A tender delivery fit assessment compares one dated delivery scenario with the resources the bidder can substantiate. It turns the lot scope into demand by work package, skill, quantity, place, working pattern and period. It then deducts contracted client work, approved leave, training, operational cover and protected contingency from gross internal or partner capacity. The assessment joins that net supply to a resource-loaded schedule, exposes bottlenecks and records whether each gap is covered, conditional, unresolved or blocking. Its work product is a delivery_fit_record with the source tender version, scenario, demand rows, supply evidence, current commitments, schedule logic, gap treatments, decision state, owners and expiry triggers. It is a pre-bid operating decision. It does not prove tender eligibility, approve a bid, reserve people, promise key personnel or replace a detailed mobilisation plan.
A bid can describe a persuasive delivery model while the required people are already committed elsewhere. Company headcount hides timing and concentration. Ten engineers on payroll do not cover a four-week period that needs two cleared integration specialists at the same site. A named project manager may appear free in a resource tool while a signed client extension awaits entry. A partner may have the skill but no approved allocation. A recruitment request may be funded but still has no accepted candidate or start date. If these different states are all labelled available, the team can pass a tender condition and still lack an executable delivery plan. Averaging demand over the contract creates the same error: spare months after go-live cannot repair a shortage on the cutover weekend.
Start with the tender’s delivery clock and the obligations of the chosen lot. Build demand before looking at names. Measure concurrency, location, shift coverage, qualifications, equipment and dependencies in the periods where they are required. Then ask the accountable resource owner for dated supply evidence and deduct every protected commitment. Treat blank calendars, unsigned partner statements, proposed hiring and expected releases as conditions rather than capacity. Test schedule logic after resources are assigned because an unresourced critical path is only a sequence of dates. Keep buyer-facing conditions of participation separate from the supplier’s internal delivery test: satisfying a qualification rule does not prove that the same people are free. An agent may extract public obligations, normalize time periods, compare approved capacity and draft scenarios. It may not inspect private personnel data without authority, reassign staff, commit a partner, contact the buyer, authorize overtime, approve the bid or turn a conditional resource into a tender promise.
Scope
Capacity has no meaning until the delivery scenario is fixed
Begin with one decision object. Record the procedure identifier, notice or invitation version, lot and the tender documents that control delivery. Add the expected award date, earliest mobilization activity, contractual commencement, service term, sites, buyer working windows and any required cutover or readiness milestone. If the buyer offers alternative start dates or volumes, create separate scenarios. A blended scenario cannot tell the approver which commitment was tested.
Separate source values from supplier assumptions. A date printed in the contract schedule is source evidence. A four-week standstill allowance, an assumed 15-day buyer data handover or a forecast award slip is a planning assumption unless the procurement documents say otherwise. Store each assumption with an owner, reason, range and impact. If the feasibility result depends on it, the assumption is a condition of the result.
Delivery scope comes from more than the statement of work. Include service levels, opening hours, response times, security conditions, reporting, acceptance, buyer dependencies, transition duties, contract options and the submitted solution where they change demand. Keep optional volumes separate until the contract says whether the bidder must hold capacity for them. The goal is a defensible operating scenario, not the largest imaginable workload or the most convenient base case.
This scope step stops short of interpreting disputed legal obligations. Record the clause and its operational reading. Send a material ambiguity to the authorized clarification or legal owner. An agent may show two scenarios for two plausible readings, but it should not choose the cheaper reading and describe it as the contract.
| Field | Value to retain | Why it changes the test |
|---|---|---|
| decision_id | Stable local identifier | Keeps later reviews attached to the same assessment |
| procurement_scope | Buyer, procedure, notice version and lot | Prevents capacity from being transferred between different obligations |
| delivery_scenario | Named base, alternative or stress scenario | Shows which combination of dates and volumes was tested |
| time_basis | Award, mobilisation, start, milestones, term and calendars | Locates demand in the periods where work must happen |
| place_basis | Remote, site, region, travel and coverage assumptions | Determines which capacity can reach the work |
| scope_sources | Exact clauses, schedules and approved response choices | Allows another reviewer to reconstruct demand |
| assumptions | Value, range, owner, rationale and effect | Keeps missing buyer facts visible |
Demand
Build demand by activity, period and constraint before assigning names
Break the scenario into work packages that finish with an observable output or readiness condition. For each activity, capture the earliest start, duration range, predecessors and milestone consequence. Then describe the resource demand. A role label alone is too weak. The row needs the skill or authorization, productive quantity, minimum simultaneous presence, location, working pattern, equipment and the reason that the resource is needed.
