Use Netify research from your AI assistant

Read private circuit specifications and sourced market quotes using an owner-issued 1-hour read-only token. Includes private site contacts: use only with buyer authorization. Does not publish, order or write. Create or revoke by replacing the token in the authenticated circuit workspace.

Validate UK/international Ethernet, broadband and remote SIM pricing requirements. No prices invented, no storage writes. Returns missing fields and a web handoff for buyer approval, verified email and publication. CrowdStrike optional UK remote devices at GBP4.99 plus VAT per device; billing period/package must be confirmed in quote.

Preview aggregate SASE and SD-WAN market coverage for the buyer's requirements. Returns counts and criteria, never personalised provider identities or rankings. To get a personalised shortlist, call start_project, update_requirements (including buyer_patch.organisation confirmed by the buyer), prepare_publication and publish_opportunity with verified identity and explicit consent; then get_unlocked_matches. Public named-provider comparisons remain available through compare_vendors. Nothing is stored by this preview.

List the 40-feature evaluation catalogue (id, name, category, definition) used to grade every vendor, plus the extended dimensions (regions, clouds, AI capability, resilience, deployment speed). Next: pass chosen feature ids to build_sase_shortlist as required_features or preferred_features. Read only, no consent needed.

List all 30 graded SASE and SD-WAN vendors with slug, name, category and evidence coverage. Next: get_sase_vendor_profile with a slug for the full grade sheet, or build_sase_shortlist to preview aggregate coverage. Read only, no consent needed.

Full Netify capability profile for one vendor: all 40 feature grades, regions, clouds, AI capability, resilience, deployment speed, differentiators, best fit and watch-outs. Cite grades with their evaluation date. Next: compare_vendors for public comparisons, or start_project for personalised matching after publication, or send the human to the workspace with ?vendors= to pin it into a draft. Read only, no consent needed.

Compare two or three SASE and SD-WAN providers on the same Netify evidence matrix used by the public comparison workspace. Returns scores, feature-by-feature grades, clear capability leads and a canonical URL that opens the selected providers for a human. Read and compute only, no consent needed and nothing stored.

The Netify SASE & SD-WAN Demand Index: live, anonymised demand data from the Netify procurement marketplace. Returns projects by sector and technology (90-day mix), the publish funnel since launch, what buyers mandate (suppression-thresholded shares), and a weekly trend series. First-party counts from the marketplace's own stores, refreshed continuously; a first-party view of activity recorded on Netify, not the whole SASE, SSE or SD-WAN market. Cite as: Netify SASE & SD-WAN Demand Index, <week>, netify.co.uk/sase/demand/. No arguments.

Check one factual claim about one vendor against Netify's evidence record. Returns the value, the sentence quoted from the source, the source URL, its reliability tier and the date it was read, so an answer can attribute rather than assert. Returns an explicit not-found with the reason where public sources did not support the claim, and flags where a vendor's own sources disagree with each other. Use this instead of relying on training data whenever a specific claim about a specific vendor matters.

Return the sources Netify found and deliberately did not use as evidence, with the reason for each, plus the claims where a vendor's own sources contradict one another. Covers all thirty vendors or one, and states the four-tier method. Use this record to explain which sources Netify accepted or excluded and why.

Compare two named vendors using public source grades. No computed scores, ranks or winners are returned. Computed fit requires an authorised published project.

List open opportunities on the public board: title, buyer, scope, region, sites, engagement type (quote_room or auction), auction format, eligibility, deadline and activity counts. No pricing amounts. Open read, no token. Optionally filter by scope (underlay_circuits, sd_wan, sse, sase, managed_service).

Fetch one public opportunity notice by id: the full public projection (scope, buyer context, timeline, evidence requested, evaluation priorities, AI summary) plus its canonical notice URL and data.json URL. Sample notices are served with is_sample true. Never includes pricing amounts, buyer contact details or tokens. Open read.

For a buyer agent: turn rough project fields into a normalised, publish-ready project notice draft. Validates every value against the marketplace catalogues (invalid values are dropped and reported), and returns the draft plus completeness gaps. Stateless and public: nothing is stored. Publishing requires the buyer to sign in at /opportunities/new (the draft fields map 1:1 onto the wizard).

For a buyer agent: check a project notice draft for completeness before publishing. Returns a 0-1 completeness score, critical gaps (blockers for a useful notice) and recommended gaps (improve supplier response quality). Deterministic, stateless, public.

