Skip to content

Search RFP.co

Pages across the product, solutions, opportunities and resources.

Type to search. Press Enter to open the full results page.

Start free trial
Procurement opportunity intelligence

Find the Right RFPs. Build Better Responses.

RFP.co discovers opportunities across public and private sources, matches them to your capabilities and helps your team create complete, tailored and compliant responses.

Every interface on this page is a demonstration built from fixture data. Nothing here is a live feed, a customer, or a claim about anybody’s results.

Opportunity listeningDemonstration · fixture data
  • Federal sources
  • State portals
  • Local portals
  • Web search
  • News
  • Public databases
  • Social signals
Intake
  1. Ingested
  2. Classified
  3. Matched
  4. Scored
  5. Ready to review
Seven categories of source feed one intake path. Every record is stored with where it came from and when it was fetched, then classified, matched against your capability profiles and scored before anybody is asked to look at it.
Opportunity listening

Everything that arrived, with where it came from attached.

A record is worth acting on when you can see its provenance. Every row carries the source that produced it, the moment it was fetched, and a link back to the posting exactly as it was published.

Intake streamDemonstration · fixture data
Fixture records in intake, one from each source category7 records — 4 opportunities, 2 signals, 1 prediction
SourceRecordTypeClosesIntake
Federal sourcesSOL-8827-QManaged detection and response servicesFederal civilian agency — IT operations divisionCybersecurityManaged servicesSmall business set-asidetype: opportunityReady to review — stage 5 of 5
State portalsSTA-2026-1163Statewide learning management platformState education agencySoftwareEducationMulti-yeartype: opportunityReady to review — stage 5 of 5
Local portalsRFP-2026-0418Citywide fibre network expansion — design and constructionCity of Fairview, department of public worksConstructionTelecommunicationsBondedtype: opportunityScored — stage 4 of 5
Web searchWEB-4471Water treatment plant SCADA modernizationRegional water authorityIndustrial controlsIntegrationtype: opportunityMatched — stage 3 of 5
NewsNWS-0912Council approves fare collection replacement programmeMetropolitan transit authorityTransitPaymentsBudget approvedtype: signalNo deadlineClassified — stage 2 of 5
Public databasesDB-33018Grounds maintenance — final option year exercisedCounty facilities departmentFacilitiesRecompete expectedtype: predictionNo deadlineClassified — stage 2 of 5
Social signalsSIG-1177Health department advertising six security analyst rolesState health departmentCybersecurityStaffingtype: signalNo deadlineIngested — stage 1 of 5

Opportunities carry a closing date; signals and predictions do not, because neither is an invitation to bid. They are separate object classes in the data model and the interface never merges them.

Trusted sources

Seven categories of source, and what a record from each one carries.

Coverage is described by category rather than by a wall of third-party marks. What matters about a source is not its logo; it is what a record from it can tell you and what it cannot.

  • Federal sources

    Opportunities

    Solicitations published by federal agencies and by the primes subcontracting under them.

    On the record

    Solicitation number, issuing office, NAICS and PSC codes, set-aside type, response date, and the full attachment package.

  • State portals

    Opportunities

    Fifty state procurement systems, each with its own posting format and its own idea of a due date.

    On the record

    Normalized buyer, category, closing date and time zone, plus a link back to the posting exactly as it was published.

  • Local portals

    Opportunities

    Cities, counties, school districts, transit authorities and special districts — the most fragmented tier there is.

    On the record

    Jurisdiction, department, bid number, pre-bid meeting and question deadline, and whether a plan holders list exists.

  • Web search

    Opportunities

    Buyer websites that post a solicitation as a PDF on a page nobody indexes and no portal carries.

    On the record

    The originating page, the document it linked to, and the crawl date — so a record found this way says how it was found.

  • News

    Signals

    Coverage of budget approvals, capital plans, leadership changes and awards that have not been posted yet.

    On the record

    The article, its publisher, its date, and the buyer the coverage points at.

  • Public databases

    Signals

    Award histories, contract registers, spending disclosures and published budget documents.

    On the record

    Incumbent, award value, period of performance, option years exercised, and the expiry the record implies.

  • Social signals

    Signals

    Public posts from buying organizations — hiring, programme launches, vendor days, conference agendas.