Use units that reflect the obligation. Person-days work for ordinary project activity when the day length and calendar are defined. Named shifts are better for a service desk. Two engineers for a six-hour joint cutover is a concurrency requirement, not twelve interchangeable hours. One qualified safety lead present at each of three simultaneous sites is three people even when the total inspection effort is small. Equipment can be constrained in the same way: a test environment, vehicle or certified device may be needed by several tasks at once.
Show demand source and derivation separately. “Two trainers per site” may come directly from a buyer schedule. “Twenty-four trainer-days” may be a calculation from six sites, two trainers and two days per site. Keep the formula and rounding rule. If an agent extracts a table from a PDF or converts weeks to working days, retain the raw statement, the transformation and the calendar used. A clean total without that lineage is difficult to challenge.
Do not lower demand to match the people found later. If delivery leadership proposes a different method, create a revised scenario with the change, operational rationale, buyer compliance check and approval state. The original demand remains available for comparison.
| Work package | Resource condition | Time and place | Demand evidence |
|---|---|---|---|
| Data discovery | Migration lead, 20 person-days, at least 0.6 FTE | Weeks 1 through 6, remote plus two buyer workshops | Implementation schedule clauses 2.1 and 2.4 |
| Interface build | Two integration engineers working concurrently, 72 person-days total | Weeks 3 through 10, buyer test window applies | Six interfaces multiplied by estimated build and joint test effort |
| Depot training | Two approved trainers per session and six parallel site windows | Weeks 9 through 11, on site | Training requirement plus provisional rollout sequence |
| Cutover | Two engineers, service lead and one rollback decision owner | Saturday 20:00 to Sunday 08:00 at the control site | Buyer cutover window and supplier continuity design |
| Service acceptance | Evidence owner and operational signatory | Week 12 before service commencement | Acceptance schedule and internal authority model |
Supply
A person counts only in the periods and conditions where they can be used
Build supply from the same dimensions as demand. Record a named person or controlled pool, employer or supplying entity, verified skills, base location, approved working pattern, relevant equipment access and the periods under review. Add who supplied the evidence and when. If a qualification or clearance expires before the activity, it cannot cover that row without a separately approved renewal condition.
Net available capacity for a period is approved gross capacity minus signed client work, internal operational duties, leave, training, travel and protected contingency. Keep every deduction visible. Do not infer availability from a blank calendar, a missing timesheet or the absence of a resource booking. The resource owner may know about an extension, incident rotation or priority programme that has not reached the planning system.
Classify supply before using it. Reserved capacity has an authorized allocation to this scenario. Available capacity is confirmed free but not yet reserved. Conditional capacity depends on a stated event such as release from a current contract, signed partner allocation, accepted job offer or clearance. Unverified capacity has no current owner evidence. Unavailable capacity is protected elsewhere or fails a required condition. Only reserved and approved available capacity can support fit confirmed, and the record should still say whether reservation must happen before bid approval.
Protect the information. A delivery decision usually needs role, capability, availability band, entity, location and owner confirmation. It rarely needs health details, private calendar entries or a full personnel file. Keep personal evidence in its authorized system and link to an access-controlled reference. Publish only the minimum decision facts.
| State | Evidence | May support fit confirmed? |
|---|---|---|
| reserved | Authorized allocation to this scenario for exact periods and demand | Yes |
| approved_available | Resource owner confirms net capacity and conditions; allocation still pending | Yes, with the reservation requirement shown |
| conditional | Named dependency, owner, due date and closure proof exist | No, supports conditional fit only |
| unverified | Planning record or assertion lacks current owner confirmation | No |
| unavailable | Protected commitment, failed condition or time conflict is evidenced | No |
| authorization_missing | The reviewer cannot access or obtain the required confirmation | No decision until the authorized owner responds |
Evidence
Passing a participation condition does not reserve the delivery team
Public procurement rules explain why buyers ask about resources, but the buyer’s test and the bidder’s operating test are different. Section 22 of the UK Procurement Act 2023 allows proportionate conditions about legal and financial capacity or technical ability to perform the contract. Current Cabinet Office guidance, updated on 17 August 2026, separates those conditions from award criteria and notes that evidence may be independently verifiable. A supplier that satisfies the condition may still have a conflict in the proposed start month. The delivery fit record therefore cites the buyer condition, records the submitted evidence and performs a separate dated capacity calculation.
