This is the methodology piece. There is a real process for getting to a unique position worth defending, and most companies have skipped it. Here are four ideas that have shaped how we think about positioning in the age of AI.

It’s not what you make, it’s where you stand.

In simplest terms, positioning is what the product does and who it is for. Taken one step further, it is the organizing idea for everything your company does, the product, the user experience, the sales motion, communications, even hiring. Get it right and decisions start to flow like water.

Subscribe now

One iconic example of positioning was from ad legend David Ogilvy. Rather than positioning Dove as a detergent bar for men with dirty hands (like everyone else), Ogilvy positioned the soap as a beauty bar for women. He believed that “the most important decision is how to position your product.” His campaign for Dove soap became one of the most successful and enduring ads of all time.

That was in the 50’s, so what’s changed since? Getting from a brief to a locked position used to take a creative agency the best part of 6 months, endless rounds of research, reviews, and revisions. It also cost a lot.

Today, you can prompt your way to a new tagline, positioning statement, or even a logo over your morning coffee. The bad news is that by missing the messy middle, you bypass the part that actually differentiates the business. The result is a sea of sameness, visually and verbally.

BLANDING 2.0

We’ve seen this before. Today’s brand AI homogenization is what the industry called ‘blanding’ in 2018, the millennial pastels and friendly tone of voice you couldn’t escape from. The excellent essay ‘The Age of Average’ called this out, specifically the way brands seem content to converge on a single safe mean. Back then, every brand was built on ‘purpose’ and wanted to make an ‘impact’. The sale of transparency pioneer Everlane to climate destroyer Shein last month felt like a closing note on the era of brands built on this movement.

So here are four ways to stand out today.

1. START WITH CONSTRAINTS

In a world of abundant options, the most useful thing you can do is take some away.

In David Epstein’s latest book ‘Inside the Box’, he argues that boundaries are what produce breakthroughs. From Dr. Seuss restricting himself to only using 50 words to Victor Hugo hiding all his clothes so he would focus on writing vs going out. The constraint was the engine.

For companies, the brief has been a strong guide for this. During my time at Saatchi & Saatchi, I was taught to strip away and subtract as much as possible, to distill the ask. The tighter the brief, the better the work. It was always uncomfortable to reduce it down, but it always paid off.

Most AI-led work fails here, because the sources are endless, so the brief becomes wide, and the output is generic.

Volvo is proof of this at scale. In a car market obsessed with speed and style, it chose the unfashionable corner, safety, and owned it so completely that the word still belongs to them decades later. Whilst everyone else was competing for the same words, Volvo took one nobody wanted and made it theirs. The constraint was the strategy.

2. GO IRL

If your impulse is to prompt your way to a positioning, then your response will be average. To change things up, you have to go after unique sources. Two are our religion:

Primary Research

Interviews, event interactions, even customer complaints, which are unscrapeable by chatbots and uniquely yours. In the early days of Airbnb, founder Brian Chesky went door to door asking hosts what the product actually felt like to use. Paul Graham called this ‘doing things that don’t scale’. The result? You get a much deeper understanding that you’d never achieve purely online.

Ethnography.

This is the art of observing customers in their own world rather than asking them direct questions. Intuit’s founder Scott Cook built a multi-billion-dollar company on this practice, one he called ‘Follow Me Home’, waiting in the store until someone bought the product, then sitting in their kitchen to watch them actually use it. The idea is that what people do and what they tell you they do are rarely the same thing.

3. Change the Perspective, Change the Answer

Most positioning is done from inside your office. Where actually a sharper position can come from looking through someone else’s eyes.

There is one classic framework we use that positions every brand through four lenses: company, category, consumer, and culture (the 4 C’s). The obvious answers often live in the first three: what we make, who we compete with, who buys us. Culture is the one that pays off, because it tells you what’s happening around the business, not inside it.

Another approach is older and simpler. In 1985, with Intel’s business collapsing, Andy Grove asked co-founder Gordon Moore a question: “If the board fired us and brought in a new CEO, what would he do?” Moore said, “get us out of memories”...So they walked out the door, came back in as outsiders, and abandoned the position that built them (memory chips) and claimed a new one (microprocessors). Same company, just seen with a fresh pair of eyes.

This is also why we tell clients to remove “ER” words. Better, faster, smarter, cheaper. Every one is a comparison that tethers your position to your competitors. The head of Rolex understands this and when recently asked how the watch business was going, replied “I don’t know, I’m in the luxury business”. He refuses to be in the category everyone else is fighting in. That is the move that makes you stand out.

4. Truth & Tension

Once you have some positioning options, the last step is to figure out which one is actually worth holding onto.

