What We Found
Research synthesis across 16 competitors - VN and global - covering market structure, user behavior, pricing, and strategic implications.
Findings
- 1 The entire market - both VN-made and global - is chat-first or assistant-first. Every tool surveyed requires the user to know what to ask. None offers a guided Skill-based flow to ready-to-use document output for specific Vietnamese office roles. This is not a niche gap; it is a structural absence across all competitors surveyed.
- 2 VN-localized tools have distribution but no structure. Kiki/Zalo reaches 80M users for free; AI Hay has 15M downloads. Both are chat Q&A tools. AIVA has templates but they sit on top of a chat interface - the user still drives the session. Being Vietnamese-first and being solution-first are treated as separate design choices; no one has combined them.
- 3 Global tools have deep VN penetration but leave office document workflows unserved. ChatGPT commands 81% of the VN AI market; Gemini has 51%. Yet both tools require prompt engineering, produce generic output with recognizable AI style, and lack role-specific templates for TA, HR, Admin, or Sales tasks. High adoption does not equal solved need.
- 4 Pricing mismatch is real and documented. USD-billed subscriptions ($20/month) are a documented barrier for Vietnamese users. OpenAI launched a Go plan at local pricing (132k VND/month in VN) across 16 Asian markets to address this. No tool offers per-output pricing. The irregular usage pattern of non-power users - heavy some weeks, light others - is unaddressed by every subscription model in the market.
- 5 The enterprise AI segment (FPT.AI, Lac Viet) and the individual office worker segment are cleanly separated. Enterprise tools require procurement, IT integration, and cost 30-80 million VND to set up. No lightweight, self-serve, individual-use AI tool for Vietnamese office workers exists that also offers structured, professional-quality document output.
- 6 Claude Cowork (January 2026) is the first global product to target document-heavy knowledge work autonomously - an adjacent but distinct segment from Klerio's. Cowork's advertised audience is researchers, analysts, legal, and finance teams doing synthesis, data extraction, and document preparation from local files. Klerio targets HR, TA, Admin, and Sales workers who need role-specific output (JDs, performance reviews, recruitment emails) in Vietnamese. Both remove iterative prompting, but for different tasks and different users. Cowork closes the "no autonomous document AI" gap but not the guided Skill + Vietnamese office-role output gap Klerio targets.
- 7 No VN-native tool owns its underlying model - the entire market runs on the same small set of third-party APIs. AIVA, AI Hay, Kiki, and LovinBot all run on GPT, Gemini, or Claude via API. This means raw model quality is not a defensible moat for any VN player, including Klerio. Differentiation must come from what is built on top of the model - Skill design, role-specific depth, Vietnamese localization, pricing structure - not from the model itself.
- 8 Retention in the VN AI market is driven by lock-in, not output quality - but the mechanism differs by tool. ChatGPT holds 81% share through habit lock-in: it was the first mass-market AI tool users encountered, became the category default, and benefits from behavioral inertia that has nothing to do with integration. Gemini and Copilot hold through workflow lock-in: the AI lives inside tools users already open every day - switching requires abandoning a familiar workflow, not just learning a new interface. The competitive question for Klerio is not "which tool produces better output" but "which type of lock-in can Klerio interrupt - and with what."
Surprising / Unexpected
- 1 Even the closest VN competitor (AIVA) chose chat-first. AIVA is Vietnamese-made, targets non-technical office workers, uses templates, and is active in marketing/HR/sales - the highest overlap with Klerio. Yet its interface is still chat-driven with templates bolted on top. No VN tool has inverted the paradigm to Skill-first.
- 2 Marketing, HR, and Admin workers in VN - exactly Klerio's target audience - already use Canva at confirmed high adoption rates and are template-comfortable. Their office document work (JDs, performance reviews, recruitment emails) sits entirely outside what Canva does and is unserved by any current tool. This is the most specific early audience signal the research produced: a large, identifiable cohort that already understands guided-workflow tools and has unmet document needs no competitor is serving.
- 3 Vietnamese workplace stigma suppresses open AI adoption - not usage itself. No competitor addresses this. VietnamNet research found workers fear being perceived as "lazy, incompetent, or less capable" when using AI. The result is covert usage: people use ChatGPT privately, copy the output, and present it as their own work - which is why 81% usage and documented stigma coexist without contradiction. No competitor has built a positioning response to this. Klerio's opportunity: frame Skill-based output as applying a professional skill ("I used our JD Writing Skill") rather than asking AI to do the work ("I asked ChatGPT to write this"). This framing is legitimate only if Skill design genuinely requires professional inputs and judgment from the user - a product design requirement, not just a messaging claim.
- 4 Kiki/Zalo - free and embedded in an 80M-user platform - added government digital citizen assistance features in Sep 2025. This shows VNG is investing in Kiki's capabilities. If Zalo adds structured document output features, it becomes a formidable competitor with zero acquisition cost and massive reach.
- 5 Claude Cowork's January 2026 launch signals that the global AI race has moved beyond chat into autonomous document execution. Cowork is included in Claude Pro at no extra cost ($20/month), meaning any existing Claude subscriber already has access to an autonomous document agent. The window in which "no global tool handles office document work autonomously" has closed - but the guided Skill + Vietnamese output position Klerio targets remains uncontested.
- 6 No competitor - VN or global - offers client-side PII masking, despite users routinely sending sensitive workplace documents through AI tools. HR records, performance reviews, contracts, and recruitment data pass through ChatGPT, Gemini, Kiki, and AIVA with no user-facing control over what is sent. All data protections found in the market are server-side policies or enterprise admin controls - invisible to the individual worker. This absence is complete and universal across all competitors surveyed.
What for Klerio
Strategic implications across positioning, go-to-market, and product priorities - grounded directly in the research findings above.
Positioning
- Skill-first is the clear and uncontested position. No competitor leads with guided Skills. Anchor all messaging here.