Article 58 of Directive 2014/24/EU similarly concerns the human and technical resources and experience required to perform to an appropriate quality standard. Annex XII and the European Single Procurement Document include evidence such as technical staff, qualifications, average annual manpower, managerial staff, equipment and intended subcontracting. These are useful evidence categories. Historical workforce and equipment declarations do not identify the people or assets that remain free for a future milestone.
Third-party capacity needs two checks. Under Article 63 of the Directive, a bidder relying on another entity must prove that the necessary resources will be at its disposal; where qualifications or professional experience are relied upon, that entity must perform the work for which those capacities are required. The exact tender and governing rules control the submission. Internally, the fit assessment also needs dates, quantities, location, work package, authority and conflicts. A generic letter about cooperation may address none of them.
Treat minimum eligibility as a gate and delivery fit as a current scenario result. Keep both. Do not use an internal shortage to rewrite a buyer criterion, and do not use a passed criterion to overrule the delivery owner.
Schedule
The bottleneck sits where scarce capacity meets an inflexible activity
Assign eligible supply to activities and recalculate the schedule. Do not preserve a milestone by allowing the same person to occupy two tasks at once. Use the working calendars, travel limits and shift rules from the supply view. If an activity can use a pool, set the skill and concurrency requirements before the scheduler chooses people. If a task requires one named specialist, model that person directly.
The GAO Schedule Assessment Guide describes a reliable schedule as one that captures all activities, sequences them logically, assigns resources, validates the critical path, identifies reasonable float, analyzes schedule risk, updates actual progress and maintains a controlled baseline. A pre-bid assessment need not copy a 240-page audit method. It does need the same discipline where the conclusion depends on time. Resource loading may create a different critical path from the unresourced bid plan.
Inspect peak demand and near-critical paths. A role with 100 person-days of free capacity across a year may still fail a 30-day window that requires two people continuously. Float is usable only when the predecessor logic, buyer dates and resource calendars allow it. A ten-day delay on a noncritical activity may be harmless, while one missing day from the only clearance holder can move service commencement.
Use ranges when duration or volume is uncertain. Test a base case and a named stress case with the same method. Do not convert the probability output of a schedule risk analysis into a promise to the buyer. It informs the internal feasibility judgment and the amount of time or capacity contingency that the authorized owner may choose to protect.
| Dimension | Pass condition | Blocking evidence example |
|---|---|---|
| scope | Demand is based on the controlling lot and an explicit scenario | Material buyer dependency has no tested assumption range |
| skill | Every critical activity has an eligible skill source | Only proposed hire has the mandatory certification |
| capacity | Net supply covers quantity and concurrency in every critical period | Integration pool has one person free where two are simultaneous |
| place and calendar | Resources can work at the required location and window | Travel makes the same-day site sequence impossible |
| schedule | Resource-loaded logic meets milestones with approved contingency | Critical-path migration slips beyond the start date |
| authority | Resource and treatment owners have confirmed their decisions | Partner sales contact lacks allocation authority |
Treatments
A cure becomes capacity only after its dependency closes
A gap can have a credible treatment without being closed. Record the treatment, affected demand row, owner, approval authority, earliest decision date, lead-time basis, dependencies and the evidence that will change the supply state. Examples include an authorized release from current work, a signed subcontractor allocation, a completed internal transfer, an accepted employment offer with a feasible start date, an approved delivery redesign or a buyer-issued change to the timetable.