The first test is truth.

Start by asking: what is the customer actually buying? Nobody buys AI in isolation. People usually buy for one of the following reasons:

• a problem solved (Uber)

• a task made easier (Calendly)

• a founder’s unique perspective of the world (Elon)

• a promise of a new version of themselves (GLP1’s)

Hold each potential position against these, and you’ll reveal the strongest options.

The second test is tension.

Most brands try so hard to be inoffensive (see 2020’s playbook for ref) and therefore blend in and remain invisible.

Your positioning has to be willing to be disliked by people it is not for.

Liquid Death is a great example here. ‘Murder your thirst’ + provocative heavy-metal branding in a category of soft, clean-living is impossible to ignore. And now they own a sizeable slice of the entire category.

Water Bottle Interest Over Time - How Bold Positioning Pays Off

Try to optimise for distinctiveness. When you create tension, it’s a sign you are onto something.

Now what next?

One caution before you fall in love with your new positioning. The sharpest position does nothing until the product, experience, and operation can actually deliver on the promise it makes.

The tools we have today have made it free to sound like everyone else, which makes it more valuable to stand somewhere specific.

Next…a deeper dive to find out whether every brand in the AI era is starting to look and sound the same, and what to do about it.

From farm towns fighting back against data centers to students booing commencement speakers, it’s clear something has shifted.

During his Odyssey press tour, Christopher Nolan put it plainly, “I’ve never seen a technology that’s been so successfully adopted by Wall Street and by investors and by tech companies that the public has so thoroughly rejected.”

This is a story about how fast the public perception has turned, and how slowly most companies have responded.

The brands paying attention are already ahead.

Anthropic won this year’s Super Bowl cultural conversation by showing chatbots interrupting personal conversations, mirroring their earnest tone and overeagerness. Claude moved from 41st to 7th on the App Store within days.. They have quickly become the anti-AI AI company. By taking a position against the drift of their own category, they are winning.

They are not alone. Nikon’s Natural Intelligence campaign mimics the format of AI image prompts and then reveals each photograph was taken by a real person in the real world. Its tagline reads: don’t give up on the real world. While writing this, Substack introduced a feature to show how much content was written by AI (this post was verified as 100% human), these are good signals.

A new class of start-ups exists to strip AI out of the experience. Sinceerly removes the telltale signs, the em dashes, and even suggests typos. Extra’s founder put it simply on LinkedIn during their launch: “The fastest way to make a consumer AI product feel boring in 2026 is to keep reminding everyone it’s AI.”

The companies closest to consumer trust are racing to remove the residue of the machine, and it’s bearing fruit.

Meanwhile, walk through any business conference and the inverse is happening. Every one has a variant of “AI-powered.” Every deck has the same messaging. Even the logos are drifting into the same floating orb shape.

It is worth being clear about what’s actually happening here. On one side, the media is writing about AI on every page, because AI is the story right now. However, that doesn’t mean a company has to put AI on every page of its website. Those are two different jobs. One is reporting, the other is positioning. The companies confusing the two are paying for it in trust.

The market is correcting in public.

331 S&P 500 companies cited “AI” on their Q4 ‘25 earnings calls. Through the first ten weeks of ‘26, those companies returned an average of 1.5%. The companies that did not mention AI returned 5.6%. For three years, the word operated as a premium phrase on earnings calls, but now that has flipped.

The public is further ahead. In March 2026, NBC published a report where AI achieved a rating lower than ICE! Among voters aged 18-34, the rating was minus more than double the national figure. These are people who grew up with the technology and that brands most want to reach.

Then there is the cultural reception. Friend.com’s subway campaign in New York was defaced within days of going up; commuters scrawled “surveillance capitalism” and “AI wouldn’t care if you lived or died” across the posters. The New York Times ran a piece on the backlash. This was a reality check for founders deep in the San Francisco bubble.

It is worth saying that the AI industry has not helped itself. The dominant narrative has been about job losses and the end of humanity, reinforced by billionaires being interviewed in flattering (and extremely long) podcasts by interviewers with no stomach to push back. The technology has been positioned, notably by its own architects, as an existential threat, and, surprise, surprise, people are not up for it. Now, under a Zuckerberg-esque PR makeover, they are walking it back.

Two markets, the equities market and the consumer market, both correcting on the same word at the same time.

Brand is dead. Brand is everything. Both are right.

Two of the sharpest commentators in business have read this moment in opposite directions. Scott Galloway has argued that brand is dead, that the cost of building an audience has collapsed and that traditional brand advantage is dissolving. Paul Graham argued the opposite: that we are entering ‘The Brand Age’, when technology converges and every product is functionally identical, brand becomes the only remaining differentiator.

