Inference Advisor · Private beta
Meet Ergo. A marketing advisor in your Slack.
Inference’s frameworks deliver more coherent marketing — pressure-testing plans, monitoring strategic drift, and steering the calls that matter — with a human in the loop for the judgement you can’t prompt for.
First cohort — limited places
@ergo rewriting our homepage hero. Again. “The AI-native CRM for modern teams” — better?
11pm on a Tuesday. Still in the room.
The problem, stated properly
You’re already making marketing plans with AI.
But the AI you’re using agrees with everything, remembers nothing, and is trained to revert to averages.
You don’t need another AI tool. You need on-demand CMO-level judgement in the loop.
A generic LLM has read everything, so it believes nothing. It has no view on your positioning, no memory of why you made the calls you made, and no stake in whether the work is right. That’s fine for drafts. It’s a liability for decisions.
What Ergo is
Senior operator judgement, steering the work 24/7.
Ergo lives in your Slack. It knows your positioning, your ICP, your competitors, your brand voice — and the reasoning behind every decision you’ve worked through together.
It runs Inference’s actual frameworks — positioning, messaging, audits, competitive strategy — the same methodology used in five-figure engagements. Not marketing knowledge scraped from the internet. A working method, encoded.
And it has a view. Ask it for ad copy before your positioning is settled, and it will tell you no — and tell you why.
@ergo board wants the AI assistant feature launched next month. Can you draft the launch messaging?
Before I draft anything — this collides with a decision you made in March. You moved away from “AI-powered” language because all twelve competitors lead with it (decision log, 14 Mar). Two routes:
1. Lead with the outcome — close the books in three days — and let the AI be the how. Consistent with your positioning. Strongest play.
2. Lead with “AI assistant” and re-enter the noise. I’d argue against it, but it’s your call.
My recommendation is route 1. Want the messaging house for it by tomorrow morning?
ok but does the March call still hold? feels like the market’s moved since then
Why it’s different
Three things a generic AI
can’t give you.
Not features. Structural differences — the things that can’t be added to a chatbot with a better prompt.
Judgement, not agreement.
It has a point of view and it will disagree with you. A generic LLM is trained to please — it will find merit in whatever you bring it. Ergo gives you a straight recommendation, sequenced strategy-first, and the reasoning to challenge it.
A direct line to a human.
When a call is too important to leave to software, Ergo says so — then packages the question, with context and its own read, and routes it to the Inference team. A considered answer comes back within 24 hours. Experienced humans stand behind the hard decisions.
A living brain.
Inference’s methodology and worldview, encoded as working frameworks and refreshed as the thinking evolves. Not a model frozen in its training year. Not a chatbot scraped from a blog. The practice behind it is live — and so is the advisor.
Memory, trained on your context.
The foundation under all three: Ergo learns your business during onboarding and remembers every decision you make together — and why. Generic AI forgets. Ergo keeps the strategic spine.
See it work
Proactive guidance, not just chat.
No blank chat box. Ergo ships with named routines — ways of working you can run on day one.
positioning teardown — acme.com
1 / What you’re really saying.
“The AI-native CRM for modern teams” names a category, not a position. It tells me what shelf you’re on — not why I’d pick you off it.
2 / Sameness check.
“AI-native” appears in 9 of your 12 nearest competitors’ heroes. Three also use “modern teams”. You are camouflaged in your own category.
3 / What only you can claim.
The 14-day implementation. No competitor can say it. It’s currently in your footer, in 12px grey.
4 / Proof gaps.
“Loved by thousands” — no logos, no numbers, no names. Either prove it or cut it.
5 / Questions to resolve.
Two for your founders, one for your last five lost deals. Then we rewrite the hero — properly.
campaign critique — q3 launch plan
Diagnosis.
This is a channel plan wearing a strategy’s clothes. Spend is allocated before the core message is chosen — that’s backwards.
What’s generic.
The “thought leadership” track could belong to any company in your category. No thesis, no enemy, no stake in the ground.
What’s missing.
Nothing here uses your strongest asset — the win-loss data from March. Your customers tell you why they buy. The campaign ignores them.
Recommendation.