- Vietnamese language by design - not by prompt. State this explicitly as a product guarantee.
- Privacy Guard must be front-and-center, not buried as a technical feature. Every target role handles sensitive data daily.
- Differentiation vs. Claude Cowork must be explicit: guided Skill per role (vs. natural language goal), web-based (vs. desktop-only), Vietnamese by design (vs. English-first).
Go-to-Market
- Marketing/HR/Admin users who already use Canva are the highest-priority early audience. Template-comfortable, in Klerio's target roles, and their document work is completely unserved by current tools.
- Pay-per-output pricing is Klerio's hypothesis for the documented subscription mismatch. The problem is confirmed: flat-rate subscriptions don't fit irregular usage. Whether VN users are willing to pay per output is the key commercial question to validate before launch.
Product Priorities
- Skill specificity by role is the depth moat. AIVA has 350+ agents but they are broad. Klerio's defensibility = depth on 3-5 specific Skills (Write JD, Performance Review, Recruitment Email).
- First-mover window on TA/HR depth is short. Go deep before going wide.
- Close the context switch with lightweight export: "copy to Google Docs" and "paste to Gmail" remove the primary friction without requiring full integration.
Big Questions
Open questions without a definitive answer yet - for discussion and further validation.
Big Question #1
If Skill-first and pay-per-task have the potential to win, why haven't the global giants and current VN players done it already? They are certainly capable of thinking of it - so what is the real reason?
Global players
- They are making good money from their current model. ChatGPT, Claude, and Gemini have millions of paying subscribers. Shifting to pay-per-use means deliberately disrupting their own stable revenue stream.
- The VN office worker market is too small for them. OpenAI builds for billions of global users. A JD-writing template with correct Vietnamese HR tone is not worth the engineering effort at their scale.
- They build platforms, not specific products. ChatGPT and Gemini are general-purpose platforms. Klerio is a specific product built on top - the right role in the ecosystem.
VN players
- ChatGPT succeeded with chat-first, so everyone copied it. When AIVA was built, the most successful model was ChatGPT. Copying what works is rational with limited resources. Klerio's approach is a complete inversion - it takes conviction to bet on a direction no one has taken.
- Pay-per-use revenue is less predictable. Fixed monthly subscription is easier to forecast and easier to present to investors. Usage-based billing fluctuates month to month.
- Going deep on content requires more work than expected. Building a JD Skill with correct Vietnamese HR tone is not a technical problem - it is a market knowledge problem.
Real risks to watch
- AIVA can pivot at any time. They already have a template library - adding role-specific forms is not out of reach for them.
- Pay-per-task has not been validated with VN users. The entire argument above assumes users want to pay per completed document - this needs user research to confirm.
- Microsoft Copilot Cowork is moving in a similar direction at the enterprise level - not form-first, not Vietnamese, but the trend is shifting toward paying for output rather than access.
The opportunity exists not because competitors lack the capability - but because their priorities and resources are elsewhere. That window can close if AIVA pivots or Kiki/Zalo adds structured output.
Update vs Recommendation v1
Recommendation v1 approved by Founder on 12 June 2026. This section documents what changed since then - from market research (June 23-24) and weekly AI updates (June 10-25).
Existing user base
Klerio is not starting from zero. The previous version of the product (LacViet AI) has 100+ active beta users on a chat-first, basic-feature version. The app is still active. The transition to the Skill-based model defined in PRD v1 needs to account for converting these users from chat-first behavior to Skill-first.
What changed and what stayed
Confirmed from PRD v1
- Prompt barrier is the #1 unmet need - confirms H1
- Subscription mismatch (irregular usage unfit for flat-rate) confirmed as real pain - confirms H2 problem; pay-per-task as solution needs user research to verify
- Skill-first position is structurally unoccupied
- Privacy Guard: zero competitors with equivalent client-side PII masking
Added by market research
- AIVA is the highest-threat VN competitor specifically - not tracked prominently in PRD v1
- Kiki/Zalo 80M platform reach: distribution threat if it adds structured output
- Marketing/HR/Admin Canva users are the clearest early-adopter cohort
- Entire VN market runs on third-party APIs - model is not a moat for anyone
- AIVA has a confirmed prompt improvement feature at input level
- Workplace AI stigma documented - Skill-first framing has potential to normalize open usage
Changed by weekly AI updates
- Microsoft Copilot Cowork GA June 16 - new agentic competitor, not in original research
- Fable 5 suspended June 12 - invalidates PRD v1.1 cost reduction assumption; recalculate on Opus 4.8 pricing
- 15 Claude API outages in 18 days - multi-model fallback must be an architecture decision before September beta
- Google Workspace AI (Gmail Live, Docs Live) deploying summer 2026 - deepens Gemini workflow lock-in in Google Workspace
- Copilot Credits ($0.01/credit) shows pay-per-task is commercially viable at enterprise level - does not verify VN user willingness to pay per output; differentiate on VND pricing and Skill depth
- Vietnam AI Law (effective March 1, 2026) - user notification mandatory for all AI products; build into beta
Next research priorities
Deep research on 5 priority competitors: features, UI/UX, approach to VN users, how they PR and communicate updates, what to learn vs. what to avoid.
Global - agentic
Global - chat dominant
Global - Workspace lock-in
VN - highest threat
VN - 15M downloads
Market Research
Full data behind the findings above. Use this section to verify a specific claim or go deeper on a competitor.
Market Overview
Vietnam AI market scale and adoption context as of June 2026.
81%
ChatGPT VN market share
among Vietnamese AI users
51%
Gemini VN adoption
second highest in market
15M+
AI Hay downloads
most-downloaded VN-native AI app
80M
Zalo MAU platform
Kiki AI embedded - 1M+ Kiki Info users
Market Structure
The Vietnam AI market has three distinct, well-served segments with a gap underneath.