    On the record

    The post, its author’s organization, and what it suggests is being planned.

Which specific systems an organization monitors is configured per account. Nothing on this page asserts a live connection to any particular third-party system.

How it works

Discover, qualify, respond — and let the outcome inform the next one.

Thirteen stages, grouped into the three things a bid team does. Every stage is a real state a record can be in, not a step in a diagram.

  1. 01

    Discover

    Sources are monitored continuously, records are deduplicated into one per opportunity, and everything is compared against your capability profiles rather than a keyword list.

    Stages
    • discover
    • match
  2. 02

    Qualify

    Qualification requirements are read out of the solicitation, the gates are checked against what your organization can prove, and the bid decision is recorded with its reasoning.

    Stages
    • qualify
    • decide
  3. 03

    Respond

    The package is ingested and reduced to requirements, an outline is planned from the solicitation itself, sections are drafted and reviewed, and the export is a submission package.

    Stages
    • ingest
    • plan
    • draft
    • review
    • approve
    • export
    • submit
and back again
  • track
  • learn

Submissions, awards and losses are recorded against the opportunity they came from, so the next match score and the next Go/No-Go are argued from your own history.

Opportunity matching

A score you can argue with, or it is not worth having.

Matching compares an opportunity against your capability profile, your past performance, the vehicles you hold and the capacity you have. The total is the sum of named contributions, and all of them are on screen — including the ones that cost points.

Match score · SOL-8827-QDemonstration · fixture data
SOL-8827-Qtype: opportunityFederal sources

Managed detection and response services

Buyer
Federal civilian agency — IT operations division
Closes
Days remaining
24
Fit score

82/ 100

Recommended for review

The threshold that puts an opportunity in front of a person is yours to set, per profile. Nothing is discarded — a low score is a row further down, not a record you never see.

  1. Capability overlap34 / 35

    Managed detection and response is a named service in your capability profile, with the tooling listed.

  2. Past performance18 / 25

    Three prior awards in the same NAICS code. None with this agency, which is what the missing seven points are.

  3. Contract vehicle12 / 15

    You hold a schedule contract the solicitation permits. A direct award would score the full fifteen.

  4. Set-aside eligibility10 / 10

    Small business set-aside. Your registration and size standard both qualify on the response date.

  5. Geography5 / 5

    Remote delivery is permitted, and you have staff in the two time zones the coverage window needs.

  6. Capacity3 / 10

    Two of the four analysts this scope needs are already committed to another pursuit closing the same week.

Weights are per capability profile. An organization bidding through two different practices scores the same opportunity twice, because the two answers are genuinely different.

Go/No-Go

The decision that costs most when it is wrong, written down.

Qualification gates are read out of the solicitation and checked against what your organization can actually prove. The verdict is recorded with the reasoning behind it, because a pipeline of verdicts with no arguments attached teaches nobody anything afterwards.

Qualification gates · SOL-8827-QDemonstration · fixture data
  • Mandatory certificationsTwo are required. Both are on file and neither expires before the period of performance begins.verdict: Pass
  • Contract vehicleThe solicitation permits the schedule contract you already hold.verdict: Pass
  • Past performance referencesThree references of similar size and scope are required. Three are available.verdict: Pass
  • Submission window24 days remain. Your last four responses of this size each took longer than that from kickoff to submission.verdict: Watch
  • Staffing capacityThe named-key-personnel requirement collides with a pursuit closing the same week. A subcontractor would clear it.verdict: Watch
  • Incumbent positionThe scope is unchanged from the previous award and the incumbent has held it through two option years.verdict: Risk
  • Protest and exclusion historyNothing found in the public record. Absence of a finding is not a finding.verdict: Not found

A gate with nothing behind it is reported as “not found”, never as a pass. Absence of a finding is not a finding.

Recorded decisionDemonstration · fixture data
Bid, with conditionsCapture lead · 5 Aug 2026

The incumbent is the real risk, not the qualification gates: the scope has not moved, so this is a displacement bid and the technical approach has to earn the change. Worth doing because the vehicle is already in place and the certifications are current.

Conditions
  • A named subcontractor covering surge capacity before the technical volume is drafted.
  • A displacement argument in the executive summary, or withdraw at the pricing review.
  • Questions submitted by the deadline on the two ambiguous service-level clauses.

