{"id":1000,"date":"2026-11-07T07:15:19","date_gmt":"2026-11-07T07:15:19","guid":{"rendered":"https:\/\/rimsift.com\/?page_id=1000"},"modified":"2026-11-07T07:15:19","modified_gmt":"2026-11-07T07:15:19","slug":"home","status":"publish","type":"page","link":"https:\/\/rimsift.com\/","title":{"rendered":"Home"},"content":{"rendered":"

Draft copy imported from the RimSift designer handover. Codes such as G4 refer to the graphics list. Text in [square brackets] is a placeholder or an open decision. Replace this block with the designed layout.<\/em><\/p>\n

Page title (browser tab):<\/strong> RimSift | The control point for agentic AI Meta description:<\/strong> RimSift checks every AI tool call and chat: who is asking, what they may do, what’s hidden inside, and a full record of what happened.<\/p>\n

6.1 Hero<\/h3>\n
    \n
  • Eyebrow: THE CONTROL POINT FOR AGENTIC AI<\/li>\n
  • Headline: Every AI tool call, checked at the rim<\/strong>. (“rim” in amber)<\/li>\n
  • Sub-headline: RimSift sits between your AI agents and the tools and models they use. It checks who is asking, enforces your policy, inspects every call and keeps a complete record, so your teams can use AI without losing control.<\/li>\n
  • Buttons: Book a demo \u00b7 See how it works (scrolls to 6.5)<\/li>\n
  • Visual: G1<\/strong> rim animation, right half on desktop, below the text on mobile.<\/li>\n<\/ul>\n

    6.2 Proof strip<\/h3>\n

    One row directly under the hero. Design two versions:<\/p>\n

      \n
    • Now:<\/strong> three short statements with icons: “Controls mapped to the OWASP MCP Top 10” \u00b7 “Deploy in the UK, the EU or your own cloud” \u00b7 “Built by the team behind Parrera”.<\/li>\n
    • Later (hidden):<\/strong> greyscale customer-logo row, “Trusted by security teams at” + up to 8 logos.<\/li>\n<\/ul>\n

      6.3 The problem<\/h3>\n
        \n
      • Eyebrow: THE PROBLEM<\/li>\n
      • Headline: Agents connect. Then nobody watches.<\/li>\n
      • Body: Your teams are already wiring AI agents into email, code, customer records and internal systems. Most security tools decide whether an agent may connect, then stop looking. What the agent does next, which tool it calls, what data leaves and whether someone slipped it an instruction, goes unchecked.<\/li>\n
      • Three cards (“The options today”):
        \n 1. Block AI.<\/strong> People use it anyway, on personal accounts you can’t see.
        \n 2. Allow it.<\/strong> You approve the tool once and have no idea what it does after.
        \n 3. Review everything by hand.<\/strong> Your team becomes the bottleneck and still misses the calls that matter.<\/li>\n
      • Line under the cards: “There’s a better option: check every call, automatically.”<\/li>\n<\/ul>\n

        6.4 Four checks (tabbed)<\/h3>\n
          \n
        • Eyebrow: THE PLATFORM<\/li>\n
        • Headline: One control point. Four checks on every call.<\/li>\n
        • Four tabs; each shows a short paragraph, three bullets and a mock screen.<\/li>\n<\/ul>\n\n\n\n\n\n\n\n\n
          Tab<\/th>\nParagraph<\/th>\nBullets<\/th>\nVisual<\/th>\n<\/tr>\n<\/thead>\n
          Identity<\/td>\nEvery call is tied to a real person and the agent acting for them. Tools see the user, not a shared service account.<\/td>\nWorks with your existing identity provider \u00b7 Per-user access to each tool \u00b7 No raw API keys on laptops<\/td>\nG6<\/td>\n<\/tr>\n
          Governance<\/td>\nNothing runs unless your policy allows it, and the policy is checked on every call, not once at sign-off.<\/td>\nDefault-deny tool access \u00b7 Different tool sets for each team \u00b7 Approval step for risky actions<\/td>\nG5<\/td>\n<\/tr>\n
          Inspection<\/td>\nCalls and results are checked for hidden instructions, changed tools and sensitive data on its way out.<\/td>\nPrompt-injection detection \u00b7 Alerts when an approved tool changes \u00b7 Masks or blocks sensitive data<\/td>\nG7<\/td>\n<\/tr>\n
          Audit<\/td>\nA complete, searchable record of what every agent did, for whom and why.<\/td>\nEvery call logged with its verdict \u00b7 Sends to your SIEM \u00b7 Evidence ready for auditors<\/td>\nG9<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
            \n
          • Link under the tabs: Explore the platform \u2192<\/li>\n<\/ul>\n

