diff --git a/app/[lang]/(hyperjump)/oneai/page.tsx b/app/[lang]/(hyperjump)/oneai/page.tsx
index ef8afc26..fdc9d20b 100644
--- a/app/[lang]/(hyperjump)/oneai/page.tsx
+++ b/app/[lang]/(hyperjump)/oneai/page.tsx
@@ -42,7 +42,6 @@ import {
   oneaiHeroEyebrow,
   oneaiHeroHeading,
   oneaiHeroLede,
-  oneaiHeroTagline,
   oneaiHeroProofs0,
   oneaiHeroProofs1,
   oneaiHeroProofs2,
@@ -267,11 +266,6 @@ export default async function OneaiPage({ params }: OneaiPageProps) {
                     </li>
                   ))}
                 </ul>
-                <p
-                  className="mb-5 inline-block rounded-lg bg-yellow-300/15 px-3 py-1.5 text-base font-semibold text-yellow-100 ring-1 ring-yellow-300/35 ring-inset md:text-[17px]"
-                  data-testid="oneai-hero-tagline">
-                  {oneaiHeroTagline(lang)}
-                </p>
                 <Button
                   asChild
                   className="bg-hyperjump-blue hover:bg-hyperjump-blue/90 hidden h-12 rounded-full px-8 text-base font-semibold shadow-lg shadow-[#635BFF]/25 md:inline-flex">
@@ -324,13 +318,13 @@ export default async function OneaiPage({ params }: OneaiPageProps) {
       </section>
 
       <section className="bg-[#F6F8F9] py-16 md:py-24">
-        <div className="mx-auto max-w-5xl px-4 md:px-20 xl:px-0">
+        <div className="mx-auto max-w-6xl px-4 md:px-8 xl:px-0">
           <SectionReveal>
             <h2 className="text-hyperjump-black mb-10 text-center text-3xl font-semibold tracking-tight md:text-4xl">
               {oneaiCapabilitiesLabel(lang)}
             </h2>
           </SectionReveal>
-          <StaggerContainer className="grid gap-6 sm:grid-cols-2 lg:grid-cols-3">
+          <StaggerContainer className="grid gap-6 sm:grid-cols-2 lg:grid-cols-4">
             {capabilities.map(({ text, title }) => (
               <StaggerItem key={title}>
                 <article className="h-full rounded-2xl border border-black/6 bg-white p-6 shadow-sm">
diff --git a/locales/en/oneai.json b/locales/en/oneai.json
index 2f709fee..b26c0734 100644
--- a/locales/en/oneai.json
+++ b/locales/en/oneai.json
@@ -6,8 +6,7 @@
   "hero": {
     "eyebrow": "Unified enterprise AI",
     "heading": "OneAI",
-    "tagline": "It's like ChatGPT but better!",
-    "lede": "One AI platform for your entire company. On your servers, or isolated on ours. Controlled, audited, and cost-efficient.",
+    "lede": "One workspace your company owns, with the models your teams already want: OpenAI, Claude, Gemini, DeepSeek, and others. It runs only for you, on your servers or isolated on ours.",
     "proofs": {
       "0": "Centralized dashboard",
       "1": "One budget for everyone",
@@ -35,42 +34,42 @@
     "2": {
       "kicker": "03 · Capacity",
       "title": "Latest models, fast",
-      "text": "Access current AI models as they launch: OpenAI, Anthropic, DeepSeek, and more. Admins choose which models are available to each team."
+      "text": "Use OpenAI, Claude, Gemini, DeepSeek, and other models as they launch. Admins choose which ones each team can access."
     }
   },
   "capabilities": {
     "label": "What your team can do",
     "0": {
-      "title": "Read & create office files",
-      "text": "DOCX, PPTX, XLSX, and PDF, generated directly from AI commands."
+      "title": "Ask your files",
+      "text": "Upload a file in the chat, pick the document, and ask. Follow up in the same thread."
     },
     "1": {
-      "title": "Write & run code",
-      "text": "AI writes and executes code for more complex tasks."
+      "title": "Shared knowledge",
+      "text": "A VolunteerNow Helpdesk agent answers from one shared knowledge base. Nobody re-uploads the same files."
     },
     "2": {
-      "title": "Browse the web",
-      "text": "Search for up-to-date information so answers stay accurate."
