Semji vs MentionLab
MentionLab SRL — mentionlab.io — is a deep AI-visibility and AI-listening platform, and on measurement it is ahead of us: sentiment, Share of Voice, themes, shopping, YouTube citations, fan-outs, and the full text of every answer a model gave. What it never does is change any of it. There is no brief, no editor, no score, no agent and no CMS. Semji measures the same ground and then produces and publishes the page that moves it.
Last verified 1 August 2026
The short answer
MentionLab explains what AI thinks of you in more dimensions than Semji does. Semji does something about it. Which matters more depends on whether the report is the deliverable.
- Measurement is the deliverable and you want more of it — sentiment, Share of Voice, themes, shopping, YouTube and query fan-outs.
- You want to read what the model actually said: their Response History keeps the full text of every individual run.
- You want a starting price on a public page — “From €250/month” — where Semji publishes none at all.
- You are an agency or a group — linked parent and child organisations, a rollup of every child's balance and usage, and token purchases the parent approves and pays for.
- You are measured on visibility that improves, not visibility that is reported.
- SEO still drives most of your traffic and you refuse to run two disconnected tools.
- Your team publishes at volume and needs briefs, editing, approvals and a CMS connection.
- You want AI agents that produce and publish the work, governed at workspace level.
- You need AI visibility on the same page record as that page's rankings and traffic.
- Procurement will ask for SOC 2 Type II.
Feature by feature
Every MentionLab cell links to the page we read it on, with the date. Where MentionLab is stronger the row says so — and in the measurement section, three of them are.
From insight to published page
The decisive section, and the one MentionLab rules out by construction. Their own Terms scope the product as “a SaaS platform for AI visibility and listening” — everything in it is read-side.
| Capability | Semji | MentionLab |
|---|---|---|
| Detected gap becomes a content brief | SemjiYes. One click from an LLM gap into Planning, or into the editor to optimise the page that should already own it | MentionLabPartial. Their API's Source Topics group runs “topic extraction runs and content-gap analysis”, so you learn which topics to write. Nothing in the endpoint inventory briefs, drafts or scores the pageSource for Detected gap becomes a content brief, checked 1 August 2026 |
| AI writing grounded in your brand | SemjiYes. Brand Voice and Knowledge Base applied at generation, not after | MentionLabNo. The Terms list the whole Service: visibility tracking, share-of-voice, sentiment, benchmarking, citations, fan-outs, shopping analytics, tagging, reporting — all read-sideSource for AI writing grounded in your brand, checked 1 August 2026 |
| Page-level optimisation score | SemjiYes. Content Score across seven optimisation families — intents, questions, related terms, links, length, title, meta | MentionLabNo. The scored objects are AI answers, not pages: mention rate, sentiment, source rate. No page is graded because no page is being writtenSource for Page-level optimisation score, checked 1 August 2026 |
| Editorial planning, assignments and approvals | SemjiYes. Planning calendar, custom draft statuses, threaded comments with assignment, version history, bulk actions | MentionLabNo. Five organisation roles and three project roles, all access control. The only approval flow documented is a child organisation's token purchaseSource for Editorial planning, assignments and approvals, checked 1 August 2026 |
| One-click publishing to your CMS | SemjiYes. WordPress (incl. Yoast meta), Drupal, Contentful, Strapi, a custom MCP server and a send-to-CMS webhook | MentionLabNo. Six named exports run in the background and arrive by email, links expiring after 7 days — “Output is always CSV.” No CMS connector or publish action is documentedSource for One-click publishing to your CMS, checked 1 August 2026 |
| Visibility tied to the page that earns it | SemjiYes. GEO score, search position, Search Console performance and the draft itself on one page record | MentionLabPartial. Every cited URL carries its citation share, the entities found on it and the queries that produced it — but no position, click or conversion figure anywhere on the sheetSource for Visibility tied to the page that earns it, checked 1 August 2026 |
AI visibility measurement
MentionLab's home ground, and they are ahead of us in it. Three rows here go to them outright; a fourth would, if Share of Voice were its own line.