Enterprise AI - high cost, IT-gated, B2B
FPT.AI and Lac Viet require IT procurement and setup costs of 30-80 million VND. They serve 3,000+ enterprise clients and VN universities. Individual workers at these companies have no self-serve access.
Global general AI - free tier accessible, USD paid plans
ChatGPT (81% VN share), Gemini (51%), Claude, and Copilot are widely used via free tiers. Paid plans are USD-denominated and represent a documented barrier - ChatGPT Plus at $20/month prompted OpenAI to launch a separate VND plan for Vietnam. Output requires prompt engineering and is not role-specific or Vietnamese-optimized.
VN consumer AI - free or VND-priced, chat-first
Kiki/Zalo (free, 80M platform), AI Hay (VND, 15M downloads), AIVA (VND, 350+ templates), and LovinBot serve Vietnamese users natively at local pricing. Language quality is strong. None offers structured office-role document output - all are chat-first or template-on-top-of-chat.
The missing segment - Skill-first, individual, office-role AI
No tool leads with a guided form-to-document flow for specific Vietnamese office roles (TA, HR, Admin, Sales). AIVA comes closest - affordable, self-serve, Vietnamese-native, 350+ templates - but remains chat-first and content/marketing-centric. Claude Cowork (January 2026) handles document tasks autonomously but requires natural language goal input, is English-optimized, and is desktop-only. The gap Klerio targets is not the absence of document automation, but the absence of a guided form to single ready-to-use Vietnamese output flow for specific office roles.
Pricing Landscape
- 1 Freemium + monthly subscription is the universal model across all 14 competitors surveyed.
- 2 USD-billed subscriptions create documented friction for VN users. ChatGPT Plus at $20/month prompted OpenAI to launch a Go plan at local pricing (132,000 VND/month in VN) across 16 Asian markets in October 2025. Claude, Notion AI ($20/user/month), and Gamma remain USD-only.
- 3 Only VN-made tools price in VND. AIVA (399k-1.2M VND/month), LovinBot (3.2M VND/6 months), Kiki (free). USD pricing is a real accessibility barrier for Vietnamese office workers on local salaries.
- 4 No tool offers pay-per-output pricing. Sentiment data identifies subscription mismatch as the top documented pain point across ChatGPT, Claude, and Notion AI. Irregular users - heavy some weeks, light others - pay the same flat rate as power users.
- 5 Enterprise tools do not disclose pricing and require contact-sales. Individual workers cannot self-serve even if they wanted to.
- 6 Subscription usage caps create mid-session disruptions beyond just cost. ChatGPT Plus limits ~160 messages per 3 hours; Claude hits usage windows that void ongoing sessions. Sentiment data documents how these interruptions translate to lost productivity.
Pricing by Tool
Specific plan prices as of June 2026. Sorted VN tools first, then global.
| Tool | Free tier | Entry paid | Higher tiers | Notes |
|---|---|---|---|---|
| AIVA | Yes - 10 credits/day | ~399,000 VND/month (Basic) | Pro ~699,000 VND/month · VIP ~1,199,000 VND/month | VND 50% discount currently active |
| Kiki / Zalo AI | Free only | No paid tier | No paid tier | VND Embedded in Zalo |
| LovinBot AI | Yes - 50 credits/month | 3,245,000 VND / 6 months (PRO) | 12,000,000 VND / 12 months (TEAM) | VND Credits-based |
| AI Hay | Yes - limited queries/day | Pro tier (VND price not disclosed) | Not disclosed | VND |
| Vbee | Not offered | $25/month (600k chars, 10 concurrent) | Enterprise (undisclosed) | USD only |
| Lac Viet AI | No | Enterprise only - contact sales | Enterprise only - contact sales | B2B only |
| FPT.AI | No | Enterprise only - contact sales | Enterprise only - contact sales | B2B only |
| ChatGPT | Yes | 132,000 VND/month (Go, VN local pricing) | Plus $20/month · Pro $100/month | VND Go plan launched VN Oct 2025 |
| Claude / Cowork | Yes - Claude.ai only (Cowork requires paid plan) | $20/month (Pro - includes Cowork) | Max 5x $100/month · Max 20x $200/month | USD only |
| Gemini | Yes | $7.99/month (AI Plus) | AI Pro $19.99/month · AI Ultra $99.99/month | USD only |
| Canva AI | Yes | ~$12-15/month (Pro) | Teams: custom per-seat | USD only |
| Microsoft Copilot | Limited (Windows/web only) | $18/user/month (Business) | Enterprise $30/user/month | USD only Requires M365 |
| Copilot Cowork NEW | No free tier | $0.01/credit (pay-as-you-go) + M365 Copilot subscription | Included in M365 Copilot tiers | USD only Pay-per-credit viable at enterprise; VN user willingness unvalidated |
| Notion AI | Yes - limited AI | $10/user/month (Plus - no AI) | Business $20/user/month (AI included) | USD only AI only in Business+ |
| Gamma | Yes - 400 AI credits (Gamma branding on output) | $8-10/month (Plus) | Pro $15-20/month | USD only |
Behavioral Context
- 1 Prompt barrier is the #1 documented unmet need. Non-technical users hit a skill wall: they cannot extract professional-quality output without knowing how to write effective prompts. Documented across ChatGPT, Claude, and Gemini reviews and VN user forums.
- 2 Vietnamese language inconsistency is a shared weakness of global tools. ChatGPT, Claude, and Gemini all show mid-response language switching, tone errors, and mismatched formality in Vietnamese output. VN-made tools (AIVA, AI Hay, Kiki) handle this natively.
- 3 AI workplace stigma suppresses open AI adoption, not usage itself. VietnamNet research (2026): workers fear being perceived as "lazy, incompetent, or replaceable" - so they use AI privately and hide it. That is why ChatGPT holds 81% usage while stigma remains strong. No competitor has addressed this in positioning. Klerio's opportunity: "applying a professional Skill" vs. "asking AI to do it" is a framing that could normalize open use - but only if Skill design requires enough professional input to justify the claim.