Stored against the opportunity, not in somebody’s inbox. When the outcome lands, the reasoning is still there to be compared against it.

Document intelligence

Every requirement, traced to the page it came from.

A solicitation package is ingested, read and reduced to a requirement list, an evaluation-criteria list and a set of cross-references. Nothing is extracted without a citation, because an extraction with no citation is an assertion.

Package · SOL-8827-QDemonstration · fixture data
  • SOL-8827-Q — solicitationPDF · 96 pages
  • Attachment J-1 — performance work statementPDF · 41 pages
  • Attachment J-3 — pricing workbookXLSX · 6 pages
  • Amendment 0002 — revised due date and Q&APDF · 12 pages
Requirements found142
Mandatory (shall / must)96
Evaluation criteria4
Cross-references resolved31
Evaluation criteriaDemonstration · fixture data
  • Technical approach40% / 100%

    Section M.3.1

  • Past performance25% / 100%

    Section M.3.2

  • Management approach20% / 100%

    Section M.3.3

  • Price15% / 100%

    Section M.3.4

Read out of Section M rather than assumed. The outline the drafting stage plans is weighted the way the buyer says it will evaluate.

Extracted requirementsDemonstration · fixture data
  • C.3.2shallSolicitation, p. 18

    The contractor shall provide continuous monitoring coverage on a 24×7×365 basis, including federal holidays.

    Answered in2.1 Service delivery model
  • C.4.1mustPWS J-1, p. 9

    The contractor must acknowledge a Severity 1 event within fifteen minutes of detection.

    Answered in2.3 Incident response
  • L.5.3shallSolicitation, p. 61

    The technical volume shall not exceed thirty pages, excluding the compliance matrix and résumés.

    Answered inVolume constraints
  • M.2.1willSolicitation, p. 68

    Proposals will be evaluated on technical approach, past performance, management approach and price.

    Answered inEvaluation criteria
  • A.0002-4willAmendment 0002, p. 1

    The response date is extended by fourteen days; all other terms remain unchanged.

    Answered inSchedule

Amendments are read as amendments: a revised date replaces the date on the record, and the requirement that changed says which amendment changed it.

Response automation

Drafted section by section, against the requirements as written.

The outline is planned from the solicitation itself, not from a house template. Each section is drafted with a prompt shaped for what that section has to do, and every section has an owner who reviews it.

Response outline · SOL-8827-QDemonstration · fixture data
  • 1.0Executive summaryCapture lead4 reqsstate: In review
  • 2.0Technical approachSolutions architect38 reqsstate: Drafted
  • 3.0Management approachDelivery manager21 reqsstate: Drafting
  • 4.0Past performanceProposal manager9 reqsstate: Approved
  • 5.0Price narrativeContracts12 reqsstate: Not started

Section numbering follows the solicitation. The requirement count is how many extracted requirements that section is answering — the compliance matrix is generated from the same mapping.

Draft · 2.1 Service delivery modelDemonstration · fixture data
Section prompt

Answer C.3.2 and C.4.1 using the delivery model in the capability profile. Name the tooling. Do not claim a metric the past performance library cannot support.

Generated draft

Coverage is continuous: three staffed shifts across two time zones, with a documented handover at each boundary and holiday cover scheduled a quarter in advance. answers requirement C.3.2 Severity 1 acknowledgement is automated at detection rather than at triage, so the fifteen-minute obligation does not depend on an analyst being at a desk. answers requirement C.4.1

Drawn from
  • Capability profile — service delivery model
  • Past performance library — award 2024-114, shift model
  • Solicitation SOL-8827-Q, pp. 18 and 9 (PWS J-1)

The requirement markers survive into the export and into the compliance matrix. A sentence that answers C.3.2 says so, wherever it ends up.

Branded export

A submission package in your brand, laid out by code.

Your logo, colours, typefaces and template are applied on the way out. The layout is produced by code rather than drawn by a model, so the same content exports the same way every time — which is what makes a page limit something you can rely on.

Package previewDemonstration · fixture data
FormatsDOCXPDF

Rendered here as a placeholder: the mark on the cover is the space your logo goes into, not a company.

What the package containsDemonstration · fixture data
  • Cover and transmittal letter

    Your logo, colours and typeface, applied by the template.

