four | Compete the scope, then buy the result; measure delivered outcomes, not the rate |
| Total customer contact ban | Requirements arrive through a translator; rework and slipped schedule | Defined technical channel, prime present, business and contractual matters routed to the prime |
Pass-through pricing pressure
The request: match the labor rates in our rate card, or the price analyst will not clear it.
Why it backfires. Comparing a specialist firm's rates to a large prime's rate card compares two different products. The specialist's price on a bounded, fixed-price scope includes the risk it is carrying, the tooling it brings, and the fact that the work is done by people who have done it before. Driven to a rate card, the firm staffs to the price: junior engineers, thinner coverage, less senior review. The prime gets the rate it asked for and the delivery it did not want, and discovers the difference in month four when the increment fails acceptance.
The related error is treating a specialist's price as a pass-through to be minimized. On a fixed-price milestone the partner is carrying schedule and technical risk. That risk transfer has a price, and removing the price does not remove the risk; it moves it back to the prime, where it is more expensive.
What works instead. Compete the scope, not the rate. Ask two or three qualified firms to price the same defined increment with the same acceptance criteria, and compare total price against the same result. Where a rate structure genuinely has to be reconciled to a prime contract requirement, say so plainly and work the scope down to fit the money rather than asking for the same scope cheaper. A partner told the budget honestly will usually propose a smaller first increment that fits, which is a better outcome than a full scope staffed thin.
Customer contact bans that prevent delivery
The request: no contact with the customer, all questions through the prime's program manager.
Why it backfires. There is a legitimate version of this concern. Primes own the customer relationship, they carry the contractual interface, and uncontrolled contact creates real problems: scope creep, mixed messages, and confusion about who directs the work. But a total ban on technical contact makes engineering impossible. Requirements for a data system are discovered by talking to the people who know the data. A specialist team that cannot ask a government data steward why a field is populated inconsistently will guess, build to the guess, and rework it after the demonstration. That rework lands on the prime's schedule.
What works instead. Separate technical communication from business and contractual communication. Technical questions flow through a defined channel, in the program's normal forums, with the prime present or informed. Anything touching scope, schedule commitments, cost, or the contract goes through the prime without exception, and the partner should honor that without being reminded. Introduce the partner's engineers to the government technical leads early, in the program's regular working sessions, so the relationship exists before there is a problem to solve. Primes that do this get better delivery and lose nothing, because the customer relationship is not held by whoever answers a technical question about a schema.
The pattern underneath all six
Each request substitutes a general claim on the partner for a specific commitment to it. Unpaid proposal work claims senior engineering without committing a pursuit. Staffing claims labor without committing a boundary. Blanket assignment claims everything the partner owns without committing a price. Open-ended key personnel claims years without committing a period. Rate-card pressure claims a price without committing to a defined scope. A contact ban claims control without committing a channel.
The general claim always looks cheaper because it costs nothing to write. It is more expensive in execution because a partner facing a general claim protects itself generally: by assigning the people it can spare, by building around its own tooling, by pricing the uncertainty, by declining to invest ahead of the pursuit. None of those responses is visible in the subcontract. All of them are visible in the delivery.
The specific commitment, by contrast, is usually cheap. Naming a period, a scope boundary, an interface, a license, an allocation percentage or a communication channel costs the prime almost nothing and changes what the partner can safely commit in return. That asymmetry is the whole practical lesson: precision on the prime's side buys commitment on the partner's side at a very favorable exchange rate.
Specific commitments that cost the prime least and change the partner's behavior most
Editorial weighting, illustrative rather than measured. The last row is deliberately low: an unnamed future changes nobody's staffing decision.
How we work inside a prime's program
Precision Federal builds AI systems, data platforms, cloud infrastructure and full-stack web and mobile applications, and delivers them into production inside federal agencies. We are a small business and we work as a specialist subcontractor and teaming partner to large primes. We are direct about scope, price and risk at the start, because everything above comes from arrangements where those were left ambiguous.