Hold the paid spend two weeks. Fix the core message first. I’ve drafted the revised sequence below — and yes, I know the board wants it sooner.
messaging pressure test — homepage draft v3
Weak claims.
“Blazingly fast” — unfalsifiable. “Enterprise-grade security” — table stakes presented as differentiation.
Borrowed language.
“Single source of truth” is in 7 competitor homepages. You decided in April to own “the operating layer” — this draft has drifted off it.
ICP mismatch.
Para 2 speaks to engineers. Your buyer is a RevOps lead. They don’t care about the API — they care about Monday morning.
Verdict.
Strong bones, wrong skin. Three rewrites attached, each laddered to the positioning. Take the second one.
market weather — w/c 8 june
1 / Your largest competitor repriced.
Usage-based, 30% cheaper at entry. This will squeeze you from below within a quarter. Your counter isn’t price — it’s the implementation-speed claim. Time to make it loud.
2 / The category language is shifting.
“Agentic” is replacing “AI-powered” in your space — 4 launches this fortnight. Don’t chase it. You own an outcome, not an adjective.
3 / One thing worth doing this week.
Your March win-loss themes match the repricing story exactly. A pointed comparison page would land right now. I can draft it today.
Who it’s for
Built for the team that’s
outgrown winging it.
Ergo is for you if
And honestly, not for you if
Where it sits
A senior brain on tap, for the price of one day of one.
Ergo isn’t priced like software, because it isn’t software with a thin wrapper of marketing knowledge. It’s senior judgement plus strategic drafts, on retainer — a different category from the AI tools below it, and deliberately distinct from Advisory above it, where Thom does the work himself.
One flat monthly rate. All frameworks, accumulated context, proactive monitoring, and human escalation included. Pricing to be determined during beta phase — get in touch if you want to put Ergo through its paces.
Above · collaborations
Advisory & Fractional CMO
Thom doing the work,
agents supporting
this · ergo
Ergo, in your Slack every day
Ergo advising, monitoring, and shaping work 24/7.
Humans on demand for the hard stuff.
Below · the tool ceiling
Generic AI tools
Fast, agreeable,
and they forget you
The private beta
The first cohort is forming.
Cohorts are capped in beta phase because every escalation gets reviewed by one human, and that bandwidth is finite. Applications are reviewed personally; places go to the teams Ergo can genuinely support.
What beta members get
What’s asked in return
You’re on the list.
Thom reviews every application personally — you’ll hear back within a few days. No pitch, no pressure, no drip sequence.
FAQ
Fair questions.
How is this different from just using Claude or ChatGPT?
A general assistant — even one loaded with your documents — is trained to be agreeable and to average across everything it has read. Ergo is built around Inference’s marketing methodology and worldview, then loaded with your business during onboarding. It runs frameworks rather than improvising, disagrees with you when it should, and brings a human in on the calls that matter.
What happens when we join?
Onboarding starts with context, not configuration: a structured intake, your existing material — site, decks, brand docs, win-loss notes — and a working session to load your strategic picture. Within the first week you’ll be up and running and seeing Ergo arrive with its first impact.
Where does our data live?
In an isolated, per-client instance — no shared model training, no cross-client bleed. A data processing agreement covers what’s stored and where, and if you leave, your context is exported to you and deleted within 30 days.
How fast are the answers?
Most arrive in the thread within minutes, any hour you’re working. When something deserves a human’s judgement, Ergo says so explicitly and routes it to the Inference team — with an answer back within 24 hours.
What does it cost?
One flat monthly rate — no seats, no usage meters, everything included. The number comes up early in the application conversation, with founder pricing locked in for the first cohort.
What if we outgrow it?
Then it’s done its job. The natural next step is Advisory — Thom doing the work directly, with everything Ergo has learned about your business carried across. No re-briefing, no starting over.
From the annotations archive
How — and why — I encoded twenty years of marketing judgement into an agent.
The thinking behind Ergo — what AI genuinely changes about marketing judgement, what it doesn’t, and why a human stays on the hard calls.
Want an advisor with your back — and an opinion?
Apply for the first cohort. If it’s a fit, Thom will tell you. If it isn’t, he’ll tell you that too — that’s rather the point.