Both are right. Brand has been compressed. Today, it’s less about awareness and more about the specificity of your position. When every AI product can do roughly the same, the position “AI-powered” stops carrying weight. The cost of saying it is now greater than the cost of staying silent.

Ultimately, it’s not about whether to use AI, everyone uses AI. The question is whether the audience can clock it in the work you put out into the world.

Can you “clock it?”

A few weeks ago, I sat in the audience at an event and listened to a speech that had clearly been written by AI. He started with a negative to positive (it’s not x, it’s y), then outlined 3 examples (I was x, y, and z). By the end, most people were on their phones. It lacked any personal touch or roughness; it was too obvious and polished. Ultimately, people just don’t talk like that in real life.

Look through the hellscape that is LinkedIn, or even your emails and you’ll see it everywhere. What was once a fun experiment has started to become disrespectful.

Great work subverts your expectations. That is why people clock AI, because there is no friction in it. LLMs are particularly bad at making content that surprises you, because they are designed to produce the average of what has been said or done before.

This is now measurable. Cornell researchers sampled 20,000 stories from four leading models and found that the same eleven words appeared in 88% of them. The most common protagonist, a man usually called Elias, now fronts AI-generated books and health guides sold on Amazon.

We are all developing this pattern recognition faster than the tools have learned to evade it.

In our work at Atmosphere, Scott and I have stripped this kind of language out of client websites, ads, landing pages and decks, and watched the numbers move. At the very least, the exercise forces us to ask the crucial question of what is actually unique about this business. “AI-powered” isn’t an answer.

The presence of a machine implies the absence of care, and the absence of care is what the customer now penalises.

The discipline behind the work

There is a craft behind getting positioning right in this market, and it is increasingly important. In The Future of Focus we called this moment The Imprint: the taste, judgement, and conviction that remains in the work after the tools have done their part. The Imprint is what cannot be generated.

The discipline behind it is both intellectual and emotional. It involves inputs that the tools cannot scrape from the web, gleaned from in-depth interviews and ethnography, and it involves the willingness to hold a tried and tested framework and a gut feeling at the same time.

In the AI era, advantage will accrue to the companies bold enough to refuse the easy path and the obvious output. Everything else is the residue.

This is the first in a two-part series.

Next week: the science behind getting positioning right, the inputs and processes that unlock uniqueness and the tests that separate a real insight from a confident sentence.

Entire companies are being formed around agents such as Sierra for customer support. Amazon is building synthetic environments so agents can learn new skills and improve. Your founder friend runs an entire business development function off her Mac Mini next to the dog bed under her desk.

How can you get your own?

Today we’ll try to demystify the concepts. It’s deliberately simple and aimed at the business leader who wants to get started.

What is an agent?

Agents aren’t magic.

They’re models with context, capabilities, instructions and a system for keeping them on track. The reason people use agents vs. building a new piece of software:

• Agents can figure things out. Pursue an objective without prescribing every step.

• Agents can take action. Use tools, interact with your systems, actually do things.

• Agents can do things you can’t. Call skills you don’t possess.

• Agents operate at a massive scale. Create 1,000 variants to your 5.

Core Pieces of Agentic System

• Context & Data: what the agent knows

• Model: how it reasons

• Tools & Skills: what it can do and how it knows how to do it

• Harness: how the whole thing runs

• Evals: how you know it actually works

Context & Data

Agents need context and data to act. The nuance here is how agents access data. They can use normal sources like direct database reads or APIs. They can also read documents that you provide such as a Word Doc, PDF, Markdown files etc.

The most common standard today is MCP (Model Context Protocol). Anthropic created this. A slightly incorrect but easy way to think about this is a wrapper around an API. It gives the agent the ability to call tools and access data. You might see an MCP connection to Notion or Google Docs for example. It’s a standardized format an agent understands.

The MCP connection will give the agent access to your data in another platform.

Model

We covered this in How to Choose Your Frontier Model. This is your agent’s brain. Frontier models, hosted models, open source, it doesn’t matter. It’s how your agent reasons.

Tools & Skills

A tool gives an agent capability. A tool might be to read a PDF or build a spreadsheet.

A skill tells an agent how to perform a task. A skill could be to perform an analysis and write a report. A skill is ultimately a file. Anthropic’s documentation has a concise description.

“Each Skill packages instructions, metadata, and optional resources.”

Here’s an example skill.md file. It’s easy to read and straightforward.

You’ll face a trade-off between how granular to make a skill. The more granular, the more precise. But that means headaches in debugging and maintenance. It’s more art than science.