Keep commercial optimism out of the state. “We can probably hire” is unverified. A funded requisition with a market lead-time range is conditional. An accepted candidate may still depend on notice period, screening, equipment and onboarding. An affiliate has no capacity in the record until the supplying entity and allocation owner confirm the work, period and quantity. The same rule applies to a subcontractor whose capability is well documented but whose delivery team has not committed.
Treatments must respect the tender. Moving work offshore, replacing a named person, changing on-site coverage, reducing transition activity or delaying a milestone may need buyer permission or may conflict with the proposed solution. The internal team can model the alternative, but only the tender documents or an authorized buyer clarification can change the obligation.
Set a closure gate before the bid becomes binding. If the tender permits the condition to be satisfied by award or commencement, record that exact date and source. Otherwise use the internal commitment gate. A conditional fit should never arrive at submission with no owner watching its unresolved dependencies.
Worked example
A twelve-week depot rollout fails in weeks three through six
Consider a fictional single-lot tender for a logistics platform across six depots. The buyer expects mobilisation to start on 11 January 2027, site training in weeks nine through eleven and one coordinated cutover in week twelve. The response outline looks credible. It lists a programme lead, migration lead, integration team, partner trainers and service lead. The first capacity pass also looks positive because the supplier employs three integration engineers and the training partner has ten trainers.
The dated supply view changes the result. The programme lead has approved 0.6 FTE capacity, which covers the scenario. The migration lead remains committed to a signed client extension until the end of week five. Only one integration engineer is free in weeks three through six, while the interface plan requires two working concurrently. The partner’s letter says it intends to support the bid but gives no sites, dates, quantities or allocation authority. Two trainers are tentatively available, yet the delivery plan needs paired trainers at simultaneous depot windows. The service lead and cutover engineers are covered.
Resource loading puts migration preparation and interface testing on the critical path. Moving the migration lead to week six delays the earliest tested cutover by fifteen business days. Splitting the interfaces across one engineer avoids double booking but creates the same date failure. The training row remains unresolved because a corporate capability statement cannot show who will cover the site windows.
The record therefore returns not_fit on current evidence. It carries three blocking rows: migration timing, integration concurrency and partner allocation. A later buyer start, an approved internal specialist release or a dated partner commitment could produce a different scenario. None is counted yet. The bid authority receives the blockers and treatment gates rather than a green score diluted by the roles that passed.
| Demand | Current evidence | Result | Closure proof |
|---|---|---|---|
| Programme lead, 0.6 FTE weeks 1 through 12 | Named owner confirms 0.6 FTE after protected duties | covered | Reserve allocation before bid approval |
| Migration lead, weeks 1 through 6 | Named lead committed elsewhere through week 5 | blocking conflict | Authorized release or compliant revised start |
| Two integration engineers, weeks 3 through 6 | One eligible engineer free; two others protected | blocking concurrency gap | Second approved engineer for exact period |
| Paired trainers at six site windows | Partner capability known; allocation statement has no dates or quantity | unresolved | Authorized partner roster and dated allocation |
| Cutover and service lead | Named resources and weekend pattern confirmed | covered | Final reservation and fatigue check |
Decision
Return the evidence and stopping point, not a single capacity score
Use five top-level states. fit_confirmed means every critical dimension is covered by reserved or approved_available supply and the schedule meets the scenario with the approved contingency. conditional_fit means the scenario works only if named conditions close by their gates. not_fit means at least one blocking demand cannot be covered under the current scenario. indeterminate means material scope or supply evidence is missing. authorization_missing means the facts may exist, but the reviewer cannot obtain a decision from the person entitled to provide it. Keep the dimension results underneath the top-level state.
Make the record inspectable by another person or agent. Include stable IDs, source locators, observed times, raw demand statements, derivations, calendars, demand and supply units, resource states, owner assertions, schedule version, conflicts, treatments, decision rationale and expiry triggers. Avoid publishing confidential client allocations or personal details. A public or broadly shared summary can expose demand, coverage state, evidence class and responsible function while the underlying authorized system retains the sensitive proof.
An agent can extract tender dates and resource conditions, create demand rows, detect overlaps, run arithmetic, load a scenario and draft the decision. It must preserve the source and every assumption. It should abstain when identity, scope, authorization or private evidence is missing. It cannot make a blank calendar authoritative, disclose personal data, move someone off a client, bind a partner, approve overtime or promise a treatment in the tender.