- 4 Tool integration drives retention at the top. Gemini (Google Workspace) and Copilot (Microsoft 365) have the strongest enterprise retention because the AI lives inside tools users already open every day. VN-native tools have no equivalent integration depth.
Competitive Segment Map
Classified by interaction model - what the tool does, not where it comes from. All 15 tools placed.
Open Chat
Free-form prompt - user must know what to ask
Embedded in Existing Tools
AI lives inside tools users already open daily - strongest retention
Template / Guided Workflow
Pick a format, fill details, get output - guided but still chat-driven or design-first
Autonomous Agent
User states a goal; agent handles end-to-end execution - no guided form, no VN role output
Enterprise B2B
IT-gated, procurement required, 30-80M VND setup cost
Specialist Vertical
TTS/voice only - different output medium, not a competitor
Skill-First Guided Output
Guided form per role - single ready-to-use professional output in Vietnamese
Klerio - unoccupied position
Competitor Landscape
Sorted by threat level (High to Low) within each group.
VN Competitors
| Name | Year | Positioning | Pricing | Skill-based? | VN adoption | Threat |
|---|---|---|---|---|---|---|
| AIVA | ~2024 | All-in-one VN AI for content, marketing, and automation | Freemium/Sub (VND) | Partial - templates on top of chat | Native VN - user count undisclosed | High |
| Kiki / Zalo AI | Jan 2025 | Free general AI embedded in Zalo | Free | No | 1M+ Kiki Info users; 80M Zalo platform | Medium |
| LovinBot AI | ~2024 | White-label AI for VN agencies and resellers | Freemium (VND) | Partial - 54+ content templates | Native VN - user count undisclosed | Low-Medium |
| Lac Viet AI | Post-2020 | B2B HR chatbot integrated with SureHCS HR software | B2B/Enterprise | No | VN SMEs + universities | Low-Medium |
| AI Hay | 2022 | Vietnamese student Q&A with source-cited answers | Freemium (VND) | No | 15M+ downloads, 120M+ queries/month | Low |
| Vbee | 2018 | Vietnamese TTS/voice cloning/ASR | Sub ($25 USD/mo) | No | 2M+ end-users, 300+ business clients | Low |
| FPT.AI | 2017 | Enterprise conversational AI for VN banking, insurance, retail | Enterprise | No | 3,000+ enterprise clients | Low |
Global Competitors
| Name | Year | Positioning | Pricing | Skill-based? | VN adoption | Threat |
|---|---|---|---|---|---|---|
| ChatGPT | Nov 2022 | General-purpose AI assistant; most recognized AI brand globally | Freemium + Sub (VND 132k available) | No | High (81%) | High |
| Canva AI | Oct 2023 | All-in-one creative AI suite for non-designers | Freemium + Sub | Partial - guided design workflows | High (Marketing/HR/Admin) | Medium |
| Gemini | Feb 2024 | AI embedded in Google Workspace (Gmail, Docs, Sheets) | Freemium + Sub | No | High (51%) | Medium |
| Claude / Cowork | Mar 2023 / Jan 2026 | Claude chat: long-context AI. Cowork: autonomous desktop agent for document-heavy work | Freemium + Sub (USD); Cowork in Pro $20/mo | No | Low-Medium | Medium |
| Gamma | 2022 | AI-guided presentation and website builder | Freemium + Sub | Partial - guided flow for presentations only | Low-Medium | Low |
| Notion AI | Feb 2023 | AI workspace for knowledge workers; model-agnostic | Freemium + Sub ($20/user) | Partial - templates within Notion | Low | Low |
| Microsoft Copilot | Nov 2023 | AI embedded in Microsoft 365 (Word, Excel, Outlook, Teams) | Add-on Sub ($18-30/user + M365 license) | No | Low-Medium | Low |
| Microsoft Copilot Cowork NEW | Jun 2026 (GA) | Agentic workspace AI inside M365 - user states a goal, agent handles multi-step execution; runs on Claude Opus 4.8 + Sonnet 4.6 | Pay-per-credit ($0.01/credit) + M365 Copilot subscription | No - goal-level natural language only, no guided form, no VN role templates | Low-Medium (enterprise M365) | Medium |
UI / Interaction Patterns
How each tool actually looks and behaves at the screen level. Matters for Klerio because Skill-first is a UI-level claim, not just a product category claim.