+      "title": "Create/Edit Documents",
+      "text": "Word, PowerPoint, Excel, or PDF. Type the brief, then download the file."
     },
     "3": {
-      "title": "Shared knowledge",
-      "text": "Upload internal documents. Teams get answers from your company knowledge base."
+      "title": "Run code",
+      "text": "Python or JavaScript for scraping, a website proof of concept, or charts. Review what it ran."
     },
     "4": {
-      "title": "Speech to text",
-      "text": "Turn voice into text, ideal for meeting notes and documentation."
+      "title": "Web search",
+      "text": "Turn web search on when an answer needs the live web. Off, and the model stays at its knowledge cutoff."
     },
     "5": {
-      "title": "Text to speech",
-      "text": "Convert text to natural speech for audio content and virtual assistants."
+      "title": "PII Handling",
+      "text": "National ID, email, and more. Detect it, redact it, or block the message."
     },
     "6": {
-      "title": "Scheduled automation",
-      "text": "Tell a selected model what to do: hourly, daily, weekly, or on your own schedule."
+      "title": "Budget",
+      "text": "Limits by provider, virtual key, model, organization, team, or user. Cap by cost, tokens, or requests."
     },
     "7": {
-      "title": "PII Redaction",
-      "text": "Automatically detect and mask sensitive data such as NIK, NPWP, phone and email before it reaches the model, with audit logs."
+      "title": "Multiple Providers",
+      "text": "Connect the providers you already use. Everyone works in one workspace."
     }
   },
   "comparison": {
diff --git a/locales/id/oneai.json b/locales/id/oneai.json
index 6347f681..da9ea38a 100644
--- a/locales/id/oneai.json
+++ b/locales/id/oneai.json
@@ -6,8 +6,7 @@
   "hero": {
     "eyebrow": "AI enterprise dalam satu platform",
     "heading": "OneAI",
-    "tagline": "Seperti ChatGPT, tapi lebih baik!",
-    "lede": "Satu platform AI untuk semua tim di perusahaan, berjalan di server Anda sendiri (on-premise). Terkendali, teraudit, dan hemat biaya.",
+    "lede": "Satu workspace milik perusahaan Anda, dengan model yang tim sudah mau pakai: OpenAI, Claude, Gemini, DeepSeek, dan lainnya. Berjalan hanya untuk Anda, di server Anda atau terisolasi di server kami.",
     "proofs": {
       "0": "Dashboard terpusat",
       "1": "Satu anggaran untuk semua",
@@ -35,42 +34,42 @@
     "2": {
       "kicker": "03 · Kapasitas",
       "title": "Model terbaru, tanpa menunggu",
-      "text": "Pakai model AI terkini begitu diluncurkan: OpenAI, Anthropic, DeepSeek, dan lainnya. Admin bebas menentukan model mana yang bisa diakses tiap tim."
+      "text": "Pakai OpenAI, Claude, Gemini, DeepSeek, dan model lain begitu tersedia. Admin yang menentukan model mana yang bisa diakses tiap tim."
     }
   },
   "capabilities": {
     "label": "Yang bisa dikerjakan tim Anda",
     "0": {
-      "title": "Baca & buat file kantor",
-      "text": "Buat DOCX, PPTX, XLSX, hingga PDF langsung lewat perintah ke AI secara native."
+      "title": "Tanya file",
+      "text": "Unggah file di chat, pilih dokumennya, lalu tanya. Lanjutkan di percakapan yang sama."
     },
     "1": {
-      "title": "Tulis & jalankan kode",
-      "text": "AI menulis dan mengeksekusi kode dalam terminal yang disiapkan untuk masing-masing indvidu sehingga tugas yang lebih kompleks dapat dilakukan tanpa tambahan harness coding seperti Claude code, OpenCode, dan lain-lain."
+      "title": "Pengetahuan bersama",
+      "text": "Agen VolunteerNow Helpdesk menjawab dari satu basis pengetahuan. Tidak perlu unggah ulang per orang."
     },
     "2": {
-      "title": "Jelajahi internet",
-      "text": "Cari informasi terbaru supaya jawabannya selalu akurat dan mutakhir."
+      "title": "Buat/Edit Dokumen",
+      "text": "Word, PowerPoint, Excel, atau PDF. Tulis brief-nya, lalu unduh filenya."