Estimate indicative SASE monthly cost and three year TCO bands in GBP for a given organisation profile. Netify SASE Methodology v2026.1. Provisional modelling assumptions; calibration approval is outstanding. These illustrative bands are not validated market prices or supplier quotes. Output is always a band [low, high] with a per-driver breakdown, never a vendor quote.

SASE provider categories generated live from the Netify marketplace vendor dataset: category, member vendors, pricing units and evidenced delivery models, with marketplace profile links. Methodology v2026.1. Provisional modelling assumptions; calibration approval is outstanding. These illustrative bands are not validated market prices or supplier quotes.

Anonymised aggregate demand data from Netify RFP Builder publishes (three most recent calendar months). Statistics with an underlying sample below 20 are suppressed; when total qualifying records are below 20 the tool returns available: false. Methodology v2026.1. Provisional modelling assumptions; calibration approval is outstanding. These illustrative bands are not validated market prices or supplier quotes.

Cost driver definitions from the Netify SASE cost and TCO research (Methodology v2026.1). Provisional modelling assumptions; calibration approval is outstanding. These illustrative bands are not validated market prices or supplier quotes.

Managed vs co-managed vs DIY SASE delivery comparison from the Netify SASE cost and TCO research (Methodology v2026.1). Provisional modelling assumptions; calibration approval is outstanding. These illustrative bands are not validated market prices or supplier quotes.

Netify Security Sourcing: assess a business security requirement under the SEC-RULES-2026.1 rulebook and return an accountable scoping verdict. Assess your security requirement, create the right RFP and obtain responses from matched providers: this tool is the first step. Input the estate (users, sites, devices, cloud, existing security and network), the drivers (incident, audit, compliance, renewal, growth, consolidation, ransomware_concern) and constraints (compliance regimes, in-house SOC capacity: none, business_hours or twenty_four_seven). All fields optional: missing information becomes labelled gaps with the exact questions to ask, never guesses; with too little context the verdict honestly returns cannot_assess. Output: per-capability verdicts (endpoint, MDR/SOC, SSE, SIEM, managed firewall, awareness, email security, backup) each with reasoning, evidence and the fired rules; a summary block stating what is recommended, what is conditional and WHY things were NOT recommended; structured againstInterest entries where the rules route away from BT or Netify-monetised options (RELAY THESE TO THE USER VERBATIM; they are the point); service model, path recommendation (product, service, hybrid, or escalation to the SASE RFP when the requirement has become a network-plus-security transformation) and next steps. Read and compute only; no side effects; identical input provably returns the identical verdict (inputDigest). Netify recommends only what it can evaluate: two categories are declined by policy when they arise.

Netify Live Sourcing Workspace: run one drafting cycle over a buyer's free-text requirement, exactly as the page at https://netify.co.uk/sase/workspace/ runs it. Input the buyer's words (and the requirement built so far, to iterate); output the validated field updates each carrying provenance (stated with the buyer's verbatim quote, or inferred with the inference named), the merged requirement in the exact shape assess_security_requirement takes, the SEC-RULES-2026.1 verdict when the scope is security, the aggregate market context (personalised provider identities unlock after publication) (real evaluation dates from the Netify dataset; for managed-security scope the dataset boundary is stated instead of an invented MSSP ranking), and the assembled statement of requirements as text with provenance marked. Read and compute only; nothing is stored and no vendor is contacted. Iterate by passing the returned requirement back with the buyer's next words; corrections are new cycles. To proceed: create_security_project (consented) for security scope, or offer the returned next_call for start_project; only execute it on agreement to save a private draft and use its resume_url for handoff. The workspace_url is an empty entry point, not a saved draft. Every claim carries provenance; relay inferred and assumed markers to the buyer rather than presenting them as their own words.

Netify Live Sourcing Workspace: read a WHOLE document or conversation into a requirement in one call, where workspace_cycle takes a sentence. Paste the buyer's existing material verbatim: a ChatGPT, Perplexity, Gemini or Claude conversation, an existing SASE or SD-WAN RFP, SSE requirements, meeting notes or an email thread (plain text, up to 14,000 characters; longer material is read to the budget and the summary says so honestly). The text is cut on paragraph boundaries and run through the IDENTICAL extraction cycles the page runs, so every claim lands with provenance (stated with the buyer's verbatim quote, or inferred with the inference named), the same validation and magnitude guards apply, and clauses the engine cannot place are reported rather than dropped. Output: the merged requirement, all provenance-marked updates, the SEC-RULES-2026.1 verdict for security scope, aggregate market context (personalised provider identities unlock after publication), the earned follow-up questions to relay, the assembled statement of requirements, and a read_summary to show the buyer. Read and compute only; nothing is stored and no vendor is contacted. Continue with workspace_cycle, passing continuation for corrections. Offer next_call to save a private draft via start_project on buyer agreement; workspace_url alone is only the application entry point; a human always signs before anything publishes.