VN Competitors
| Name | Entry point | Core interaction model | Input method | Output format |
|---|---|---|---|---|
| AIVA | Web (app.aiva.vn) + Chrome extension + mobile app | Chat-first; left sidebar navigation to 350+ agent/template categories; sparkle button at chat input bar for "cai thien cau hoi" (prompt improvement); browseable agent library accessible before starting a task | Text chat; agent/template selected from side library | Text content (articles, scripts, social posts, marketing copy); image and video generation on separate tabs |
| Kiki / Zalo AI | Embedded Zalo mini-app (mobile); Kiki Auto (in-car) | Chat thread inside Zalo messaging app; no standalone UI or website - entirely embedded in Zalo | Text or voice message inside Zalo chat | Text answers; task completions (payments, ride booking) via Zalo mini-app actions |
| LovinBot AI | Web platform (lovinbot.ai) - no mobile app | Hub-style: tabbed sections for AI Store (branded tool catalog), Chat Pro (multi-model chat: GPT/Gemini/Claude switcher), AI Marketing (54+ content templates), SEO Writer Pro; templates browseable before starting | Template picker from library + text fill-in fields (for templates); free-form text prompt for Chat Pro | Content marketing text (articles, ad copy, social posts, SEO articles); white-label output for resellers |
| Lac Viet AI | Deployed chatbot widget on client HR platforms (SureHCS, university websites) | Pre-scripted HR Q&A chatbot; no standalone UI accessible to individual workers - all interactions happen inside the client's deployed system | Text chat in HR software or company website widget | HR policy answers; CV/contract classification summaries delivered inside HR software |
| AI Hay | Mobile app (iOS/Android) | Single-column chat Q&A thread; source citations displayed below each answer | Free-form text query; image upload for equations/objects | Text answer with numbered citations; step-by-step solution list |
| Vbee | Web tool (vbee.vn) | Minimal three-step UI: paste text, pick voice/dialect, click generate. No chat, no back-and-forth | Plain text paste into input box; voice + dialect dropdown selector (Northern/Central/Southern VN, 50+ languages, 700+ voices) | MP3/WAV audio file download |
| FPT.AI | Enterprise web console + API | Admin/developer dashboard for bot configuration; end-users interact via chatbot widgets deployed by enterprise clients - no standalone consumer UI | Enterprise admin: form-based bot builder. End-user: chat widget on third-party site | Automated chatbot responses; form fills; document classification outputs delivered inside client systems |
Global Competitors
| Name | Entry point | Core interaction model | Input method | Output format |
|---|---|---|---|---|
| ChatGPT | Web + mobile + desktop app | Centered single-column chat thread; model selector at top bar (GPT-5, 4o, etc.); left sidebar for conversation history and Projects; file attachment, voice mode, memory panel | Free-form text prompt; voice; file/image upload; built-in DALL-E image prompting | Text, images, code, file summaries; no structured templates or forms |
| Canva AI | Web + mobile (canva.com) | Canvas design editor as primary interface; Magic Write = AI text generation inside any design text element; Magic Studio = dedicated AI tools side panel (text-to-image, background removal, Magic Expand, AI presentation builder); template library is the entry point when creating a new design | Template selection from gallery or blank canvas; Magic Write: click text element + type prompt; Magic Studio: open panel + describe what you want | Visual designs (social posts, presentations, flyers, reports) with AI-generated copy embedded inside; text-only output via Magic Write is possible but secondary to visual output |
| Gemini | Web (gemini.google.com) + mobile + right-side embedded panel in Gmail/Docs/Sheets/Slides | Chat thread on gemini.google.com; Gems (saved custom AI personas) accessible from sidebar; inside Google Workspace: floating sparkle button opens AI side panel - context-aware of current document/email content | Text prompt; inside Workspace: highlight text then choose action, or open side panel and prompt | Text, images (Imagen); inside Gmail: email drafts; inside Docs: inline paragraph suggestions; inside Sheets: formula and data generation |
| Claude (chat) | Web + mobile | Clean single-column chat thread; Projects in left sidebar; file attachment; Artifacts panel (shows generated documents separately from chat); voice mode (no Vietnamese support) | Free-form text prompt; voice; file/image upload | Long-form text, code, document analysis; Artifacts shown in side panel; no templates |
| Claude Cowork | Desktop app (macOS + Windows) - separate from claude.ai web | Goal-input then autonomous execution: user describes desired outcome in natural language; Cowork works independently across local files/folders/apps and delivers a finished output; no guided form, no back-and-forth prompting required | Natural language goal description; no structured form; user sets cadence for scheduled tasks | Finished local files (organized folders, drafted documents, synthesized research reports, extracted data tables) delivered to user's desktop |
| Gamma | Web (gamma.app) | Three-step guided flow: (1) text prompt or paste outline, (2) theme/style picker (cards showing visual layouts), (3) AI generates full presentation; Step 4: slide editor with AI chat sidebar for post-generation revisions | Single text prompt or outline paste; theme visual selection | Complete presentation deck (slides + auto-generated images + layout); exportable to PowerPoint, PDF, Google Slides |
| Notion AI | Embedded inside Notion workspace - no standalone interface | "/" slash command triggers AI inline within a Notion page; or highlight text to get floating action menu (summarize, rewrite, translate, etc.); Notion AI reads full workspace for context | "/" command followed by prompt; text highlight + action selection; side chat panel (newer) | Inline text edits, new document drafts, database rows, page summaries - all appear directly in Notion pages; no external output |
| Microsoft Copilot | Embedded side panel inside Word/Excel/PowerPoint/Outlook/Teams; also at copilot.microsoft.com | Side panel docks to the right inside each M365 app; context-aware of the open document/email/spreadsheet; command input field at bottom of panel; no guided form - user types what they want | Natural language prompt in side panel; document/email context loaded automatically | Drafts and edits in Word; formulas and table fills in Excel; new slide decks in PowerPoint; email summaries and reply drafts in Outlook; meeting transcription + action items in Teams |
| Microsoft Copilot Cowork NEW | Dedicated Cowork workspace inside M365 - separate from general Copilot side panel; GA June 16, 2026 | Agentic workspace: user describes a goal or delegates a multi-step task; Cowork works across M365 apps and connected tools (9 partner plugins post-GA), returns completed output; runs on Claude Opus 4.8 + Sonnet 4.6; pay-per-credit billing ($0.01/credit) | Natural language goal description; M365 context (files, emails, calendar) loaded automatically; no structured form | Completed documents, spreadsheets, email sequences, data analyses - delivered inside M365 environment |
- -Gemini and Copilot are the only tools where the AI is genuinely invisible to the user - it appears as a feature of tools they already have, not a new tool to learn.
- -Gamma is the only tool that uses a multi-step guided flow (prompt then theme selection then generation) - closest in concept to Klerio's Skill-first approach, but limited to presentations only.
- -AIVA is the only VN tool with a pre-submission input improvement feature (sparkle button at chat input) and a browseable agent library - the most structured VN competitor UI.
- -Vbee has the simplest UI in the set: three inputs, one output, no iteration. High task specificity = minimal interface.
