0101 · One-time · 10 days
Diagnostic Sprint
By application
A baseline AEO and AI ops scan across 5 AI models, plus a full technical, content, and off-site audit. You get a scored report and a 60-90 day roadmap. Most brands start here.
AI search is the
new front door.
We make sure your brand is the answer ChatGPT, Claude, Perplexity, Gemini, and Google AI give.
AI visibility and AI ops advisory for brands in Singapore and Australia.
Built by the team that ran Omnidesk to SGD 10M+ for 8 years.
How we engage
AEO + AI ops engagements.
Request access →
Most Common
0202 · 3 months · Most common
Pilot
By application
Two interventions per month across AEO and AI ops. Weekly citation tracking. Monthly report. Designed to prove the model works on your brand before a longer commit. Converts to Retainer at month 3 or ends cleanly.
0303 · 6 months · Retainer
Retainer
By application
Full execution of the intervention framework across AEO and AI ops. 2-3 interventions per 4-week cycle. Access to the sector playbook library. For brands ready to own AI search and operate AI-native.
Q2 cohort opening. Limited slots. Pricing on application.
Built
We ship what we sell.
Before advising on AI operations, Firstpress ships them. Here are the products, tools, and case work running in the wild.
Product
The Conversion Engine
8 Claude Code skills built from running a SGD 10M+ DTC brand. Email CRO, retention flows, ad strategy, conversion copy, launch planning. Install them, point them at your store, and get operator-level thinking on Monday.
Live on Gumroad · USD 297 · 100% eval score vs 79% baselineSee what's inside →
Tool
SG AI Grants Calculator
Interactive tool that helps Singapore businesses find AI-related grants they actually qualify for. Free, no email required to start.
Live at firstpress.ai/grants-calculatorTry the calculator →
Tool
BattleBot
AI-powered competitive intelligence for ecommerce sales teams. Pulls live competitor data into the sales conversation.
Internal tool
Case
Klaviyo Revenue Recovery
Segmentation rebuild that suppressed 29,000 dead profiles and recovered email revenue for a large DTC account.
Operator case work
App
Leo's First Words
Toddler flashcard PWA preserving dialect, stretched to 15 decks and 300+ cards. Part of a 52-week app-shipping challenge.
10 apps shipped this year
Every intervention Firstpress recommends is one we've already run on our own stack.
The wedge
If buyers ask AI, will they hear about you?
01
Extractability
Can AI pull a clean answer from your site? Structured content, schema markup, answer-first pages. The first gate.
02
Authority
Does the model trust you? Author bios, consistent entity signals, publication history. The second gate.
03
Corroboration
Do Reddit, press, and peers agree? Off-site mentions, reviews, citations in authoritative sources. The third gate.
The playbook
Part of the framework.
Technical, content, off-site. A preview of the levers we pull.
AI crawler access
Make sure GPTBot, Claude-Web, PerplexityBot can fetch you.
llms.txt deployment
Publish a canonical index of what AI should read first.
Schema markup
FAQ, Article, Organization. Machine-legible claims.
Internal linking
Answer hubs connected by meaningful anchors.
12 more revealed under engagement.
Who
I'm Ben. I run this practice.
I co-founded a Singapore DTC brand in 2018 and ran it as CEO for 8 years. We scaled to 8-figure annual revenue without outside capital.
I now work as operator-in-residence for a small number of brands building the same way. Firstpress is where I ship that work publicly: AEO, AI ops, and the infrastructure that makes both run.
After 8 years operating a real DTC brand, the parts of the stack moving fastest are whether AI tells a buyer your name, and whether your team runs AI-native day to day.
Writing
Signal Before Noise.
Field notes on AI ops, AI builds, and the infrastructure behind them. Launching soon.
Coming soon
AI ops in real businesses
How an 8-year operator runs AI inside a working company, not a demo.
Coming soon
AI builds & build-in-public
Working logs from shipping skills, tools, and apps, week by week.
Coming soon
AI infrastructure
The stack I run for myself and clients. Models, agents, hooks, memory.
Founding readers
Get on the list before the doors open.
Founding readers get the methodology drops first, the field notes the public posts won't have, and a private invite when the Q2 cohort opens. No spam. Unsubscribe any time.
- · Early access to every new essay before it goes public
- · Operator-only field notes (numbers, not summaries)
- · First call when the Q2 cohort opens
Before you ask
Questions worth asking first.
- AEO (Answer Engine Optimization) is the practice of making your brand the answer a generative AI gives. When a buyer asks ChatGPT or Perplexity "who makes the best standing desk in Singapore," the model picks one or two names. Those are the only brands in the room. AEO is how you become one of them.
- AI citation patterns move on a 4 to 8-week lag from intervention. The framework is not a one-month project, it is a body of work. Six months is the floor where the signal compounds enough to measure. Shorter engagements sell certainty we cannot deliver.
- No one can guarantee model outputs. Anyone who offers that is lying, and you should read it as a red flag. What we guarantee is the work: named interventions, weekly citation tracking, monthly reporting. The signal is yours to see, every week, whether it's moving or not.
- Most SG agencies bolted AEO onto an SEO playbook and rebranded. We built the practice category-first, alongside the AI ops layer most agencies don't touch. The operator behind it (Ben) ran an 8-year DTC brand as CEO to 8-figure annual revenue, he reads buyer intent the way a founder does, not the way an account manager does.
- The market we understand deepest is Singapore DTC and B2B services. If your brand sits outside that we can still run the methodology, but buyer nuance and cultural signal work better when we are close to the market.
- Every engagement starts with a baseline: citation rate across 5 models, position score, competitive index, model coverage. Every week during execution we re-run the measurement. Every month we ship a report. If a lever is not moving the metric, we swap it. No vanity numbers.
- Anonymized and aggregated intervention results may be used in public research and writing. Your brand name, domain, and specific data are never shared without written consent. You can opt out entirely in the MSA.
- We are happy to run alongside them. AEO is a distinct stack: different crawlers, different content primitives, different off-site plays. Most SEO agencies do not yet have the measurement infrastructure to prove AEO impact, so we often work as the missing half of the stack, not the replacement.
Contact
Already running a brand and want to talk?
Email me directly. Otherwise, the early-access list above is the cleaner path until the cohort opens.