Expire the decision on a tender amendment, lot or volume change, revised award or start date, new current-work commitment, resource departure, leave change, failed hiring gate, changed partner allocation, qualification expiry, schedule redesign or passage of the review interval. Recompute affected rows first, then the critical path and top-level state. The historical record remains unchanged so the team can see which assumption or commitment moved the decision.
This assessment hands evidence to the bid authority. A positive result does not prove eligibility, attractive economics, acceptable contract terms or strategic value. A negative result does not select a partner or authorize a buyer clarification. Those decisions need their own records and owners.
| Object | Required content | Safe use |
|---|---|---|
| scope | Procurement, version, lot, scenario, dates, locations and sources | Reconstruct the obligation being tested |
| demand | Activity, resource condition, quantity, concurrency, period, calendar, place and derivation | Compare like-for-like resource needs |
| supply | Controlled resource reference, entity, skill, net availability, state, owner and observed time | Use only within approved access and decision scope |
| schedule | Activities, dependencies, assignments, critical path, float, range and version | Recalculate after any material resource change |
| gaps | Affected demand, consequence, treatment, authority, gate and closure proof | Keep conditions visible until evidence changes |
| decision | Top-level state, dimension states, rationale, approver and next review | Inform but never replace the authorized bid decision |
| validity | Assessed time, review interval and event triggers | Prevent reuse after the evidence moves |
What good looks like
Useful outcomes from tender delivery capacity assessment
- One lot, tender version and delivery scenario define the scope of the decision.
- Every material work package has dated demand for skills, quantity, location, working pattern, equipment and dependencies.
- Gross headcount is replaced by net capacity after current obligations and protected cover are deducted.
- Internal, affiliate, subcontractor, candidate and unfilled-role capacity remain visibly different.
- Peak concurrency and critical-path demand are tested instead of relying on contract-wide averages.
- Each resource assertion names an accountable owner, evidence date, confidence state and reservation status.
- Blocking gaps, permitted treatments and proof required for closure appear in the same record.
- The result states what is feasible now, what depends on a condition and what remains unknown.
- Existing client and operational commitments stay protected from an unsupported bid promise.
- The assessment expires when scope, dates, availability, partner commitments or current work changes.
Operating model
How to run the work
- 01
Fix the delivery scenario
Record the procedure, current tender version, lot, scope basis, expected award and start dates, service locations, term, volumes, working windows and every assumption that changes resource demand.
- 02
Convert obligations into dated demand
Break the work into packages and milestones. For each period, state the skill or asset, quantity, concurrency, location, working pattern, predecessor and evidence source.
- 03
Build the evidenced supply view
Ask authorized owners for the people, pools, equipment and partner capacity that could serve the scenario, together with qualifications, base location, calendars, current commitments and evidence dates.
- 04
Calculate net capacity
Deduct contracted work, approved leave and training, operational duties, realistic travel and protected contingency. Keep proposed releases, hiring and unsigned partner support conditional.
- 05
Load resources onto the schedule
Assign eligible supply to activities without double booking. Recalculate start, finish, critical path and float under the same calendars used for the capacity view.
- 06
Test treatments and authority
For every gap, identify a permitted change, its owner, decision date, prerequisites and closure proof. Do not assume the buyer will accept a later date, different location or substituted person.
- 07
Issue and expire the fit record
Return fit_confirmed, conditional_fit, not_fit, indeterminate or authorization_missing, with blocking facts and the next authorized review. Attach recheck triggers instead of allowing a timeless green status.
Evaluation
Questions that change the decision
- Which procedure, notice version and lot define the delivery obligation?
- What start date, mobilisation period, service term and buyer milestones control the scenario?
- Which demand is explicit, derived from the response design or still assumed?
- Where do simultaneous activities create a higher need than total effort suggests?
- Which roles require a named person, qualification, clearance, language or previous experience?
- Which work must occur at a site, in a coverage window or under a specific travel assumption?
- What internal work, client obligation, leave, training or operational reserve is already protected?
- Who has authority to confirm each person, pool, asset or partner allocation?