- -Claude Cowork is the only tool that removes the iterative prompting loop entirely via autonomous execution - but it does this through goal-level natural language, not a guided form. Users still need to know what outcome they want and describe it clearly. No role-specific Vietnamese templates exist.
- -Klerio's planned UI (guided form per Skill, guided input fields, single-shot output) has no direct equivalent in any of the tools surveyed.
Feature Comparison Matrix
Key capabilities across all 16 competitors. ✓ confirmed yes - ✗ confirmed no - ~ uncertain/partial.
| Tool | Guided input - single-shot output | Ready-to-use output | Vietnamese language quality | VND / local pricing fit | Native tool integration | Privacy Guard / PII control | Prompt Rewrite Suggestions |
|---|---|---|---|---|---|---|---|
| Klerio | ✓ | ✓ | ✓ | ✓ | ~ | ✓ | ✓ |
| AI Hay | ✗ | ✗ | ✓ | ✓ | ✗ | ~ | ✗ |
| AIVA | ~ | ~ | ✓ | ✓ | ~ | ~ | ~ |
| FPT.AI | ✗ | ✗ | ✓ | ✗ | ✓ | ~ | ✗ |
| Kiki / Zalo AI | ✗ | ✗ | ✓ | ✓ | ✓ | ✗ | ✗ |
| Lac Viet AI | ✗ | ✗ | ✓ | ✗ | ✓ | ~ | ✗ |
| LovinBot AI | ~ | ~ | ✓ | ✓ | ~ | ~ | ~ |
| Vbee | ✗ | ✗ | ✓ | ✗ | ✓ | ~ | ✗ |
| ChatGPT | ✗ | ✗ | ~ | ~ | ~ | ~ | ✗ |
| Claude / Cowork | ✗ | ✗ | ~ | ✗ | ~ | ~ | ✗ |
| Gemini | ✗ | ✗ | ~ | ~ | ✓ | ~ | ✗ |
| Microsoft Copilot | ✗ | ✗ | ✓ | ✗ | ✓ | ~ | ✗ |
| Copilot Cowork NEW | ✗ | ✗ | ~ | ✗ | ✓ | ~ | ✗ |
| Notion AI | ~ | ~ | ~ | ✗ | ✓ | ~ | ✗ |
| Gamma | ~ | ~ | ~ | ~ | ~ | ✗ | ✗ |
| Canva AI | ~ | ~ | ✓ | ~ | ✓ | ~ | ✗ |
- -Guided input / single-shot output: ✓ only if the tool converts the user's problem via a guided form into a ready output without multiple rounds of back-and-forth. AIVA/LovinBot/Notion AI/Gamma rated ~ because they have templates or structured flows but still require iterative chat interaction.
- -Ready-to-use output: ✓ only if output is professional-grade and role-specific without further editing. Sentiment data downgrades ChatGPT, Claude, Gamma, and Canva AI from potential ✓ to ✗/~.
- -VND / local pricing fit: ✓ if tool has VND pricing or free tier accessible to individual VN office workers. ~ for USD-priced tools with a free tier. ✗ for USD-only subscription or enterprise-contact-sales models.
- -Privacy Guard / PII control: Klerio ✓ - on-device client-side PII masking + transparency/control layer showing users what data is used (planned P1). No competitor has equivalent client-side masking or user-facing transparency. ~ = server-side policies or enterprise admin controls only. ✗ = confirmed no controls or confirmed problematic data practices (Kiki: forced data collection expansion Dec 2025; Gamma: confirmed trains on user data, no opt-out in policy).
- -Prompt Rewrite Suggestions: Klerio ✓ - opt-in feature at input bar; system suggests multiple rewritten versions before Skill matching. AIVA confirmed ~: has a "cai thien cau hoi" sparkle button at chat input (user-triggered, pre-submission; scope unverified). Source: app.aiva.vn, researcher observation 2026-06-23. AIVA is the only VN competitor with a confirmed input-level prompt improvement feature.
- -Klerio is the only tool with ✓ on both Guided input and Privacy Guard columns.
Competitive Positioning Map
X axis: Chat-first vs Skill-first / Solution-first. Y axis: Global-optimized vs Vietnam-localized. Klerio targets the unoccupied top-right quadrant.
SWOT Analysis
Stars = top-priority items.
Strengths
- ★ Only tool with guided Skills to ready-to-use document output for Vietnamese office roles. No competitor occupies this position.
- ★ Pay-per-output pricing addresses documented subscription mismatch. No competitor offers it. Willingness to pay per output has not been validated with VN office workers - requires user research.
- Built natively for Vietnamese - language quality by design, not by prompt.
- No prompt engineering required - addresses the #1 documented unmet need.
- Individual, self-serve model fills the gap between enterprise tools and generic chat tools.
- Privacy Guard is an uncontested differentiator - no competitor offers client-side PII masking.
Weaknesses
- ★ No brand recognition in a market where ChatGPT has 81% share.
- ★ Dependent on third-party LLM APIs - no proprietary model.
- No existing user base, community, or network effects.
- Skill library depth is unproven at launch. A shallow Skill library will drive churn.
- No native tool integration at launch - Klerio requires a context switch.
Opportunities
- ★ Template-comfortable Canva users are the clearest early-adopter signal - Marketing/HR/Admin VN workers with confirmed Canva adoption and completely unserved document needs.
- ★ Prompt barrier + subscription cost are the two most-cited unmet needs - Klerio addresses both simultaneously.
- Individual HR/TA workers at FPT.AI/Lac Viet enterprise clients are unserved on personal document output.
- AI workplace stigma may be reduced by Skill-based framing vs. chatbot positioning.
- Vbee (TTS/voice only) - different output medium, not a competitor and not currently relevant to Klerio scope.
Threats
- ★ AIVA - highest-threat VN competitor; could pivot from content-first to office-document-first.
- ★ Kiki/Zalo - free, 80M distribution; if Zalo adds structured document output, acquisition cost spikes.