     },
     "3": {
-      "title": "Basis pengetahuan bersama",
-      "text": "Unggah dokumen internal, lalu tim bisa langsung mendapat jawaban dari basis pengetahuan perusahaan."
+      "title": "Jalankan kode",
+      "text": "Python atau JavaScript untuk scraping, bukti konsep website, atau grafik. Anda bisa meninjau apa yang dijalankan."
     },
     "4": {
-      "title": "Speech to text",
-      "text": "Ubah rekaman suara jadi teks. Pas untuk notulen rapat dan dokumentasi."
+      "title": "Pencarian web",
+      "text": "Nyalakan pencarian web kalau jawaban butuh internet. Kalau mati, model tetap di batas pengetahuannya."
     },
     "5": {
-      "title": "Text to speech",
-      "text": "Ubah teks jadi suara yang natural untuk konten audio maupun asisten virtual."
+      "title": "Penanganan PII",
+      "text": "NIK, email, dan data sensitif lain. Deteksi, samarkan, atau blokir pesannya."
     },
     "6": {
-      "title": "Otomasi terjadwal",
-      "text": "Serahkan tugas rutin ke model pilihan Anda: tiap jam, harian, mingguan, atau sesuai jadwal yang Anda tentukan."
+      "title": "Anggaran",
+      "text": "Batas per penyedia, virtual key, model, organisasi, tim, atau pengguna. Batasi biaya, token, atau jumlah permintaan."
     },
     "7": {
-      "title": "Redaksi PII",
-      "text": "Deteksi dan samarkan data sensitif seperti NIK, NPWP, no. HP dan email otomatis sebelum dikirim ke model, lengkap dengan audit log."
+      "title": "Multiple Providers",
+      "text": "Sambungkan penyedia yang sudah Anda pakai. Semua orang bekerja di satu workspace."
     }
   },
   "comparison": {
diff --git a/public/documents/oneai-leadership-memo-en.docx b/public/documents/oneai-leadership-memo-en.docx
index ceeb7b53..0724bd58 100644
Binary files a/public/documents/oneai-leadership-memo-en.docx and b/public/documents/oneai-leadership-memo-en.docx differ
diff --git a/public/documents/oneai-leadership-memo-id.docx b/public/documents/oneai-leadership-memo-id.docx
index 4d1c1328..4d0feccd 100644
Binary files a/public/documents/oneai-leadership-memo-id.docx and b/public/documents/oneai-leadership-memo-id.docx differ
diff --git a/public/documents/oneai-promo-en.pdf b/public/documents/oneai-promo-en.pdf
index 99d57eda..011b31ff 100644
Binary files a/public/documents/oneai-promo-en.pdf and b/public/documents/oneai-promo-en.pdf differ
diff --git a/public/documents/oneai-promo-id.pdf b/public/documents/oneai-promo-id.pdf
index 27e55002..5d568ec8 100644
Binary files a/public/documents/oneai-promo-id.pdf and b/public/documents/oneai-promo-id.pdf differ
diff --git a/scripts/generate-oneai-leadership-memos.py b/scripts/generate-oneai-leadership-memos.py
index 736f2c15..5c601446 100644
--- a/scripts/generate-oneai-leadership-memos.py
+++ b/scripts/generate-oneai-leadership-memos.py
@@ -47,9 +47,9 @@
     "request_heading": "1. Request",
     "request": (
         "I recommend we evaluate OneAI from Hyperjump Technology as our "
-        "company-wide AI platform, or at minimum schedule a 30-minute briefing. "
-        "This is not about replacing personal ChatGPT for a few people. It is "
-        "about governing AI now that 20-40 of us already use it."
+        "company AI platform: one workspace for OpenAI, Claude, Gemini, "
+        "DeepSeek, and other models, running only for us. Please at least "
+        "schedule a 30-minute briefing."