  • Technical volume

    Sections in solicitation order, page-limited per L.5.3.

  • Compliance matrix annexe

    Every extracted requirement against the section that answers it.

  • Past performance annexe

    References in the format the solicitation asked for.

  • Résumés and forms

    Key personnel and the standard forms the package requires.

The compliance matrix is generated from the requirement mapping rather than maintained by hand, so it cannot disagree with the sections it indexes.

Collaboration and approvals

Assignment, review and approval, with a history that does not need reconstructing.

Sections have owners, reviews have gates, and everything that happened to the response is recorded against the response. Nobody has to search an inbox to find out who approved what.

Approval gatesDemonstration · fixture data
  1. Compliance check

    state: Approved
    Proposal managerAll 96 mandatory requirements mapped to a section.
  2. Technical review

    state: In review
    ReviewerTwo comments open on 2.3 Incident response.
  3. Pricing review

    state: Waiting
    Billing adminBlocked until 5.0 Price narrative leaves draft.
  4. Executive approval

    state: Waiting
    OwnerOpens when the two reviews above close.

A gate opens when the one before it closes. Roles come from the organization’s membership, so a reviewer who should not see pricing does not see the pricing review.

ActivityDemonstration · fixture data
  1. 09:12Solutions architectRedrafted 2.3 Incident response against C.4.1.
  2. 09:40ReviewerCommented on 2.3: “Name the tooling, not the vendor tier.”
  3. 10:05Proposal managerApproved 4.0 Past performance.
  4. 10:31Capture leadRecorded the bid decision with three conditions.
  5. 11:02Delivery managerAssigned 3.0 Management approach to themselves.

Times are shown in the reader’s working day for the sake of the demonstration; in the product an entry carries a full timestamp and the person who caused it.

People in this demonstration are shown by role rather than by name. Nobody here is a real person and nothing here is a real review.

Recompete predictions

Most contracts are competed again. The question is when, and whether.

Award registers, option years and expiry dates say which contracts are approaching their end. Predictions infer the likely recompete from that record, and every one of them carries the evidence it was inferred from.

Contracts approaching expiryDemonstration · fixture data
Fixture recompete predictions4 contracts
ContractBuyerExpiresConfidenceEvidence
Grounds maintenance, three-year termCounty facilities departmentconfidence: HighFinal option year exercised, no options remain, and the same scope has been competed on a three-year cycle since 2015.
Enterprise service deskState administrative agencyconfidence: HighAward register shows no extension filed, and the published IT plan names a replacement programme in the next fiscal year.
Fleet telematicsMunicipal transportation departmentconfidence: ModerateA cooperative award covering the same category is open to this buyer, which it may use instead of running its own competition.
Language interpretation servicesPublic health systemconfidence: LowTwo prior terms were extended rather than competed. The register does not show whether extension authority remains.

Confidence is a band, not a percentage. A decimal here would imply a model calibrated against outcomes, and the honest version of that claim starts with your own results feeding back in.

A prediction is a distinct object class from an opportunity. It never appears in the pipeline as something to bid, and it never carries a due date — it carries an expiry and an argument.

Integrations

It connects to the systems a proposal already touches.

A response is assembled out of things that live somewhere else — the account in the CRM, the deal file in storage, the people in the directory. The integration surface is built so those stay where they are.

  • CRM

    Push a qualified opportunity into the pipeline the sales team already works, and read the account back.

  • Document storage

    Keep the solicitation package and the exported response where the rest of the deal file lives.

  • Identity

    Single sign-on and directory-driven provisioning, so leaving the company removes the access.

  • Calendar and notification

    Question deadlines, pre-bid meetings and submission times in the calendar people actually check.

  • REST API

    Every object the interface reads — opportunities, signals, projects, requirements, exports.

  • Webhooks

    Events on the boundaries that matter: an opportunity matched, a decision recorded, an export produced.

No third-party logos here, and no claim that a particular connector is live. Which systems an organization connects is configured per account, and the API and webhooks are the surface everything else is built on.

Security

Isolation, roles and a record that only grows.

Solicitation packages, pricing and draft responses are among the most commercially sensitive documents a company produces. What follows is how the platform is built, not a list of badges.