| Capability | Semji | MentionLab |
|---|---|---|
| Prompt tracking, mention and citation rates | SemjiYes. AI Visibility Score, mention rate, citation rate and average mention position | MentionLabYes. Mention rate, Avg. position and mentions per result — plus a first-class Share of Voice, which Semji does not have. Their glossary states the denominator for every oneSource for Prompt tracking, mention and citation rates, checked 1 August 2026 |
| Cited-source and competitor analysis | SemjiYes. GEO Sources by platform, source type, page type and root domain, with the prompts citing each source | MentionLabYes. Citation rate, citation share, Presence and True reach per domain, every cited URL and the queries behind it. The four-axis source-type card is marked betaSource for Cited-source and competitor analysis, checked 1 August 2026 |
| Sentiment and brand perception | SemjiNo. Not available | MentionLabYes. Every mention scored on a sentiment index — Positive 60+, Neutral 40–59, Negative below 40 — broken down per entity, so you can compare how models describe you and your competitorsSource for Sentiment and brand perception, checked 1 August 2026 |
| The raw AI answer, kept and readable | SemjiNo. Aggregate GEO metrics and sources, not the response text | MentionLabYes. Response History gives “every individual run, with the full response text, entity chips and its source list”, plus semantic response tags applied over past answersSource for The raw AI answer, kept and readable, checked 1 August 2026 |
| Surfaces beyond the chat answer | SemjiNo. Web answers only | MentionLabYes. Shopping visibility and retailer distribution, YouTube channels cited, and fan-out occurrence rates that identify “a question the model considers unavoidable in your category”Source for Surfaces beyond the chat answer, checked 1 August 2026 |
| Engines included in your plan | SemjiChatGPT, Google AI Overviews, Gemini, Claude, Perplexity and DeepSeek — no per-model add-on | MentionLabDocumented per model per country across 187 markets, Google AI Mode included — and enforced, not degraded. Cost scales with choice: tokens per run = queries × AI models × iterationsSource for Engines included in your plan, checked 1 August 2026 |
Search and AI in one system
AI answers have not replaced Google. MentionLab has no keyword, no search ranking and no visitor in its data model, so it stays a second subscription beside whatever you already run.
| Capability | Semji | MentionLab |
|---|---|---|
| Prompt prioritisation by real search demand | SemjiYes. Real monthly search volume on the prompt, with topic, persona and funnel stage | MentionLabNo. A query carries text, language, country, tags and an optional Comparative / Informative / Perception type. No volume, difficulty, funnel stage or persona field existsSource for Prompt prioritisation by real search demand, checked 1 August 2026 |
| Keyword research and rank tracking | SemjiYes. Daily positions, search intents, keyword difficulty, cannibalisation detection, competition duels | MentionLabNo. No keyword object and no SERP data. Their guide coaches you the other way: “Write queries the way a customer would”Source for Keyword research and rank tracking, checked 1 August 2026 |
| Traffic and performance attribution per page | SemjiYes. Clicks, impressions and average position per page, branded vs non-branded, estimated traffic | MentionLabNo. Every metric in the glossary divides by responses MentionLab collected itself. Nothing in it observes a visitor, a session or an orderSource for Traffic and performance attribution per page, checked 1 August 2026 |
| Search Console and analytics integrations | SemjiYes. Search Console, Google Analytics 4, Matomo, Adobe Analytics, Piwik Pro | MentionLabNo. Their own documentation index runs guides, analysis, account management and the API — and carries no analytics-connector section at allSource for Search Console and analytics integrations, checked 1 August 2026 |
AI agents and automation
Both ship an MCP server. The question is whether the platform does the work, or hands the data to an assistant you then have to drive yourself.
| Capability | Semji | MentionLab |
|---|---|---|
| Prebuilt agents that produce work | SemjiYes. About 24 templates — Content Master, GEO Expert, E-E-A-T Auditor, Legal Compliance, Content Localizer, CMS Publisher — in bulk across Planning | MentionLabNo. “Agentic Readiness” points the other way: simulating third-party agent journeys against your own site. That is an audit of your estate, marketed on the homepage and absent from the docsSource for Prebuilt agents that produce work, checked 1 August 2026 |
| Agents that draft, optimise and publish | SemjiYes. End to end inside the platform, on a schedule or on demand, with an Activities log and relaunch on failure | MentionLabNo. Nothing writes anywhere. Even a Read & Write API key “can read data and create or update project data” — its own project configuration, and no furtherSource for Agents that draft, optimise and publish, checked 1 August 2026 |
| Public REST API | SemjiYes. Public API v1, org-scoped keys, Swagger and a developer portal — full draft lifecycle | MentionLabPartial. Broad and genuinely well documented across analytics, entities, queries and jobs — but “The REST API is currently in beta and is subject to breaking changes”Source for Public REST API, checked 1 August 2026 |
| Assistants and team channels | SemjiYes. MCP server with OAuth and audited tool-call logs, plus a Microsoft Teams bot and Slack that push opportunities into the channel | MentionLabPartial. A remote MCP server with OAuth 2.1 and PKCE for Claude, Cursor and Windsurf — a personal assistant rather than a shared channel, and “currently in beta”Source for Assistants and team channels, checked 1 August 2026 |
Enterprise readiness
The section procurement reads. MentionLab publishes Belgian legal-entity details and a GDPR policy citing the EEA and standard contractual clauses — and no security attestation of any kind.