- Can a relied-on entity prove both its capability and its dated commitment to perform?
- Which activity is on the resource-loaded critical path, and how much usable float remains?
- Can a gap close before the tender’s commitment gate without changing an unapproved buyer obligation?
- What change invalidates the assessment and who must review it again?
Failure modes
Where teams lose control
Annual headcount may be treated as proof of availability for a particular month.
The same specialist may be allocated to overlapping activities or pursuits.
A current contract extension may be omitted because it has not reached the resource system.
Leave, training, incident duty or business-as-usual support may be treated as movable spare time.
An FTE average may hide a shift, weekend or two-person concurrency requirement.
Travel and site access may consume the hours that were counted for productive work.
A qualification may be valid but held by a person who cannot perform in the required period.
An affiliate or subcontractor capability statement may be mistaken for an allocation commitment.
Funded recruitment may be counted before a suitable candidate accepts and clears onboarding.
A schedule may retain fixed milestone dates after resource loading moves its critical path.
A conditional treatment may be presented in the bid as if its approval already exists.
Personal availability or commercial allocation details may be exposed beyond the people authorized to review them.
Measurement
Measure the finished job
Measure the completed workflow, including review effort and exceptions. Output volume on its own is not evidence of a better process.
- critical demand rows linked to an exact tender or scenario source
- demand periods with explicit quantity, concurrency, calendar and location
- supply rows confirmed by an accountable owner within the review window
- claimed capacity supported by current commitments and reservation state
- critical roles covered by approved resources in every required period
- double bookings found before tender approval
- critical-path activities with assigned skills and usable float
- partner allocations backed by a dated authorized commitment
- conditional gaps closed with the specified proof before the commitment gate
- fit records rechecked after an expiry trigger
- client or operational commitments displaced after a fit-confirmed decision
Questions
Common questions
Does passing the tender’s technical selection criteria prove delivery fit?
No. Selection evidence can prove qualifications, experience, historical workforce or access to resources. Delivery fit also tests current commitments, exact periods, concurrency, location, schedule logic and allocation authority.
Can company headcount be used as available capacity?
Only as background evidence. The decision needs net capacity for the required skill, time, location and working pattern after protected work and contingency are deducted.
Should named people be reserved before the supplier bids?
Follow the tender, contract and internal authority rules. The record must at least distinguish reserved people from confirmed but unreserved capacity and state the gate by which reservation is required.
Can an open vacancy close a delivery gap?
It is a conditional treatment, not current supply. Record hiring approval, lead-time evidence, required start, onboarding or screening, owner and the accepted-candidate proof needed to close it.
Does a subcontractor capability statement count as capacity?
Not by itself. The assessment needs the supplying entity, work package, resource quantity, dates, location, conflicts and an authorized commitment that matches the tender rules.
Why is total person-day capacity insufficient?
It can hide simultaneous work, shifts, site presence and critical-path timing. Spare capacity after a milestone cannot cover a shortage before it.
Can overtime be used to make the scenario fit?
Only when the authorized operational owner confirms a lawful, safe and sustainable working pattern for the exact period. The assessment must not invent overtime or consume protected incident cover.
What if the tender start date or volumes are uncertain?
Create separate named scenarios or ranges and state which source or assumption controls each. Do not average incompatible cases into one green result.
May an AI agent make the final bid or no-bid decision?
No. It may assemble and test the evidence within its access and calculation authority. The accountable delivery, commercial and bid owners decide whether the supplier will make the commitment.
Sources
Primary references
- Guidance: Conditions of Participation UK Cabinet Office
- Procurement Act 2023, section 22 The National Archives
- Directive 2014/24/EU, consolidated text Publications Office of the European Union
- European Single Procurement Document standard form Publications Office of the European Union
- GAO Schedule Assessment Guide U.S. Government Accountability Office
- Project Routemap: Delivery Planning UK Infrastructure and Projects Authority
- Overview of the Project Set Up Toolkit UK Infrastructure and Projects Authority
Zelius
Managed tender intelligence and bid execution for teams that want the commercial outcome.
Suppliers, founders and commercial teams pursuing public or private opportunities. Start with the workflow, constraints and evidence you already have.