- Claude Cowork (Jan 2026) - adjacent segment overlap; included in Claude Pro at $0 marginal cost for existing subscribers.
- Microsoft Copilot Cowork (GA June 16, 2026) - agentic AI on M365, pay-per-credit. Enterprise channel with wide distribution.
- Gemini Workspace AI deepens lock-in: Gmail Live + Docs Live deploying summer 2026 at $19.99/mo. Direct overlap with Klerio's email and document use cases.
- ChatGPT's regional pricing investment (VND 132k/mo) signals long-term VN market commitment.
- Third-party API dependency: Fable 5 was suspended June 12, 2026 with no warning. Any upstream model can disappear. Multi-model fallback is now a launch prerequisite.
- VN AI market moving fast - first-mover window on TA/HR depth is short.
Strategic Quadrant Combinations
Four cross-quadrant implications from the SWOT, grounded in the research.
Strength + Opportunity - Strategic bet
Target Canva-familiar Marketing/HR/Admin users first. They are template-comfortable and already understand guided-workflow tools. Their document work (JDs, performance reviews, recruitment emails) is completely outside what Canva does and unserved by any competitor. Klerio is for document work; Canva is for design work - two separate tools for one user.
Weakness + Threat - Major risk to address
Zero brand recognition vs. AIVA's existing community and ChatGPT's 81% market share. If AIVA adds office-specific templates (JD, HR, performance review) before Klerio builds traction, Klerio's main differentiator shrinks. Mitigation: prioritize template depth in specific roles (TA, HR) that AIVA's content-centric model underserves, and build early user community fast.
Strength + Threat - Defensive play
If pay-per-output is validated by user research, it becomes structurally hard for subscription-based AIVA and free Kiki to replicate without changing their business model. This potential pricing moat only matters if VN users actually prefer paying per completed document - which needs to be tested before claiming it as a differentiator.
Weakness + Opportunity - Build or partner?
Klerio's lack of native tool integration is addressable by building lightweight export features - "copy to Google Docs," "paste to Gmail" - rather than deep integrations. Given Canva's confirmed VN adoption gap, a future Canva plugin or export integration could collapse the context-switch barrier for the highest-opportunity user cohort.
Sources
All sources checked June 2026. Organized by competitor group and sentiment research.
VN Competitors
| Competitor | Source URL | Source file | Date checked |
|---|---|---|---|
| AI Hay | vietnamnet.vn - AI Hay global top 5 | competitor-ai-hay.md | 2026-06-23 |
| AI Hay | skywork.ai - AI Hay vs ChatGPT deep dive | competitor-ai-hay.md | 2026-06-23 |
| AI Hay | App Store - AI Hay listing | competitor-ai-hay.md | 2026-06-23 |
| AIVA | aiva.vn | competitor-aiva.md | 2026-06-23 |
| AIVA | tomaz.vn - AIVA pricing | competitor-aiva.md | 2026-06-23 |
| AIVA | Chrome Web Store - AIVA extension | competitor-aiva.md | 2026-06-23 |
| FPT.AI | fpt.ai | competitor-fpt-ai.md | 2026-06-23 |
| FPT.AI | secondtalent.com - Vietnam AI companies | competitor-fpt-ai.md | 2026-06-23 |
| Kiki / Zalo | zalo.me/en/ai | competitor-kiki-zalo.md | 2026-06-23 |
| Kiki / Zalo | vietnam.vn - top AI platforms Vietnam | competitor-kiki-zalo.md | 2026-06-23 |
| Kiki / Zalo | gialainews.com - Kiki top 10 AI Vietnam | competitor-kiki-zalo.md | 2026-06-23 |
| Kiki / Zalo | saigoneer.com - Zalo debuts Kiki | competitor-kiki-zalo.md | 2026-06-23 |
| Lac Viet AI | lacviet.vn | competitor-lac-viet.md | 2026-06-23 |
| Lac Viet AI | lacviet.vn - LV Chatbot AI assistant | competitor-lac-viet.md | 2026-06-23 |
| Lac Viet AI | lacviet.vn - 3 AI integrations in SureHCS HRM | competitor-lac-viet.md | 2026-06-23 |
| LovinBot AI | lovinbot.ai | competitor-lovinbot.md | 2026-06-23 |
| LovinBot AI | bigspyai.com - LovinBot review | competitor-lovinbot.md | 2026-06-23 |
| Vbee | vbee.vn | competitor-vbee.md | 2026-06-23 |
| Vbee | tryaivoices.com - Vbee AI voice | competitor-vbee.md | 2026-06-23 |
| Vbee | capterra.com - Vbee AIVoice | competitor-vbee.md | 2026-06-23 |
Global Competitors
| Competitor | Source URL | Source file | Date checked |
|---|---|---|---|
| ChatGPT | chatgpt.com/pricing | competitor-chatgpt.md | 2026-06-23 |
| ChatGPT | vietnam.vn - ChatGPT Go plan 132k VND | competitor-chatgpt.md | 2026-06-23 |
| ChatGPT | vnexpress.net - ChatGPT Go plan in Vietnam | competitor-chatgpt.md | 2026-06-23 |
| ChatGPT | sianamarketing.com - ChatGPT usage by country 2026 | competitor-chatgpt.md | 2026-06-23 |
| ChatGPT | techcrunch.com - ChatGPT Go expands to 16 Asian markets | competitor-chatgpt.md | 2026-06-23 |
| Claude | claude.com/pricing | competitor-claude.md | 2026-06-23 |
| Claude | cloudzero.com - Claude pricing | competitor-claude.md | 2026-06-23 |