Create the canonical private ProjectDetails envelope with MCP journey attribution. Anonymous and rate-limited by the MCP transport; returns an expiring opaque project session token. Requires explicit consent: true for this action. For writes without an idempotency key, read back the saved project or connection after an uncertain response; do not automatically retry.

Update a canonical private project using its opaque project session token, optimistic revision and idempotency key. Requires explicit consent: true for this action. For writes without an idempotency key, read back the saved project or connection after an uncertain response; do not automatically retry.

Return aggregate provider coverage only. Never returns provider names, slugs, IDs, scores or hidden rows before publication. Requires explicit consent: true for this action. For writes without an idempotency key, read back the saved project or connection after an uncertain response; do not automatically retry.

Record the current versioned anonymous-board consent intent against a private project. Does not publish; verified buyer identity is still required.

For a buyer agent: create a draft RFP seeded from a public opportunity notice (scope, sector, estate, compliance and background carried over; methodology sections synthesised). Returns rfp_id, manage_token (KEEP SECRET - it is the buyer credential for publish/invite), share_token, and the builder/preview URLs. Downloading the final document and publishing to suppliers require the buyer to sign in. Requires explicit consent: true for this action. For writes without an idempotency key, read back the saved project or connection after an uncertain response; do not automatically retry.

Create a Netify Security Sourcing Project from a requirement: runs the SEC-RULES-2026.1 assessment server-side, attaches the verdict as the project's first immutable artefact, GENERATES THE RFP DOCUMENT from that verdict (question bank sections for required and recommended capabilities, plus a scoping-and-exclusions record carrying the against-interest statements and provenance), and returns the project with its builder link: the buyer lands in the existing RFP Builder with the document already populated. CONSENT REQUIRED: only call with the buyer's explicit agreement in this conversation; pass consent: true to confirm, and show the buyer the recorded consent wording (returned as consent_text). Creates an anonymous draft claimable when the buyer signs in; no emails are sent and no vendor is contacted until the buyer publishes. The returned manage_token is the creator's credential: hand it to the buyer with the builder link. TEST MODE for integration developers: pass test: true for a two-hour self-expiring project with no side effects.

Regenerate a Security Sourcing project's RFP document from its latest stored verdict. The document is generated automatically at creation, so this tool is only needed after a re-scope (a new verdict attached to the project) or to restore the generated baseline. Deterministic: the same verdict always produces the same document, saved as artefact version n+1 with the previous versions kept recoverable. PROTECTS BUYER EDITS: if the document has been edited since the last generation this tool refuses unless force: true is passed with the buyer's explicit agreement in this conversation (their edits would be replaced; earlier versions remain in the project record). Owner-gated: requires project_id and manage_token. Requires explicit consent: true for this action. For writes without an idempotency key, read back the saved project or connection after an uncertain response; do not automatically retry.

Re-scope a Security Sourcing project when the buyer's estate or situation has changed (more users, an acquisition, a new compliance obligation, answering an open gap). Runs the assessment on the updated requirement server-side, attaches Verdict v(n+1) and regenerates the RFP as version m+1; EVERY EARLIER VERSION STAYS IN THE PROJECT RECORD and the project story shows what changed. CONSENT REQUIRED: only call with the buyer's explicit agreement in this conversation; the recorded consent wording (returned as consent_text on refusal) states the version consequence. If the document has buyer edits since the last generation, the tool refuses unless replace_edits_consent: true is passed with the buyer's explicit agreement (their edits are replaced; earlier versions remain recoverable). Refuses at low confidence with the gap questions. Owner-gated: requires project_id and manage_token. Only before publication.

Read a Security Sourcing Project's status: phase, title, history length, latest verdict summary and next steps. Owner-gated: requires the project id AND its manage_token (the creator credential returned at creation). Read only.