Harness

The harness is exactly what it sounds like. It’s the runtime environment around the agent. It gives all the instructions on how to use skills, what to do / not to do, enforces policy, and importantly handles memory and prescribes actions.

Models have limited context windows (the amount of data you can shove in a prompt). They forget things between sessions and even chats. The harness can keep a running log of memory and compacts that into usable pieces that can be reinserted at specific times.

Evals

The most important part of your whole agent build is the eval loop. An eval loop takes the output of your agent and runs a series of repeatable tests.

The eval loop can check for simple things like data validity, scan for inappropriate language, and enforce policy rules. It creates a feedback loop for the agent. The eval tells the agent whether the output was good or bad and gives data signals for future improvement.

How do I build an agent?

All the major platforms have agent builders now – Google, Salesforce, Microsoft, OpenAI, Anthropic etc. For developers you can use CrewAI, LangGraph or n8n, which tend to be popular choices.

A prototype is as simple as following a tutorial. You can build one in 20 mins. At Atmosphere we did early prototypes by using Claude Code (an agent) inside of Replit (another agent) to build our own agent. A very matrix experience.

What surprised us was how well they worked. We combined AI web scans and data structuring with proprietary frameworks and instructions gleaned from 20 years of experience. We could produce much higher quality outputs than raw models alone. That process took a few months but was worth the effort in development and eval.

How do I move to production?

The easy part is getting an agent built. The harder part is getting an agent to run reliably at high quality, economically.

Moving into production requires several steps that need engineering expertise:

• Context & Data: ensuring secure connections into your production software environments. For example connecting to your Microsoft or Salesforce data.

• Model: understanding data policies of your model choice, for example no-training, enterprise SLAs, any regulatory or compliance needs from your company.

• Tools & Skills: building persistent MCP connections with reliable data flow. Documenting and debugging all your skill.md files.

• Harness: documenting your governance layers and ensuring the enforcement produces a reliable, compliant output via your evals.

• Evals: a system for storing your evals and an interface for users at your organization to run continuous evals for reliability and improvement.

That’s a wrap on the series. We’ve moved from what AI actually is to mapping workflows, choosing tools and models, and now agents. Hope you enjoyed and best of luck with your AI implementation. If we can be helpful please reach out, hello@atmospherepartners.co

You’ve decided to consolidate your AI stack. But as models become platforms, the switching costs explode. A few lines of Python is now months of work and potentially tens or hundreds of thousands of dollars. How do you make your bet?

Before we get too far, I wanted to give a pointer to Tomasz Tunguz from Theory Ventures. His analysis of platforms, models, costs, inference is the best in the industry. If you want to go deep, I’d highly recommend following him on LinkedIn.

At Atmosphere, our main model is Claude, then we use ChatGPT for specific tasks, Perplexity and Gemini for search, and Grok for realtime.

We’re not covering the whole world of model delivery. Large platforms like Microsoft Azure, Amazon Bedrock and Salesforce offer plenty of capability and model choice, including homegrown models. Data platforms like Databricks and Snowflake offer AI tooling around enterprise data. Inference platforms like together.ai are great for Open Source.

We’ll keep this focused on the Frontier Labs.

Evaluation Criteria

You want to know what the model can do, how much will it cost for a given level of performance, what you can build on it, and if it meets your enterprise requirements.

1. Capabilities - domains of expertise (coding, finance, writing, analysis) and modalities (text, voice, image, video)

2. Economics - output speed (tokens/second) and price (in/out per 1m tokens)

3. Build Environment - agents, tooling, data connectors, SDKs, integrations

4. Enterprise Readiness - data security, SOC2, HIPAA, procurement, partner ecosystem

We’ve stopped paying attention to MMLU (Massive Multitask Language Understanding) scores. These are quantitative benchmarks that researchers used to demonstrate the capability of models. They don’t matter as much anymore because: i) the big labs keep leapfrogging each other., ii) MMLU doesn’t account for user interaction or engagement, and iii) smaller focused models can outperform the large ones at specific tasks.

What matters most if what you’re trying to do and your own eval.

How the Frontier Models stack up

This table is highly variable. I wrote the first draft a couple of weeks ago and essentially everything changed. Grok launched a very competitive voice model. All the pricing shifted.

To read the table, capabilities advance with every model release. Economics depend on the model choice; we’ve listed the low and high. Output speed is the same. Build environment is quickly evolving with more capability every week. The major platforms are reaching parity on enterprise adoption with friendlier terms such as no-training and ZDR (zero data retention) and better compliance.

The enterprise ecosystems are also expanding. OpenAI launched the OpenAI Deployment Company and Anthropic established a $1.5b joint venture for deployment. Every major consultancy now has a partnership with the big labs.