| Capability | Semji | MentionLab |
|---|---|---|
| SOC 2 Type II | SemjiYes. Certified | MentionLabNo. No SOC 2, ISO 27001 or trust portal is published. The privacy policy's whole security disclosure is access controls, “encryption in transit” and regular review; encryption at rest is not claimedSource for SOC 2 Type II, checked 1 August 2026 |
| SSO / SAML | SemjiYes. SAML with automated certificate rotation | MentionLabPartial. “SSO (SAML) is available on Enterprise agreements and is configured by MentionLab, not self-serve.” No SCIM; the role is fixed at invite timeSource for SSO / SAML, checked 1 August 2026 |
| AI usage and cost governance | SemjiYes. Consumption and time saved by workspace, user and feature, with a projection to end of contract | MentionLabPartial. Strong on money, silent on people: a token ledger, daily usage charts and PDF statements, filterable by project — but not by user or feature, and no forward projectionSource for AI usage and cost governance, checked 1 August 2026 |
| How you buy | SemjiQuote-based, with scoped onboarding and a dedicated customer success manager | MentionLabTheir pricing page publishes floors, not prices: Launch “From €250/month” for “1.500 credits / month”, Scale “From €450/month”, Enterprise “Contact us”. Every CTA on it books a demo or contacts salesSource for How you buy, checked 1 August 2026 |
See where your brand actually stands
Run a free GEO audit on your own domain. No signup, results in about two minutes.
- Free
- No signup
- Results in 2 minutes
Three things measurement cannot do
Not feature gaps — structural ones. MentionLab is very good at the half of the job it does. Each of these is the other half.
From gap to published page
- A prompt shows a competitor cited and you absent
- Semji opens a brief, or the existing page in the editor
- An agent drafts in your brand voice, grounded in your knowledge base
- Scored against seven families, reviewed, and published to your CMS
SEO and GEO in one model
- One topic and competitor model across search and AI answers
- Daily rank tracking and AI visibility on the same page record
- Clicks, impressions and positions attributed to the page that earns them
AI production your legal team will sign off
- Brand voice and knowledge base applied at generation
- Approval workflow, threaded comments and version history
- AI consumption reported by workspace, user and feature
- SOC 2 Type II, SAML SSO, and audited API and MCP access
Frequently asked questions
Which MentionLab is this comparison about?
MentionLab SRL, the Belgian company at mentionlab.io — the AI-visibility and AI-listening platform whose pricing page publishes Launch, Scale and Enterprise. There is a separate and unrelated company trading as mentionLAB at mentionlab.ai, which sells an AI agent that writes and auto-publishes content. Nothing on this page describes that product. If it is the one you are evaluating, this is the wrong comparison.
Can I run Semji and MentionLab together?
Yes, and it is a reasonable stack. MentionLab reads dimensions we do not — sentiment, themes, shopping, YouTube, fan-outs — and hands data over cleanly through six CSV exports, a REST API and an MCP server. Semji takes the finding and produces the page. The overlap is prompt tracking and cited sources, so most teams eventually consolidate once they see the same prompts and the same domains in both, but nothing forces that on day one.
Which AI engines does each tool cover?
Semji tracks ChatGPT, Google AI Overviews, Gemini, Claude, Perplexity and DeepSeek, with no per-model add-on — and does not track Google AI Mode or Microsoft Copilot. MentionLab documents its models per country across 187 markets, Google AI Mode included, with availability enforced at run time rather than silently degraded; that table is worth reading market by market, since AI Mode is listed as unavailable in France. We do not quote a model count for them: their marketing logo strip and their documentation do not list the same set, and the docs concede the tables “are maintained by hand and can lag”. Commercially, MentionLab meters runs rather than models — tokens per run = queries × AI models × iterations — so every engine you add multiplies against your monthly credit allowance.
What does MentionLab do better?
More than most competitors on this site, and it is worth being specific. Sentiment, which Semji does not have at all, and the whole narrative layer behind it — themes, per-entity comparison, mention-type distribution. Share of Voice as a named metric. The raw answer: Response History keeps the full text of every run, which is often what decides a head-to-head trial. Semantic response tags that classify past answers by meaning. Shopping, YouTube and fan-out surfaces we do not measure. A 187-market availability matrix including Google AI Mode. Six named CSV exports for a team landing this in a warehouse. A published starting price, and a documented signup that analyses your website and proposes prompts for you — where Semji publishes no price and has no self-serve path at all. Their metrics glossary also states the denominator of every rate and volunteers where its own numbers mislead, which is a standard our documentation should be held to.
How does pricing compare?
MentionLab publishes floors rather than prices: Launch “From €250/month” for “1.500 credits / month”, Scale “From €450/month”, Enterprise “Contact us” — every figure prefixed From, no annual figure published, and every CTA on the page booking a demo or contacting sales. A credit is a token, and tokens per run = queries × AI models × iterations, so cost moves with markets and engines rather than with seats. Semji is quote-based, because scope varies with workspaces, markets and AI usage, and because the platform covers production as well as measurement. That is a fair point against us: MentionLab tells you the order of magnitude before the call, and we do not.
Measure everything, then change something
Book a demo and we will run your own prompts through both halves of the platform — the visibility report, and what Semji does about it.
MentionLab is a trademark of MentionLab SRL (mentionlab.io), which is unrelated to the similarly named company at mentionlab.ai. Semji is not affiliated with, endorsed by or sponsored by MentionLab SRL. This comparison is based on publicly available information as of 1 August 2026; products change, so check both vendors' current documentation before deciding.