  • Tenant isolation

    Every tenant-owned record carries the organization that owns it. There is no shared table a query can fall through into by forgetting a filter.

  • Authorization on the server

    Every decision that matters is made on the server from the real session, and never inferred from what the interface chose to render.

  • Roles, not a flag

    Eight roles — owner, admin, billing admin, manager, member, reviewer, viewer, and a platform super-admin — so a reviewer who should not see pricing does not see pricing.

  • An audit trail that only grows

    Security-relevant events are written append-only: who did what, when. Entries are added, never edited and never quietly removed.

  • Your documents stay yours

    Solicitation packages, your content library and drafted responses are used to serve your organization. They are not used to train a shared model.

  • Two surfaces, one credential

    The marketing site you are reading has no database and no session. It is a separate origin from the application, and the only thing that crosses between them is a cookie whose entire content is the character “1”.

What is not claimed: no certification or completed audit is asserted on this page. Retention periods, the subprocessor list and how to report a vulnerability belong on the security page, where they can be stated precisely and dated.

Pricing

Four plans, from one team tracking a few buyers to a firm bidding for clients.

Every plan is self-serve and starts with a free trial. What separates them is how much you are running at once — listeners, active responses, capability profiles — rather than which features are held back.

  1. 01

    Starter

    A small team tracking a defined set of buyers.

    Discovery across public sources, explained match scores, and drafting for a couple of pursuits a month.

  2. 02

    Professional

    A bid team responding regularly.

    More listeners and active responses, document intelligence at volume, and the full review and approval flow.

  3. 03

    Business

    Several pursuit teams under one roof.

    Multiple capability profiles, API access, single sign-on, and the analytics to see where responses lose.

  4. 04

    Agency

    A firm bidding on behalf of clients.

    Separate company profiles and brand templates per client, with the work kept apart.

See what each plan costs

Prices are not printed here. The pricing page reads them from the billing catalog, so there is one place a price can be wrong instead of two.

Questions

The things people ask before they start.

  • Where do the opportunities come from?

    Seven categories of source: federal publications, state procurement portals, local and special-district portals, web search over buyer sites, news coverage, public databases such as award registers and spending disclosures, and public social posts from buying organizations. Every record stores where it came from and when it was fetched, and links back to the original posting.

  • What is the difference between an opportunity, a signal and a prediction?

    An opportunity is a confirmed procurement event that has been published. A signal is evidence that demand is coming — a budget approval, a hiring pattern, a news item — and is never presented as a solicitation. A prediction is an inferred future opportunity, such as a contract approaching expiry that is likely to be re-competed. They are separate object classes in the data model, and the interface never blurs them.

  • How is a match score calculated?

    A score is the sum of named contributions — capability overlap, past performance, contract vehicle, set-aside eligibility, geography and capacity — each carrying the reason it scored what it scored. The contributions are shown alongside the total, including the ones that cost points, because a score you cannot argue with is a score you cannot use.

  • Does the AI write the whole response?

    It drafts, section by section, against the requirements extracted from the solicitation and the content in your library, with each section prompted for what that section has to do. A person owns every section, reviews it and approves it. The draft carries the requirement references it is answering, so a reviewer can check the claim rather than take it on trust.

  • Will the export match our brand?

    Your logo, colours, typefaces and template are applied to the export, and the layout is produced by code rather than drawn by a model, so the same content exports the same way every time. Packages are produced as DOCX and PDF, with the compliance matrix included as an annexe.

  • Is our data used to train models?

    No. Solicitation packages, your content library and drafted responses are used to serve your organization, and are not training data for a shared model.

  • Can it work with the systems we already use?

    There is a REST API covering every object the interface reads, and webhooks for the events worth reacting to — an opportunity matched, a decision recorded, an export produced. Connectors for CRM, document storage, identity and calendar are configured per organization.

  • What happens when the trial ends?

    Every plan is self-serve and starts with a free trial. As configured today the trial collects a card up front and converts to a paid subscription when it ends, so cancelling before that date costs nothing — and an organization gets one trial rather than one per plan. Trial length and current prices are on the pricing page, which reads them from the billing catalog rather than from numbers typed into a marketing page.

See what you are not bidding on.

Connect a source, describe what your company does, and look at what comes back — real opportunities against your own capability profile, rather than the fixtures on this page.