Skip to content
Vol. I · Issue No. 01The Marketing AI Lab
teemuno
From the Lab · Vol. I, No. 1

The most interesting decade in marketing's history is underway. We are the lab building for it.

Teemuno is the marketing AI lab. Custom agents, white-label platforms, brand intelligence systems — for the agencies and brand teams reimagining how the work gets done. Founded Singapore, 2023. Parent of gimmefy. Builder of Marcus, Frontier, and whatever you would like to be the fourth.

I.
The Thesis

A generational opportunity, quietly underway.

This is, by a wide margin, the most interesting decade marketing has ever had. The category has access to capability that would have looked like science fiction in 2018. Briefs can be drafted by machines that know the brand. Campaigns can be modelled by synthetic audiences before a single dollar is spent. Voice and tone can be encoded once and obeyed by every system that touches the work. A new kind of marketing organisation is being built.

And yet, inside most marketing teams, the capability has so far added a fourteenth tab. Briefs still live in Notion. Voice guidelines still live in a 2019 PDF. Tone-of-voice still lives in someone's head, who is on holiday. The capability is real and astonishing. The infrastructure to put it to work has not arrived in most rooms.

"Most consultancies will sell you 'an AI strategy.' We already built the platform you will be sitting next to."
Read the full thesis →
II.
The Practice

What the lab builds.

Six categories of work. Each in production today on gimmefy or a partner platform. Most engagements combine two or three.

I.

Custom Agents

multi-toolbrand-awareyour stack
II.

White-Label Platforms

4–6 wk buildsyour cloudyour billing
III.

Brand Intelligence Systems

structuredversionedqueryable
IV.

Workflow Engines

orchestratedmulti-stageauditable
V.

Creative Production Systems

visual · videoaudio · motionbrand-locked
VI.

Strategy & Audits

two-day · two-weekwalk-out plansact without us
See all six in detail →
III.
The Ladder

Four ways to engage. From a long lunch to a long partnership.

What used to take quarters, we now ship in weeks. The platform you will be sitting next to already exists.

I.

The Audit

A two-day embed with your marketing leadership. We map the workflow, model the three highest-leverage opportunities, and compare the vendor landscape. The plan stands alone — most clients can act on it without us.

2 days
II.

The Blueprint

A two-week design sprint. Architecture, prompts, tool selection, and a three-scenario ROI model for one high-leverage system. Hand it to your engineering team, or move directly into the Build with us.

2 weeks
III.

The Build

The production system itself, shipped inside your stack — a custom agent in two weeks, a workflow engine in three, a full white-label platform in six. You own the code on day one. We support it for as long as you wish.

2–6 wks
IV.

The Partnership

Your embedded AI lab. A dedicated team, new systems shipped each quarter, brand and model evaluations, cost optimisation. Outcomes-tied when the math is clean enough to allow it.

ongoing

Engagement scope is shaped to your operation.

IV.
Selected Works

Three platforms shipped. The fourth is being built now.

Each end-to-end by the lab. Each running in production today. The archive below lists smaller engagements as they came through.

No. 01The Flagship

gimmefy: the marketing operating system we built for ourselves.

— Built, operated, and continuously evolved by Teemuno since 2023

The platform that gave the lab its evidence base. A full operating system for marketing — vault, agents, workflows, multi-modal creative production, simulation, evaluation, and publishing. Now in daily use across agencies and brand teams in three continents. It is what the rest of this site is built on top of, and the proof that everything else is shipped rather than theorised.

Visit gimmefylabs.com →↗
See all selected works →
V.
Exhibits from the Lab

Eight things we have already built.

Each of these is in production on gimmefy. Each is available as a bespoke deployment for an agency or a brand. They are the materials we work from.

Exhibit · I
// In production · gimmefy

The Maestros

Twelve expert AI personas, each with their own knowledge base.

Persona-grade agents trained on a body of work. Each behaves in voice, references their own canon, and holds positions you can argue with. Used for strategic counsel, creative critique, second opinions, and the kinds of conversations that used to require a $4,000/hour consultant.

Sutherland on advertising. Sinek on positioning. Cialdini on persuasion. Plus nine more.

Deploy as › Bespoke personas for your team, your clients, or your brand house.

Exhibit · II
// In production · gimmefy

The Playbooks

Twenty-plus orchestrated workflow engines for marketing work.

Multi-stage AI systems that run the way a senior marketer would run them — research, draft, critique, refine, ship. Each one orchestrates several models, several tools, and several rounds of self-evaluation before producing an output a director would sign off on.