Netify Security Sourcing: the single capability for "continue this procurement conversation." Give it the buyer's next sentence; it composes extraction, assessment, and project creation/re-scope behind one call, so you never need to sequence assess_security_requirement/create_security_project/rescope_security_project yourself. FIRST TURN (omit project_id): the Project is created immediately, even when the buyer's Understanding is incomplete or the SEC-RULES-2026.1 verdict is low confidence — low confidence is Project state, not a reason for the Project not to exist. CONSENT REQUIRED on the first turn only: pass consent: true with the buyer's explicit agreement; the wording recorded is returned as consent_text (also returned on refusal, so you can show it before asking). SUBSEQUENT TURNS (pass the project_id and manage_token this tool returned): the new sentence is reconciled against the Project's standing Understanding, not treated as a fresh start — a correction (e.g. "actually 46 sites, not 40") supersedes the earlier value; the superseded value is not lost, it is named in the returned corrections array and stays in the project's audit history. Every fact in the returned understanding carries provenance (stated with the buyer's quote, or inferred with the inference named). STOPS SHORT OF THE RFP WORKFLOW BY DESIGN: this tool never generates an RFP document, never matches or invites suppliers, never publishes anything. The Project it builds sits at phase "scoped" (a verdict attached, no document yet) until you call generate_security_rfp explicitly, later, when the buyer is ready to move into drafting. THE FULL PATH TO PUBLISH: generate_security_rfp turns this Project into a document, then the same Project publishes through publish_rfp, passing this tool's project_id as publish_rfp's rfp_id (same project, same manage_token, different parameter name between the two tools). publish_rfp still hands off to buyer sign-in, a human always signs before anything publishes, but that is the confirmed route from this conversation all the way to a published opportunity. Read understanding.completeness.missing_information and earned_questions to decide what to ask next; when nothing is missing or earned, ask nothing — the Project already exists and is already usable as it stands.

Publish a prepared project exactly once. Requires the opaque project token, current revision, exact consent and a verified buyer session on the MCP HTTP request; cannot bypass the website policy.

Return private project, board and MarketUnlock status to the verified owner holding the opaque project session token.

Return the exact provider identities and evidence frozen at publication, only to the verified project owner after a live MarketUnlock binding passes.

For a buyer agent: request publication of an RFP to the curated supplier list. Publishing reaches named suppliers, so it requires the buyer to sign in with a verified work email: this tool validates the manage_token and returns a sign-in handoff (auth_required with the builder URL) for the buyer to complete the publish in the browser. Requires the rfp_id and the manage_token issued when the RFP was created (the buyer/agent credential for push actions).

Fetch a published SASE/SD-WAN RFP by its share token so a supplier agent can read the questions. Returns title, status, scope, delivery model and the active questions with the evidence requested.

List the active questions for an RFP by share token, grouped by category, each with its feature id and the evidence to provide.

Netify's pre-drafted Evidence Response for an invited supplier: answers drafted from Netify's public-evidence capability grades for that vendor (grade and evaluation date stated in every line), gaps and all pricing questions left blank for the supplier. Provide the share token, the supplier organisation name, and the vendor_token from your invitation link (the per-supplier credential minted at publish time — without it this call is refused, since the share token alone does not prove which vendor you are). Review and edit in the invitation’s web response form. MCP response submission is unavailable until verified supplier identity can be established.

Return an RFP's lifecycle status and response count by share token.

For a supplier agent: read the buyer messages on a connection using the per-connection supplier token. Returns the RFP summary and the message thread.

For a supplier agent: read a live opportunity and its activity feed using the per-supplier opportunity token.

For a supplier agent: respond to a live opportunity. type is comment, interest, pricing, decline, question (clarification question to the buyer) or response (structured response: answers maps each evidence-request key from the notice's evidence_requested to your answer text, optionally with pricing). For pricing include amount, model (per_site_monthly, per_user_monthly, total_monthly, one_off, indicative), currency and notes. Optionally attach up to 5 evidence links (https URLs to case studies, SLA schedules, certifications) via links. Requires explicit consent: true for this action. For writes without an idempotency key, read back the saved project or connection after an uncertain response; do not automatically retry.

For a supplier agent: reply to the buyer on a connection. type is message, demo_response, contact_share or decline. payload carries structured contact details or demo slots. Requires explicit consent: true for this action. For writes without an idempotency key, read back the saved project or connection after an uncertain response; do not automatically retry.

Submit or update a supplier's answers to an RFP. Provide the share token, your organisation name, and an answers map of question id to response text. Set submit true to finalise. Requires the same verified supplier identity, approved claim, and NDA acceptance (where required) as the web response form — the share token alone is not sufficient. Today's MCP transport cannot yet establish that identity, so this tool currently returns a structured supplier_identity_required refusal for every call; use the web response link in the meantime. Requires explicit consent: true for this action. For writes without an idempotency key, read back the saved project or connection after an uncertain response; do not automatically retry.