The first weeks are concrete. In week one we read the architecture, the data model, the current pipeline and the security package, and we sit in the program's normal technical forums. By the end of week two we return a written finding: what the workstream requires, where the technical risk sits, what we would change in the first increment, and a measured baseline for whatever acceptance will be scored against. Weeks three through eight produce a working increment in the program's own environment, against real data, with tests, a deployment pipeline and the control evidence a security review will ask for. We follow the prime's configuration management and reporting practice rather than importing our own.
The prime keeps what matters. The customer relationship and the contract stay with the prime, and anything touching scope, cost or schedule goes through the prime's line. The code, models, data, pipelines and documentation are delivered under the subcontract, with our pre-existing tooling named as background and licensed so nothing in the delivered system is blocked for whoever maintains it later. We supply performance narrative and technical detail for the prime's own volumes and evaluation responses without being chased for it.
Pricing takes one of two shapes. A bounded increment prices as a firm fixed-price milestone against written acceptance criteria, which puts schedule and technical risk on us. A continuing workstream prices as a committed team at a stated allocation, with named engineers and a substitution path in the subcontract. When a scope is not bounded enough for a fixed price, we say so and propose the smaller increment that is.
The first step is one email with a one-page brief: the program or pursuit, the workstream, the environment named by product, what data exists and who grants access, the security destination, and the date that matters. We return a scoped, priced statement of work.
Bottom line
Every request on this list has a legitimate concern behind it and a version that serves the prime better. Commit something small before asking for expensive proposal engineering. Keep a boundary the partner can own instead of dissolving scope into hours. Name background technology and license it instead of assigning everything. Commit key personnel for the period that is real, with a substitution standard. Compete the scope rather than the rate, and work scope down to a budget honestly. Route business matters through the prime and let engineers talk to engineers. Primes that operate this way get better technical volumes, faster delivery and partners that answer the first call. The alternative is a supplier base that behaves exactly as it has been asked to.
Frequently asked questions
Some pre-award effort is normal on both sides. The problem is asking for substantial senior engineering with nothing committed, because the partner will supply its available people rather than its best on the volume that decides the bid. Commit something small and specific instead: exclusivity on that pursuit for a defined period, a workshare range in the teaming agreement contingent on award, and funding for any demonstration the customer will see, which the prime then owns and can reuse.
It changes the product. A scoped workstream can be held to acceptance criteria; supervised hours cannot, so the prime loses accountability and gains a management burden. It also creates a mismatch between what the proposal described and what is executed. Specialist firms retain senior engineers by giving them ownership of outcomes, so placing them as supervised staff produces turnover the prime experiences as a partner problem. If staffing is what the program needs, buy staffing and price it as staffing.
In three named categories. Background is what the partner brings, listed specifically, excluded from assignment, and licensed to the government and the prime for use, maintenance and modification of the delivered system so no future maintainer is blocked. Foreground is what is created under the subcontract and follows the prime contract's data rights requirements. Jointly funded work has ownership settled before code exists. For software copyright, a written present assignment transfers ownership; a work-made-for-hire recital alone generally does not.
For the period the evaluation covers, usually the base period, with stated allocation percentages rather than an assumption of full time, and a substitution path with a real standard: qualifications equal or better, a notice period, and prime approval not unreasonably withheld. Open-ended multi-year commitments with no substitution produce either a refusal or an unkeepable promise. Ask separately how many people at the partner can do the work, since that is what actually predicts continuity.
They have to for technical work to succeed, but the channel should be defined. Technical discussion happens in the program's normal working forums with the prime present or informed. Anything touching scope, schedule commitments, cost or the contract routes through the prime without exception. A total ban means requirements arrive through a translator and get rebuilt after the demonstration, on the prime's schedule. Introducing the partner's engineers to government technical leads early is what prevents that.