Campaign Architect. White Paper Writer. GTM Plan. GEO Audit. Demand Generation Plan.

Deploy as › Bespoke workflow engines for your team's weekly cadence.

Exhibit · III
// In production · gimmefy

The Focus Group

A synthetic roundtable of experts, on demand.

You drop a piece of work into the room. Six experts debate it. They disagree, push back, build on each other, and produce a synthesis. It costs the time of a coffee. It is the closest thing we have built to a genuinely useful "what would marketing think" button.

Brand directors. Performance leads. Creatives. Researchers. Customers, modelled.

Deploy as › Synthetic review panels tuned to your audience or your stakeholder map.

Exhibit · IV
// In production · gimmefy

The Model Council

Side-by-side multi-model comparison for any prompt.

Five frontier models, the same brief, the same brand context, three columns of answers laid out for comparison. Useful when the answer matters enough to read three opinions, and when you want to argue with the model rather than nodding along to the first one.

Claude. GPT. Gemini. Grok. Sonar. All vault-aware. All side-by-side.

Deploy as › Multi-model decision tools for high-stakes copy or strategy work.

See all eight exhibits →
VI.
In the Lab

Specimens. Real work, set on the page.

Five recent outputs from agents the lab has shipped. Click a specimen. The page shows you what came out.

Exhibit · Lab SpecimenRECORDED · OUTPUT · UNEDITED
// agents on file

Brief:A new running shoe for women who run for sanity, not splits.

Enter the lab →
VII.
By the Numbers

What the lab has built thus far.

Numbers worth using to describe an operation. Counts of things buyers understand, not internal vocabulary. Refreshed each quarter.

20+
AI models, one interface
6·6
Image and video models
12+
Expert personas
90+
Languages, voice
3
Continents in use
3
Platforms shipped
100K+
Brand-locked generations
'23
Founded Singapore
VIII.
From the Bench

Field notes from the practice.

Short, opinionated, written by the people doing the work.

No. 06
The Build · Working Paper

Custom GPTs are not custom AI. The difference, with diagrams.

The system prompt is the floor, not the ceiling. What you actually need for a brand-grade agent — tools, memory, evaluations, the lot.

14 min · May
No. 05
The Math · Working Paper

Build versus buy for marketing AI: the math nobody puts on a slide.

What we actually charged Marcus. What it would have cost them to build it. What the SaaS shelf would have cost them at year three.

9 min · May
No. 04
The Craft · Specimen

Why your brand voice keeps drifting in AI output, and the lockfile that stops it.

The drift problem is a context problem. The schema we use across every gimmefy customer to hold voice steady across a hundred thousand generations a month.

7 min · April
Read all field notes →
IX.
Frequently Posed

The questions that come up first.

Pre-answered so the first call can spend itself on more interesting subjects.

ChatGPT Enterprise gives your team a chat window. We give your business an infrastructure layer — brand-aware agents, multi-stage workflows, governance, model routing, evaluations, observability. A chat window is a tool. We design the workshop.

An Audit completes in two days. A Blueprint in two weeks. A custom agent typically inside two to three weeks. A full white-label platform in four to six. The architecture exists; we customise it for your brand, your data, and your stack. What used to take quarters now ships in weeks.

It depends on scope, and we will not pretend otherwise. Audits begin small. Blueprints are fixed-scope. Builds and Partnerships scale with what we are building together. We share three-scenario mathematics by the second meeting and pricing is on the page of the proposal, not buried in it.

You do, from day one. Repositories sit in your organisation. Models are yours, or licensed in your name. White-label deployments run inside your cloud account if that is your preference. There is no vendor lock-in. There never has been.

A Brand Vault — a structured knowledge layer that every agent on your stack reads from before it generates. Voice rules, banned phrases, audience personas, product taxonomy, brand-on examples. Versioned, auditable, owned by your team. The drift problem becomes a versioning problem, which is solvable.

Reasonable category exclusivity is available on Partnership engagements. The Audit, the Blueprint, and the Build are open to anyone — what we build for you will not be what we built for them, and the work each agency does on top of the platform is what differentiates them.

Next.js and TypeScript on the front. Supabase or your cloud for data and authentication. LangSmith for observability. Model-agnostic on the LLM side — we route across Claude, GPT, Gemini, Grok and Sonar by job. Edge-first, server-side rendered where it matters.

The Audit. Two days. It is a real deliverable — you walk out with a plan you can act on without us. If the only question is whether to build, that is the right call to take.

T

The site has shown you a fraction of what we build.

A conversation will cover the rest. Twenty minutes. No deck on either side.

Read the Field Notes

Yours, The Lab.