Is SOMIN's 13-minute brief claim real?

It's measured -- and independently documented in Google Cloud's customer story on SOMIN (cloud.google.com/customers/somin-ai). Median time from brief upload to a stack of pre-scored creative concepts is 13 minutes, with campaign planning and creative analysis running up to 12× faster than manual workflows. The Mothercare, KPI Media, and Click2View case studies document the before-and-after numbers.

Where does SOMIN's data come from?

Public ad inventories from Meta, Google, TikTok, and YouTube. Engagement signal from live in-feed surfaces. First-party audience data from licensed partners. SOMIN does not scrape gated content or buy data it can't audit. Data handling is aligned with the GDPR and Singapore's PDPA, on Google Cloud infrastructure.

Trusted by P&G · Audi · Ferrari · Mothercare · McDonald's

Stop blending in.
Start seeing what nobody else can.

Real-time competitive insight and audience intelligence, transformed by GenAI into market-ready creative ideas -- for brand teams, agencies, creator programs, and AI agent stacks alike. From weeks to 13 minutes per brief.

Built for

Brands Agencies Creators AI Agents

Book a call See the platform

Live

CONTENT LIBRARY · PEPSI INDIA · 691 CREATOR POSTS

The full SOWISE agent trace pulling Pepsi's brand kit, creating studies, and polling 691 creator posts.

Live

BRAND TRACKER · WHITESPACE VS. PEPSI

Six culinary content territories scored against Pepsi India's own output -- volume tells you who's loud, whitespace tells you where to move.

Live

PERSPECTIVES · 3 PERSONAS CARRY THE PLAN

Every tracked creator post clustered into the personas actually driving the conversation -- ranked by volume, not guesswork.

Moments publishing calendar mapped to cultural moments for Pepsi India

Live

MOMENTS · PUBLISHING CALENDAR, PEPSI INDIA

Cultural moments mapped straight to a 13-week publishing calendar -- Onam, Diwali, and the weekly lead-pillar slot, live.

Live

PROPOSITIONS · INSIGHT → BRAND IDEA

Insights become testable propositions automatically -- three editorial pillars, weighted and ready to brief.

Expressions view, rendered deck and scored creative concept library

Live

EXPRESSIONS · DECK + CREATIVE, READY TO SCORE

A proposition becomes a rendered deck and a scored concept library -- on-brand and ready for review.

13min

Per brief

75% lower

Research cost

4audiences, 1 engine

Brands · Agencies · Creators · Agents

The reality today

You're spending 40 hours on research that should take four.

Briefs go in late. Insights arrive after the campaign window has already closed. Local teams guess. Pitches lean on last quarter's data. Creator rosters scale faster than results. SOMIN ends that workflow -- whichever seat you sit in.

The old way

Two analysts, three weeks, one slide deck that lands too late.
Competitor scraping that misses 80% of the in-feed conversation.
Persona work built on 2-year-old qual, refreshed once a year.
Creative briefs that read clean but die at scoring.

With SOMIN

Live category, brand, and audience intelligence -- refreshed daily.
Every competitor ad parsed, scored, and indexed by emotion and proposition.
Audience tensions surfaced from real engagement signal, not guesses.
Concepts pre-scored for CTR likelihood before a dollar moves.

The platform

Four modules. One end-to-end workflow.

From the moment a brief lands to the moment a concept is ready to test, SOMONITOR does the unglamorous work -- and hands you the strategic part back.

Brand tracker

Put your finger on your competitors' pulse.

Every ad, every post, every piece of content from every competitor -- tagged, scored, and clustered by what's actually working in the feed.

Live category map

See share of voice, share of engagement, and share of attention move week-on-week.

Decoded creative

Every ad parsed for proposition, persona, emotion, format, and call-to-action -- automatically.

White-space alerts

When a competitor abandons a tension or a category opens up, you hear about it the same day.

WHITESPACE MAP · PEPSI INDIA LIVE

691 creator posts decoded across six content territories, scored against Pepsi India's own output.

SOMIN SOWISE whitespace map for Pepsi India: culinary heritage and home-cooking territories scored WHITESPACE, sports fandom scored BLOCKED

From brief to concept

Four steps. Thirteen minutes.

The full SOMONITOR loop -- from the moment a brief lands to a stack of pre-scored concepts ready for review.

Step 01

Brief lands

Upload a brief, paste a paragraph, or pick a category. SOMIN classifies the ask and pulls the relevant competitive set.