Open source

Cost has become a major issue with CEOs and CTOs looking to govern and rein in spending. A rogue agent can burn through thousands of dollars before someone notices.

The world of Open Source has shifted significantly in the last few months. While there were previously only 2-3 models (Llama, DeepSeek, Mistral) there are now ~10 serious contenders to evaluate.

Prices can be dramatically less expensive than the large models, roughly $0.05-$1.75 per 1m tokens input and $0.20-$4.40 per 1m tokens output.

That’s 5-20x cheaper than the higher end Frontier models.

But again the point isn’t about high-level cost, it’s about the specific thing you’re trying to do and whether the model you’re deploying is good at the actual problems you need to solve.

Open Source Models

As of writing the major Open Source models are:

• DeepSeek (DeepSeek); strong reasoning at lower cost

• Qwen (Alibaba): broad, general purpose

• GLM (Zhipu AI): open-weight quality leader

• Kimi (Moonshot AI): long-context and coding

• MiniMax (MiniMax): efficient, agent-centric

• Llama (Meta): broad capability

• Mistral (Mistral): European leader

• Gemma (Google): Google’s Open Source effort

• GPT-OSS (OpenAI): cheap and efficient

You can run these models on 3rd party platforms like together.ai or on your own computer through a product like Ollama. If you’re going this route, you’re already at the leading edge of AI implementation.

To bring it all together, you have plenty of good options. And your evaluation should be tailored to the specific use case and workflows we mapped. That’s how to make your bet.

The last one in this series: the Agent Overview

I’m not hating on cold outreach. We use it effectively. I’ve gone through demos, bought software, even found partnerships from cold starts.

Just the complete lack of relevance. AI tool automation at its worst. Zero context on our business. I don’t even know what we’d do with 200 leads.

The Paradox of Choice

So we’ve all had those experiences but want to use AI for upside, what to do?

Today we’ll be more editorial than usual. This is a view from our experience and the clients we work with. Most sit at Seed to Series C, and everything AI.

Adoption of AI tools is highest where results are verifiable, mistakes can be caught and corrected, and the human remains in control. The spot we call the Imprint.

We discussed previously how to be surgical about choosing your AI implementation workflows.

AI Tools

Figuring out where to start can be daunting.

Separating adoption (value to customer) from a press release, even harder. When we dug deeper, the common thread was about compressing the production process in the middle of work, while still giving enough agency to humans for the final polish or review.

What we hear from operators is they use the following:

• Coding: Claude Code, Cursor, Codex

• Prototyping: Lovable, Replit, Bolt

• Voice: ElevenLabs, Deepgram

• SDR automation: Clay, Apollo

• CRM: Attio; Lightfield and Monaco emerging

• Sales & Meeting Intelligence: Gong, Granola

• Presentations: Gamma, Beautiful.ai

• Content Creation: Runway, HeyGen, Hedra, Higgsfield

• Legal: Harvey

One category is fairly universal, the AI assistant. For that, it’s Claude, ChatGPT, and Gemini. These are the individual walking around tools.

Several companies we’re excited about - Ando.work, Highlight.ai, Intangible.ai, and Blomma - solve different problems but fit the same pattern. AI multiplication around the human Imprint, not replacement.

Platforms

These will come as no surprise. Companies mostly consolidate onto one platform. Many use multiple models to compare outputs, or put them in adversarial conditions to challenge responses and improve performance.

• Claude: best all in one platform for enterprise use cases

• OpenAI: integrations improving and Codex showing strong progress

• Google Gemini: best for Google Workspace, tooling more complex

• Microsoft: best for MSFT shops, improving feature set with Copilot

• Salesforce: great for CRM-heavy use cases, requires skilled admin

Build your own

If you go out on your own, we recommend, right now, building on the Anthropic or OpenAI platforms. They have the most robust set of tools and are emerging as platform leaders.

The most sophisticated operators build completely from the ground up, either through a platform above or by integrating the tools in the background. This is an extremely effective mechanism but requires that you have your own AI engineering specialists or a 3rd party implementation partner. Common tools range from orchestration frameworks like LangGraph and the OpenAI Agents SDK, to automation layers like n8n, deployed on infrastructure like Vercel.

Perhaps you’ve heard of the magic of an agent running on a Mac Mini, reading your inbox, posting on LinkedIn, writing your emails? Sure, but only for the 0.01% of AI enthusiasts.

Next up: we’ll talk through how to choose your Frontier Model.

Because work isn’t disappearing. It’s shifting.