            6.5 How it works<\/h3>\n
              \n
            • Eyebrow: HOW IT WORKS<\/li>\n
            • Headline: Point your agents at RimSift. Everything else stays the same.<\/li>\n
            • Three numbered steps beside G2<\/strong>:
              \n 1. Connect.<\/strong> Route your AI agents, MCP servers and models through RimSift. No change to the tools themselves.
              \n 2. Set policy.<\/strong> Choose which teams can use which tools, and what needs approval.
              \n 3. See and control.<\/strong> Every call is checked live. Safe calls carry on; risky ones are stopped or held for approval.<\/li>\n
            • G4<\/strong> live call log below the steps, full width.<\/li>\n<\/ul>\n

              6.6 The gaps we close<\/h3>\n
                \n
              • Eyebrow: WHY RIMSIFT<\/li>\n
              • Headline: Eight gaps today’s tools leave open.<\/li>\n
              • 4 \u00d7 2 grid of small cards (icon, title, one line), each linking to its row on \/why-rimsift\/:
                \n 1. Nobody watching the agent:<\/strong> approved agents act with no one checking each call.
                \n 2. Tools that change after approval:<\/strong> a tool’s description is rewritten and the agent follows it.
                \n 3. Hidden instructions:<\/strong> text inside a document or web page tells the agent what to do.
                \n 4. Harmless steps, harmful total:<\/strong> each action looks fine; together they leak data.
                \n 5. Human or agent?<\/strong> Logs can’t show who really did what.
                \n 6. Shadow AI:<\/strong> staff paste company data into AI tools you never approved.
                \n 7. Data walking out:<\/strong> sensitive data leaves inside a prompt or tool call.
                \n 8. No proof for auditors:<\/strong> no record of what the AI did, or why.<\/li>\n<\/ul>\n

                6.7 Where it runs<\/h3>\n
                  \n
                • Headline: Wherever your agents work.<\/li>\n
                • Three cards with G11<\/strong>: Hosted gateway<\/strong> (for AI apps and remote MCP servers) \u00b7 Workstation<\/strong> (for coding agents and local tools on laptops) \u00b7 Browser<\/strong> (for ChatGPT, Claude, Gemini and Copilot on the web). Line beneath: “One policy, three places, one audit trail.”<\/li>\n
                • G13<\/strong> integration strip below.<\/li>\n<\/ul>\n

                  6.8 Built for regulated and sovereign environments<\/h3>\n
                    \n
                  • Headline: Your data stays where you decide.<\/li>\n
                  • Body: Run RimSift in the UK, the EU or inside your own cloud. Prompts and tool data are inspected in the region you choose, and nothing is sent to a third-party service for checking.<\/li>\n
                  • Three small stats-style blocks (text, not numbers): “UK and EU hosting” \u00b7 “Self-hosted option” \u00b7 “Fail-closed by default”.<\/li>\n<\/ul>\n

                    6.9 Customer quote (hidden until real)<\/h3>\n

                    One large pull-quote on signal-dim background, name, role, company logo. Design it; leave it switched off.<\/p>\n

                    6.10 FAQ (four questions, accordion)<\/h3>\n

                    Use the first four questions from section 9.4.<\/p>\n

                    6.11 Latest<\/h3>\n

                    Three newest items from Resources, G16<\/strong> cards. Hidden if fewer than three.<\/p>\n

                    6.12 Closing band<\/h3>\n

                    Standard closing call-to-action band (section 5).<\/p>\n","protected":false},"excerpt":{"rendered":"

                    Draft copy imported from the RimSift designer handover. Codes such as G4 refer to the graphics list. Text in [square brackets] is a placeholder or an open decision. Replace this block with the designed layout. Page title (browser tab): RimSift | The control point for agentic AI Meta description: RimSift checks every AI tool call […]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-1000","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/rimsift.com\/index.php\/wp-json\/wp\/v2\/pages\/1000","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/rimsift.com\/index.php\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/rimsift.com\/index.php\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/rimsift.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/rimsift.com\/index.php\/wp-json\/wp\/v2\/comments?post=1000"}],"version-history":[{"count":1,"href":"https:\/\/rimsift.com\/index.php\/wp-json\/wp\/v2\/pages\/1000\/revisions"}],"predecessor-version":[{"id":1341,"href":"https:\/\/rimsift.com\/index.php\/wp-json\/wp\/v2\/pages\/1000\/revisions\/1341"}],"wp:attachment":[{"href":"https:\/\/rimsift.com\/index.php\/wp-json\/wp\/v2\/media?parent=1000"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}