Step 02

Insight surfaces

Brand Tracker and Perspective Studies surface the tensions, propositions, and personas worth pursuing.

Step 03

Concepts build

Expressions generates content directions tagged for channel, audience, and creative format.

Step 04

Concepts score

Ad Scoring predicts CTR likelihood and engagement before media spend moves.

Four playbooks

One engine. Four ways to win.

Brand teams defend every campaign with data. Agencies pitch with unfair speed. Creator programs turn rosters into revenue. AI agents run the whole loop end to end. Same category engine, four motions.

For brands

Defend every campaign with data, not opinion.

Brand teams walk into concept reviews with live category evidence instead of stale data. Campaign development accelerates from six weeks to 13 minutes.

More granular segments

See the brand playbook →

For agencies

Pitch with unfair speed.

Strategy and account teams get an instant, data-backed edge -- the moves competitors are making, the white space they're missing, and a brief that's differentiated before the deck opens.

13×

Faster research

75%

Lower pitch-prep cost

£M+

Accounts won on insight

See the agency playbook →

For creators

Turn your creator roster into a growth engine.

SOMIN maps the product white space your category leaves open and auto-builds tailored micro-briefs -- in each creator's own voice -- so affiliates convert instead of just posting.

Creators activated convert

See the creator playbook →

For AI agents

Six agents. One category engine.

Strategy, paid media, influencer & GMV, social, CRM, and reporting -- each agent runs a marketing job end to end, grounded in live category signal, with a human sign-off at every gate.

Kill-switch activation

See the AI agents →

Case study · Mothercare

Mothercare saved 25+ hours a week while keeping every campaign on brand.

Singapore parents, four competitor brands, 4,000+ ads. SOMIN replaced three weeks of manual research with a continuous, decoded category view -- and gave the local team back its time.

Without SOMIN, segmentation of expecting mothers' needs would not be as comprehensive. SOMIN became the team's research department.

<b>Ms. Xueting Lin</b> · Head of Marketing, Mothercare Singapore

Time saved

80%

Reduction in weekly content research time.

Cost

74%

Lower research spend on the same brief volume.

Granularity

3×

More granular audience segmentation in every plan.

Coverage

4k+

Competitor ads decoded and indexed continuously.

More results

The numbers our customers live by.

SAMY Alliance01

12× faster research, 78% CTR-prediction accuracy.

SAMY won a multi-million pound Nightcap account on the strength of pitch insight produced in days, not weeks.

12×faster

78%CTR accuracy

£Maccount won

Read on Google Cloud Customers →

KPI Media02

Briefs in 2 days. Costs cut to a quarter. +$5K/mo retainer.

Singapore-based performance shop took every brief from a one-week cycle to a two-day cycle, and grew client retainers in the process.

5/5vendor rating

1wk→2dcycle

+$5kmonthly

Read full case study →

Click2View03

Reporting time down 9× and analysis 6× across the studio.

Content marketing studio compressed reporting from 9 hours to 1, and category analysis from 18 hours to 3. New revenue uplift projected at +100--400%.

9×reporting

6×analysis

400%upside

Read full case study →

Havas Media MENA04

Day-one onboarding, no learning curve, every team on it.

Performance Director Branislav saw the team adopt SOMIN inside a single onboarding session -- from first pitch deck to live use, same day.

Day 1adoption

0training days

100%team usage

Read full case study →

Blak Labs05

Analysis 10× faster. Monthly cost down by $4K.

Independent strategy shop replaced 10-hour analysis cycles with 1-hour cycles, and cut the monthly research bill from $10K to $6K.

10×analysis

$4ksaved/mo

1hper cycle

Read full case study →

Neo36006

3-person team to 1. Half the cycle time. Same output.

Using SOMONITOR since April 2024, Neo360 reduced briefing cycles from 4 hours to 2.5 and ran them with one person instead of three.

3→1headcount

4h→2.5hcycle

$300saved/brief

Read full case study →

In aggregate

The numbers our customers live by.

13min

Per brief

From a paragraph of context to a stack of pre-scored creative concepts.

75%

Lower research cost

Average reduction in monthly research spend across the customer base.

4audiences

One engine

Brand teams, agencies, creator programs, and AI agents all run on the same category signal.

Agentic AI

Six agents. One category engine.