| Claude | expertbeacon.com - Claude AI in Vietnam | competitor-claude.md | 2026-06-23 |
| Claude | github.com - Claude VN access discussion | competitor-claude.md | 2026-06-23 |
| Claude Cowork | anthropic.com/product/claude-cowork | competitor-claude.md (updated) | 2026-06-24 |
| Claude Cowork | claude.com/pricing | competitor-claude.md (updated) | 2026-06-24 |
| Gemini | finout.io - Gemini pricing 2026 | competitor-gemini.md | 2026-06-23 |
| Gemini | felloai.com - Gemini pricing | competitor-gemini.md | 2026-06-23 |
| Gemini | sianamarketing.com - AI usage by country 2026 | competitor-gemini.md | 2026-06-23 |
| Gemini | knowledge.workspace.google.com - Gemini Workspace add-ons | competitor-gemini.md | 2026-06-23 |
| Microsoft Copilot | microsoft.com - Copilot pricing | competitor-copilot.md | 2026-06-23 |
| Microsoft Copilot | copilot-experts.com - Copilot pricing guide | competitor-copilot.md | 2026-06-23 |
| Microsoft Copilot | pacisoft.vn - Copilot ho tro tieng Viet | competitor-copilot.md | 2026-06-23 |
| Microsoft Copilot | support.microsoft.com - Copilot supported languages | competitor-copilot.md | 2026-06-23 |
| Copilot Cowork NEW | microsoft.com - Copilot Cowork GA announcement | weekly-ai-update-2026-06-18.md | 2026-06-18 |
| Copilot Cowork NEW | learn.microsoft.com - Copilot Cowork what's new | weekly-ai-update-2026-06-25.md | 2026-06-25 |
| Notion AI | notion.com/pricing | competitor-notion-ai.md | 2026-06-23 |
| Notion AI | techjacksolutions.com - Notion AI pricing | competitor-notion-ai.md | 2026-06-23 |
| Notion AI | dancumberlandlabs.com - Notion AI guide | competitor-notion-ai.md | 2026-06-23 |
| Gamma | gamma.app/pricing | competitor-gamma.md | 2026-06-23 |
| Gamma | eesel.ai - Gamma pricing | competitor-gamma.md | 2026-06-23 |
| Canva AI | eesel.ai - Canva AI pricing | competitor-canva-ai.md | 2026-06-23 |
| Canva AI | aiworthit.com - Canva AI review | competitor-canva-ai.md | 2026-06-23 |
| Canva AI | sawankr.com - Canva Magic Studio guide 2026 | competitor-canva-ai.md | 2026-06-23 |
Weekly AI Update Sources
| Topic | Source URL | Source file | Date checked |
|---|---|---|---|
| Claude Fable 5 launch | anthropic.com - Fable 5 + Mythos 5 launch | weekly-ai-update-2026-06-10.md | 2026-06-10 |
| Fable 5 + Mythos 5 suspended | anthropic.com - access suspension notice | weekly-ai-update-2026-06-18.md | 2026-06-18 |
| Copilot Cowork GA | microsoft.com - Copilot Cowork GA blog | weekly-ai-update-2026-06-18.md | 2026-06-18 |
| Copilot Cowork Credits + plugins | learn.microsoft.com - Cowork what's new | weekly-ai-update-2026-06-25.md | 2026-06-25 |
| Claude API 15 outages in 18 days | status.claude.com - incident history | weekly-ai-update-2026-06-25.md | 2026-06-25 |
| Google Workspace AI (Gmail Live, Docs Live) | blog.google - Google I/O 2026 announcements | weekly-ai-update-2026-06-10.md | 2026-06-10 |
| Viettel + NVIDIA sovereign AI | vietnamplus.vn - Viettel NVIDIA sovereign AI | weekly-ai-update-2026-06-25.md | 2026-06-25 |
| Vietnam AI Law compliance | vietnam-briefing.com - Vietnam AI Law | weekly-ai-update-2026-06-18.md | 2026-06-18 |
Sentiment Sources
| Competitor / Topic | Source URL | Source type | Date checked |
|---|---|---|---|
| ChatGPT | vnexpress.net - su phu thuoc ChatGPT | VnExpress article | 2026-06-23 |
| ChatGPT | eesel.ai - ChatGPT reviews | Review aggregation | 2026-06-23 |
| ChatGPT | capterra.com - ChatGPT reviews | Capterra reviews | 2026-06-23 |
| ChatGPT | nucamp.co - AI prompts for VN marketing professionals | Blog / VN marketing context | 2026-06-23 |
| ChatGPT | glbgpt.com - ChatGPT Plus price in Vietnam | Pricing comparison | 2026-06-23 |
| ChatGPT | aifreeapi.com - ChatGPT Plus usage limits | Usage limits guide | 2026-06-23 |
| ChatGPT | community.openai.com - ChatGPT Plus usage limits forum | OpenAI community forum | 2026-06-23 |
| ChatGPT | atomwriter.com - ChatGPT quality degradation | Output quality analysis | 2026-06-23 |
| AI workplace stigma | vietnamnet.vn - AI workplace stigma study | VietnamNet - workplace stigma study | 2026-06-23 |
| Notion AI | reviewsavvyhub.com - Notion review 2026 | Review site | 2026-06-23 |
| Notion AI | eesel.ai - Notion review | Review aggregation | 2026-06-23 |
| Gamma | eesel.ai - Gamma reviews | Review aggregation | 2026-06-23 |
| Canva AI | eesel.ai - Canva AI reviews | Review aggregation | 2026-06-23 |
| Canva AI | medium.com - Why Canva AI struggles with text | Medium article | 2026-06-23 |
| Claude / Gemini | xda-developers.com - Gemini vs Claude vs Copilot | XDA Developers review | 2026-06-23 |
| Gemini / Claude | xda-developers.com - Why Gemini not worth recommending | XDA Developers review | 2026-06-23 |
| FPT.AI | softwarereviews.com - FPT.AI | Software reviews | 2026-06-23 |
| FPT.AI | g2.com - FPT.AI reviews | G2 reviews | 2026-06-23 |