More in planning, control, review, polish. Humans still do a remarkably good job of sorting out the nuance. That’s not going away.

How top operators use tools, platforms, and bespoke systems

For the Future of Focus we talked with 30 operators to understand their stack. The dividing line was clear: human judgment. Generic, low-judgment tasks are where today’s AI tools shine. The second question is how bespoke the solution needs to be. Off the shelf tools work quite well for generic tasks but start to crack as you need more customization. That’s where platforms and build your own start to become real alternatives.

A simple way to place your workflows.

In our previous post, we outlined how to pick apart workflows within your company and create an ROI calculation for each.

Now we want a simple framework to understand how to choose tools, platforms or build your own. You want to plot your workflows onto the matrix. You can use the precise values from the ROI calculation to accurately place these into the matrix.

1. How much judgment does the workflow take? The more nuance and human judgment required, the more you move to the right. You may use prototyping for the idea but the end product is distinctly human.

2. How customized is the workflow? Generic workflows like searching for email addresses or building customer lists sit at the bottom in tool or platform territory.

After this you need to nudge up and down based on your technical capacity or budget, and compliance and regulatory. Most companies work through a migration path of tools first, platforms with more experience, then custom solutions as ROI and capability increase. You can quantify this but it becomes more of a judgment call than a simple heuristic.

One golden rule: never use AI to automate your core competency. If you’re paid to be the best at it, use AI for the context around the work, not the work itself.

For example, enterprise sales companies can automate the market research, contact sourcing, meeting notes, and action item creation, but you never want to have high-volume cold outreach for a deal that requires warm intros, serious multi-threading or relationship-based selling.

The Atmosphere example

At Atmosphere our major workflows span strategy articulation, relationship management, content creation, deliverable creation, scheduling, and product development. We had the benefit of starting the company from scratch so had to make every SaaS and AI buying decision against a real budget and expected P/L.

We first mapped our workflows into the matrix. That gave us a sense for which pieces of work we wanted to apply AI to. In the discovery, we knew we ultimately needed a consolidated system that gave us context-lift for the keep it human piece. But deliverable creation is a ton of market research, competitive analysis, content structuring, and client context. The work is intensive and requires high AI leverage.

Because we have technical aptitude from 20+ years in consumer tech, we decided to build. We created a system of agents with an orchestration layer called Convector. The invisible engine in the atmosphere that powers flight.

Our pathway was a migration. The blue bubbles on the map above.

1. We started with buying tools (Claude, ChatGPT) to see if that could solve it. Too verbose and generic. The work required high judgment and was bespoke. The output wasn’t good enough and worse took us more time in correction than just writing it ourselves.

2. We moved into the prototype zone. We built an initial version on Replit to see if we could do better than Claude or ChatGPT on its own. Turns out we could. By tying our proprietary framework and business process IP to the data produced from AI we could deliver a step-change better deliverable than Claude on its own. It also clearly beat doing the work purely by hand.

3. We built a system of agents. We productized Convector and migrated everything over to real infrastructure. We built an eval and self-reinforcement loop so the system would get smarter and give us compounding leverage. More on this in a few weeks.

In the end it was about understanding our workflows, planning where to apply AI, and building only what we needed to.

Next post: our picks for platforms and tools.

The cost of failure is high. Think in terms of the 1-10-100 rule; $1 in prevention is $10 in correction and $100 in failure. Adapted for the AI era this gets worse:

$1 in design is $10 in engineering, $100 in tokens, and $1,000 in failure.

Catching flaws in design is cheap and easy to fix. Letting them slide into production can be catastrophic.

The Imprint

In The Future of Focus we talked about the Imprint: the point where human judgment and taste reign supreme. It’s the most critical concept for your AI implementation plan.

At Atmosphere we view AI as work augmentation. We’ve done the modeling on our actual hours spent and put sharp lines around where we expect high leverage from AI and where output is almost entirely human.

As a small firm, we need 2x blended AI leverage.

Our content, for example, is entirely human written. We use Claude for revision and proof edits, and Claude Design for visuals based off a hand curated brand kit.

We don’t use AI for any initial drafts or even concept suggestions. The quality is too low and doesn’t communicate in a voice we feel is authentic. That caps the AI leverage and we’re comfortable with the trade-off.

Three Core AI Use Cases

The AI implementations we’ve worked on all boil down to the following:

1. Personal productivity: research, writing, editing, coding, email etc.

2. Workplace productivity: augmenting specific workflows*

3. Product development: embedding AI into customer-facing products

* Many companies seek full workflow replacement. In our experience this almost never works.