SOMIN's agents each run a marketing job end to end -- strategy, paid media, influencer & GMV, social, CRM, and reporting -- the way a cross-functional team would, except the cycle takes minutes, not weeks, and there's a human sign-off at every gate. The four below are peer-reviewed and published in the ACM Digital Library; see the full six-agent lineup on the AI Agents page →

Agent 01

Tracker

Strategist

Pulls live competitive signal and decides which propositions are worth pursuing this cycle.

Meet the Strategy Agent →

Agent 02

Mapper

Planner

Surfaces audience tensions and white space the category has left unaddressed.

Meet the Influencer & GMV Agent →

Agent 03

Builder

Creative

Generates concepts and content directions tagged for channel, audience, and format.

Meet the Social Media Manager Agent →

Agent 04

Scorer

Validator

Predicts CTR and engagement on every concept and ranks the stack before review.

Meet the Paid Media Agent →

+ Social Media Manager Agent + CRM & Email Agent + Performance Reporting Agent

Orchestrated via Zapier (Silver Partner) Claude (MCP) Manus (MCP) Codex (MCP)

Under the hood

Named models. Published research. Verifiable claims.

SOMIN runs on Google Cloud, where Gemini 2.5 Pro powers large-scale ad-content understanding and idea generation, and SOMIN's own deep-learning models predict creative performance. The architecture is peer-reviewed at ACM -- the same claims you read here are documented in Google Cloud's customer story and in the ACM Digital Library.

LLM core

Gemini 2.5 Pro

on Google Cloud

Processes competitor video and ad content, clusters audiences by shared characteristics, and generates campaign insights -- the engine behind the 13-minute brief. GPT-4-class LLMs support content workflows.

Google Cloud customer story →

Prediction

SoCraft CTR models

In-house deep learning

Advertiser-level creative scoring trained on tens of thousands of labelled ads -- 78% CTR-prediction accuracy, roughly 3× the human baseline -- hosted on Google Cloud Compute Engine with Cloud GPUs. Real example, SG ride-hailing category: "Deliver with us" scored 74% (Ship), "Download the app" scored 38% (Iterate), "Grab for Pets" scored 3% (Cut) -- before a single dollar of media moved.

Every insight traces back to the ads and engagement signals that produced it -- the "Against Opacity" architecture published at ACM Multimedia 2023, extended by SOMONITOR.

Public ad inventories from Meta, Google, TikTok, and YouTube; licensed first-party partners; no gated scraping. Data handling aligned with the GDPR and Singapore's PDPA, on Google Cloud infrastructure.

How we source data →

Read the research Google Scholar · Prof. Aleks Farseev ACM DL · SOMIN Platform Paper Google Cloud · Customer Story

Backed by

The benchmarks that matter.

Gartner

Cool Vendor in Marketing AI

Google

AI First Programme

IMDA Singapore

PSG Pre-Approved Vendor

Zapier

Silver Agentic AI Partner

Questions

What teams ask first.

How is this different from a social listening tool?

Social listening tells you what's being said. SOMIN tells you what's working -- every competitor ad parsed for proposition, persona, emotion, and predicted performance, then connected to the audience tensions driving the engagement. The output is concepts ready for review, not a sentiment dashboard.

Is SOMIN only for brand marketers, or does it work for agencies and creators too?

All three, on the same category engine. Brand teams use it to defend campaigns with evidence (brand playbook), agencies use it to pitch faster with a data-backed edge (agency playbook), and creator programs use it to territory-map a roster and auto-brief affiliates in their own voice (creator playbook). See the "Four playbooks" section above, or jump straight to the page for your seat.

What's the difference between using the SOMIN platform and using the AI agents?

The platform (Brand Tracker, Perspective Studies, Expressions, Ad Scoring) is the workspace a strategist or planner drives by hand. The six AI agents run those same jobs autonomously end to end -- strategy, paid media, influencer & GMV, social posting, CRM & email, and reporting -- with a human sign-off gate at every step and a kill switch that activates in under 60 seconds. Same category signal underneath either way.

Is the 13-minute claim real, or marketing math?

It's measured -- and independently documented in Google Cloud's customer story on SOMIN. Median time from brief upload to a stack of pre-scored creative concepts across our active customer base is 13 minutes, with campaign planning and creative analysis running up to 12× faster than manual workflows. The Mothercare, KPI Media, and Click2View case studies all document the before-and-after numbers in detail.

Where does the data come from?

Public ad inventories from Meta, Google, TikTok, and YouTube. Engagement signal from live in-feed surfaces. First-party audience data from licensed partners. We do not scrape gated content and we do not buy data we can't audit. Data handling is aligned with the GDPR and Singapore's PDPA, on Google Cloud infrastructure. The Singapore ride-hailing example above is built the same way -- public ad-library data, decoded automatically, no client data.