     ),
     "why_heading": "2. Why this matters now",
     "why_intro": (
@@ -71,7 +71,7 @@
         ),
         (
             "Real capability",
-            "office files, internal knowledge, code execution, and more than one model vendor.",
+            "ask files, a VolunteerNow Helpdesk knowledge base, documents, code, web search, PII handling, budgets, and multiple providers in one workspace.",
         ),
     ],
     "cost": (
@@ -126,9 +126,8 @@
     "request": (
         "Saya mengusulkan agar perusahaan mengevaluasi OneAI dari Hyperjump "
         "Technology sebagai platform AI terpadu, atau setidaknya menjadwalkan "
-        "briefing 30 menit. Ini bukan usulan mengganti ChatGPT pribadi untuk "
-        "beberapa orang. Ini usulan mengatur AI begitu 20-40 orang di "
-        "perusahaan sudah memakainya."
+        "briefing 30 menit. OneAI adalah satu workspace untuk OpenAI, Claude, "
+        "Gemini, DeepSeek, dan model lain, yang berjalan hanya untuk perusahaan kita."
     ),
     "why_heading": "2. Mengapa ini penting sekarang",
     "why_intro": (
@@ -150,7 +149,7 @@
         ),
         (
             "Kapabilitas yang nyata",
-            "file kantor, pengetahuan internal, eksekusi kode, dan lebih dari satu vendor model.",
+            "tanya file, basis VolunteerNow Helpdesk, dokumen, kode, pencarian web, penanganan PII, anggaran, dan multiple providers dalam satu workspace.",
         ),
     ],
     "cost": (
diff --git a/scripts/generate-oneai-promo-pdfs.py b/scripts/generate-oneai-promo-pdfs.py
index 72a513c3..13406c5a 100644
--- a/scripts/generate-oneai-promo-pdfs.py
+++ b/scripts/generate-oneai-promo-pdfs.py
@@ -33,13 +33,12 @@
     "file": "oneai-promo-en.pdf",
     "place": "JAKARTA · INDONESIA",
     "eyebrow": "UNIFIED ENTERPRISE AI",
-    "lede": "One AI platform for your entire company. On your servers, or isolated on ours.",
+    "lede": "One workspace your company owns, with the models your teams already want: OpenAI, Claude, Gemini, DeepSeek, and others. It runs only for you, on your servers or isolated on ours.",
     "proofs": [
         "Centralized dashboard",
         "One budget for everyone",
         "Latest models on release",
         "On-prem, or isolated hosting",
-        "Jakarta-based support",
     ],
     "pillars": [
         (
@@ -55,17 +54,19 @@
         (
             "03 · CAPACITY",
             "Latest models, fast",
-            "Access current AI models as they launch: OpenAI, Anthropic, DeepSeek, and more. Admins choose which models each team can use.",
+            "Use OpenAI, Claude, Gemini, DeepSeek, and other models as they launch. Admins choose which ones each team can access.",
         ),
     ],
     "cap_label": "WHAT YOUR TEAM CAN DO",
     "caps": [
-        ("Read & create office files", "DOCX, PPTX, XLSX, and PDF from AI commands."),
-        ("Write & run code", "AI writes and executes code for complex tasks."),
-        ("Browse the web", "Search for up-to-date information."),
-        ("Shared knowledge", "Answers from your company documents."),
-        ("Speech to text", "Voice into meeting notes and documentation."),
-        ("Text to speech", "Natural speech for audio and assistants."),
+        ("Ask your files", "Upload in the chat, ask, and follow up in the same thread."),
+        ("Shared knowledge", "VolunteerNow Helpdesk uses one shared base. No re-upload per person."),
+        ("Create/Edit Documents", "Word, PowerPoint, Excel, or PDF. Type the brief, then download."),
+        ("Run code", "Python or JavaScript for scraping, a site POC, or charts. Review what it ran."),
+        ("Web search", "On for the live web. Off, and the model stays at its knowledge cutoff."),
+        ("PII Handling", "National ID, email, and more. Detect, redact, or block."),
+        ("Budget", "Cap by provider, key, model, org, team, or user: cost, tokens, or requests."),