Discovery

Map and quantify your workflows across the organization. Note that if you’re using AI to speed up a particular function (Sales), you need to ensure the rest of the organization stays in sync (SalesOps, Finance). If you don’t…

We call this the feral otter. The AI enhanced workflow starts to gobble up the rest of your company. Problems get worse and messier before any improvement.

The discovery work is painstaking but straightforward. Here’s a simple plan:

• Internal interviews with users on workflows (3-5 per area) depending on size

• Interview synthesis and workflow quantification

• Trade-off analysis of AI implementation vs. standard operating procedure

• Tool selection and platform choice

Your workflow scoring system needs to take several things into account at a minimum:

1. How many people touch it

2. Hours per week consumed

3. How repetitive the task is (score of 1 - 10)

4. How much human judgment is required (score of 1 - 10)

Where to apply AI

High-volume, high-repetition, low-judgment tasks pay back the fastest with AI. The graveyard for AI deployments is filled with projects that require human judgment and taste.

As a practical example, we worked with a company that needed to process ~200 leads per month to get 40% qualified with a close rate of 20%. That volume is ~2 SDRs going full tilt or ~$40k per quarter. It’s why SDR automation is such a hot topic.

TAM building and data enrichment is widely available and CRM systems have automated sequences (email etc.). That part is straightforward. We’ll cover the tools next time.

The problem was the upper funnel increase created a downstream mess in alerts, CRM data quality, human sales follow up and billing and revenue visibility. The solution was to step back and build AI enhancement into the other areas to calm the feral otter.

The lesson: own the system for your AI implementation

The end output is a structured quantification of your workflows (how many people, how repetitive a task, time spent on average etc.). The map is a spreadsheet and set of workflow diagrams. Every task in a row quantified with an assessment of AI lift. The top rows become your AI implementation plan for the quarter. The full list is your annual roadmap.

To give another practical examples for high-touch, BD type relationships. Lead sourcing gets high leverage, whereas follow up is quite specific and has a much higher judgment score. Repetition and judgment are coded scores, 1-10.

Change Management

Once you have your map, you need to spend several cycles deciding on the areas where AI can have maximum efficiency and ROI for your organization. But a plan that no one adopts is slideware. Which brings us to why most AI deployments are failing.

You need to socialize with your team. Your team members need to be bought in and frankly should be excited about removing tedious and repetitive work.

How to do this? Bring people in early. Make them part of the discovery process. Be transparent about where the judgment lines sit and where you believe that humans add the most value. You’ll be surprised during the discovery process about how nuanced work is and how hard it actually is to get a machine to repeatedly produce the right answer. It’s why you hired good people in the first place!

Build in review cycles where you sit together, review the roadmap, discuss tradeoffs, and most importantly, listen to concerns.

Fear will nuke your deployment. If the message gets delivered as ‘your job is going away’ adoption will be zero. AI without a human operator is essentially useless.

Next post: Which AI tools and platforms to actually use (or build your own)

Atmosphere (us) will do a multi-part series over the next 4 weeks to give you the language and use cases to make sense of it all, as well as some tutorials for implementing AI at your organization.

We will not cover everything. Niche experts go deeper in all of this. There is a plethora of click-bait content of ‘Be an AI expert in 30 days!’. We’ll do something different. A practical guide, direct from the operator’s seat. Tips you can actually use.

What is AI anyway?

It’s not a stupid question. Executives are all well informed, smart, read the AI news, but it’s still confusing. Today’s AI has morphed from the traditional textbooks of Machine Learning, Natural Language Processing, Computer Vision, and Generative AI. You’ll still hear these terms. But today’s AI in practice has coalesced around a smaller number of concepts.

Machine Learning

Traditional Machine Learning (ML) is very much alive. These are systems that learn from data to make predictions and decisions. They have wide adoption and are best when you have lots of historical data. For structured tasks, purpose-built ML models can be more accurate and consistent than throwing the problem at a general purpose LLM.

Dumping your data into a Frontier Model, like ChatGPT, and hoping for the best is a common (and bad) mistake. The systems, people, and techniques are different. There are other tools you should be using like R and Jupyter notebooks.

The operator question is whether the problem is better solved with ML or Frontier Models. Even that line is blurring because the models can run real analysis and code. The mistake is pasting data and trusting the answer rather than directing the model to run a specific ML analysis.

Frontier Models (LLMs)

All of the attention has shifted here. ChatGPT, Claude, Gemini. They generate new data, reason, produce text, create images and videos. They are increasingly bundled. When people say ‘AI’ in a meeting, 99% of the time they are referring to these models.

They are probabilistic, which is why they will confidently tell you the US was founded 357 years ago even as we limp toward our 250-year celebration. Then instantly backtrack when you point out the error. That’s hallucination.