How does SOMIN track competitor ad performance?

SOMIN's Brand Tracker continuously pulls every public ad and organic post competitors run across Meta, Google, TikTok, and YouTube, then uses large language models to decode each creative into a structured object: proposition, persona, tension, emotion, format, and call-to-action. Each ad is scored for predicted engagement and mapped into a live category view -- share of voice, share of engagement, and white space -- refreshed daily across 11+ markets.

How does AI reduce campaign research and pitch preparation time?

SOMIN automates the manual parts of research -- collecting competitor creative, tagging it, clustering audiences, and drafting briefs -- in one workflow. Google Cloud's published case study documents idea generation dropping from days to minutes and planning cycles running up to 12× faster. In practice: KPI Media cut pitch cycles from one week to two days, Mothercare reclaimed 25+ hours of research per week, and Click2View compressed category analysis from 18 hours to 3.

What AI models and infrastructure power SOMIN?

SOMIN runs on Google Cloud. Gemini 2.5 Pro handles large-scale ad and video understanding, audience clustering, and insight generation, alongside GPT-4-class language models for content workflows. Creative performance is predicted by SOMIN's in-house deep-learning CTR models -- published at ACM WSDM 2023 as SoCraft -- hosted on Compute Engine with Cloud GPUs and reaching 78% click-through-rate prediction accuracy, about 3× the human baseline. An explainable-AI layer keeps every recommendation traceable to the underlying data.

Is SOMIN's AI research peer-reviewed?

Yes. The platform and its models are published at top-tier ACM venues: SOMIN: An Explainable AI and LLM Platform for Real-Time, Data-Driven Digital Marketing Strategy (ACM Multimedia 2025), MindFuse: Towards GenAI Explainability in Marketing Strategy Co-Creation (ACM Multimedia 2025), Will AI Make Agencies Obsolete? (ACM Multimedia 2025), Against Opacity: Explainable AI and Large Language Models for Effective Digital Advertising (ACM Multimedia 2023), and SoCraft: Advertiser-level Predictive Scoring for Creative Performance on Meta (ACM WSDM 2023). The research line goes back to SoMin.ai's influencer-discovery marketplace (ACM Multimedia 2018) and personality-driven content generation (ACM WSDM 2021).

What does white-label deployment look like for agencies?

Your subdomain, your colour and logo on the dashboard, your team and clients seated. SOMIN runs underneath. Most agencies stand the white-label up inside a working week.

Which markets are covered?

11+ markets across APAC, MENA, and Europe today, including Singapore, Malaysia, Indonesia, the Philippines, Vietnam, India, the UAE, Saudi Arabia, the UK, Germany, and Spain. New markets onboard inside two weeks.

Guides & manuals

Integration · Anthropic

MCP Connection to Claude

Read the guide →

Integration · Manus

MCP Connection to Manus

Read the guide →

Get started

SOMIN Onboarding User Manual

Read the manual →

Real-Time Insights.
Sharper Pitches. Bigger Wins.

A 30-minute walkthrough on your category, your competitors, and the briefs you're working on this week -- whichever seat you sit in.

Book a call See the platform

You're spending 40 hours on research that should take four.

The old way

With SOMIN

Four modules. One end-to-end workflow.

Brand Tracker

Perspective Studies

Expressions

Ad Scoring

Put your finger on your competitors' pulse.

Four steps. Thirteen minutes.

Brief lands

Insight surfaces

Concepts build

Concepts score

One engine. Four ways to win.

Defend every campaign with data, not opinion.

Pitch with unfair speed.

Turn your creator roster into a growth engine.

Six agents. One category engine.

Mothercare saved 25+ hours a week while keeping every campaign on brand.

The numbers our customers live by.

12× faster research, 78% CTR-prediction accuracy.

Briefs in 2 days. Costs cut to a quarter. +$5K/mo retainer.

Reporting time down 9× and analysis 6× across the studio.

Day-one onboarding, no learning curve, every team on it.

Analysis 10× faster. Monthly cost down by $4K.

3-person team to 1. Half the cycle time. Same output.

The numbers our customers live by.

Six agents. One category engine.

Named models. Published research. Verifiable claims.

The benchmarks that matter.

What teams ask first.

Real-Time Insights.Sharper Pitches. Bigger Wins.

Real-Time Insights.
Sharper Pitches. Bigger Wins.