+        ("Multiple Providers", "Connect the providers you already use. One workspace."),
     ],
     "plan_kicker": "PLAN · UP TO 40 USERS",
     "price": "Rp12.400.000 / month",
@@ -78,7 +79,6 @@
         "Funds may not be spent on AI",
         "Extra admin overhead, higher cost",
         "No centralized control",
-        "Data leaves on 40 personal accounts",
     ],
     "p2_kicker": "ONEAI · COMPARISON",
     "p2_eyebrow": "WHY A CHATGPT SEAT IS NOT THE SAME PRODUCT",
@@ -129,13 +129,12 @@
     "file": "oneai-promo-id.pdf",
     "place": "JAKARTA · INDONESIA",
     "eyebrow": "AI ENTERPRISE DALAM SATU PLATFORM",
-    "lede": "Satu platform AI untuk semua tim—di server Anda sendiri atau terisolasi di server kami.",
+    "lede": "Satu workspace milik perusahaan Anda, dengan model yang tim sudah mau pakai: OpenAI, Claude, Gemini, DeepSeek, dan lainnya. Berjalan hanya untuk Anda, di server Anda atau terisolasi di server kami.",
     "proofs": [
         "Dashboard terpusat",
         "Satu anggaran untuk semua",
         "Model terbaru begitu rilis",
         "On-premise, atau hosting terisolasi",
-        "Support dari tim Jakarta",
     ],
     "pillars": [
         (
@@ -151,17 +150,19 @@
         (
             "03 · KAPASITAS",
             "Model terbaru, tanpa menunggu",
-            "Pakai model AI terkini begitu diluncurkan: OpenAI, Anthropic, DeepSeek, dan lainnya. Admin menentukan model untuk tiap tim.",
+            "Pakai OpenAI, Claude, Gemini, DeepSeek, dan model lain begitu tersedia. Admin menentukan model untuk tiap tim.",
         ),
     ],
     "cap_label": "YANG BISA DIKERJAKAN TIM ANDA",
     "caps": [
-        ("Baca & buat file kantor", "DOCX, PPTX, XLSX, & PDF lewat perintah AI."),
-        ("Tulis & jalankan kode", "AI menulis dan menjalankan kode."),
-        ("Jelajahi internet", "Cari informasi terbaru yang akurat."),
-        ("Basis pengetahuan bersama", "Jawaban dari dokumen internal perusahaan."),
-        ("Speech to text", "Ubah suara jadi teks & notulen."),
-        ("Text to speech", "Suara natural untuk audio & asisten."),
+        ("Tanya file", "Unggah di chat, tanya, lanjut di percakapan yang sama."),
+        ("Pengetahuan bersama", "Agen VolunteerNow Helpdesk, satu basis. Tanpa unggah ulang per orang."),
+        ("Buat/Edit Dokumen", "Word, PowerPoint, Excel, atau PDF. Tulis brief, lalu unduh."),
+        ("Jalankan kode", "Python atau JavaScript untuk scraping, POC website, atau grafik."),
+        ("Pencarian web", "Nyala untuk web hidup. Mati, model tetap di batas pengetahuannya."),
+        ("Penanganan PII", "NIK, email, dan lainnya. Deteksi, samarkan, atau blokir."),
+        ("Anggaran", "Batas per penyedia, key, model, org, tim, atau user: biaya, token, atau permintaan."),
+        ("Multiple Providers", "Sambungkan penyedia yang sudah dipakai. Satu workspace."),
     ],
     "plan_kicker": "PAKET · HINGGA 40 PENGGUNA",
     "price": "Rp12.400.000 / bulan",
@@ -174,7 +175,6 @@
         "Dana belum tentu dipakai untuk AI",
         "Administrasi bertambah, biaya membengkak",
         "Tidak ada kontrol terpusat",
-        "Data tersebar di 40 akun pribadi",
     ],
     "p2_kicker": "ONEAI · PERBANDINGAN",
     "p2_eyebrow": "KENAPA KURSI CHATGPT ITU PRODUK BERBEDA",
@@ -368,11 +368,11 @@ def page_one(c: canvas.Canvas, copy: dict) -> None:
     c.setFont("Helvetica-Bold", 8)
     c.drawString(MARGIN, y, copy["cap_label"])
     y -= 18
-    cap_w = (PAGE_W - 2 * MARGIN - 16) / 3
+    cap_w = (PAGE_W - 2 * MARGIN - 24) / 4
     for i, (title, body) in enumerate(copy["caps"]):
-        col, row = i % 3, i // 3
+        col, row = i % 4, i // 4
         x = MARGIN + col * (cap_w + 8)
-        cy = y - row * 52
+        cy = y - row * 58
         c.setFillColor(INK)
         c.setFont("Helvetica-Bold", 9)
         c.drawString(x, cy, title)