They’ll also give you a different answer for the exact same question because they are non-deterministic (you might also hear the word stochastic). Again don’t worry so much about the definitions. We just want you to be aware of the vocabulary.

It’s why high-risk situations (legal review, accounting audits) almost always have a human in the loop. You shouldn’t trust AI to close your books autonomously or give you a precise legal opinion. I could make the same argument around software architecture and data, which is why people hate on vibe-coding.

Two concepts that come hand in hand with LLMs are training tokens and parameters. Training tokens are a measure of how much data was used to train the model. Order of magnitude trillions. Parameters are the resulting weights that give the model intelligence. A bunch of numbers organized into matrices. Order of magnitude hundreds of billions. Smaller models can be in the single (7b) or double digit billions (15b) if tuned to a specific domain.

Parameter count is still a good predictor of intelligence but it’s no longer the whole story. You’ll hear terms like Reinforcement Learning from Human Feedback (RLHF), Fine-Tuning, and Retrieval Augmented Generation (RAG). These are techniques that can help a well-trained model beat a bigger raw one.

Specialized Models

These models are trained to do a specific thing extremely well. Examples here are text to speech models (ElevenLabs), speech to text models (Deepgram) or OCR models that can parse a PDF or scan a receipt. Frontier models get the press. These models do half the actual work.

AI Inference

This is the act of actually serving the model to a user. Processing a prompt, generating a response. The labs combine this all. What you see on your monthly bill is the tokens consumed and the aggregate cost. Typically in cost per 1m tokens, segmented by output and input tokens (Google, OpenAI, Anthropic etc.). But if you’re building your own with an open-source model you need to think carefully.

Inference can be a huge driver of cost and latency. More simply put, you actually have to pay for the model and compute. If you forget everything else, just remember this is where your AI bill will actually sting.

Agents

Where the industry momentum is heading. It’s a lengthy topic that we’ll cover in a separate post. You will hear these words in conversations.

• Harness: the wrapping infrastructure and evaluations

• Skills: the specific capabilities and guidelines for what the agent can do

• Tools: functions the agent can call (data fetch or CRM record updates)

• Guardrails: the instructions of what an agent can/can’t do (no processing payment)

• Model Context Protocol (MCP): the common standard to connect AI to data/tools

Agents can do things instead of just generate content which is why the industry is leaning this way. The terminology can vary a bit across providers so treat this as directional.

The alphabet soup decoded

You may hear all of these terms spat out in a meeting. You don’t need to memorize any of this. Each has a different underlying technology. Don’t spend the time trying to understand the difference between transformers, neural networks, or diffusion models unless you’re a tech leader. You won’t need it in normal operating decisions. Worse, when you use it incorrectly you instantly lose credibility.

The state of the art changes every day so by the time you learn it you’ll already be behind.

• Large Language Models (LLMs): reasoning, code/content generation, ChatGPT

• Text to Speech (TTS): converting text into speech, ElevenLabs

• Speech to Text (STT/ASR): converting live speech into text, Deepgram

• Text to image (T2I): take a prompt and synthesize a digital asset, Midjourney

• Text to video (T2V): take a prompt and create a video, Veo

• Image to video (I2V): take a reference image and generate a video, Runway

• Video to video (V2V): adjust a video for lighting, character etc., Runway

• Multi-modal LLMs: unified models across text, audio, images, and video, Gemini

• Graphics Processing Unit (GPU): massively parallel chips for training/inference

• Tensor Processing Unit (TPU): Google’s chips optimized for neural network math

• Pre-training: the initial phase of training a model on tons of data

• Post-training: everything done to the model afterward to make it useful and safe

Next up: how to start on your AI implementation. Subscribe so the next one hits your inbox!

Subscribe now

We set out to understand which tools actually stick and why. So we went straight to the source, 30+ founders, executives, and investors at the leading edge of AI.

We expected to find smarter systems, better prompts, and a few useful tools. What we found was something else entirely.

A small group of operators has pulled so far ahead that their workflows bear almost no resemblance to those inside even the most established companies. They’re not using better tools, they’ve reorganized how they work entirely.

This report is a map of what the front line now looks like. Four discoveries from the leaders who have crossed to the other side, and the principles that separate them from everyone else.

Interviews included individuals from:

The Frontier Has Split

Across our interviews, operators fell into three distinct groups.

At the top: builders, operators who have constructed proprietary AI systems trained on their own data, customers, and workflows. Every interaction makes these systems harder to replicate.

Below them: power users, sophisticated operators running deep tool stacks but still manually bridging the gaps between them.

And then: adopters, experimenting with AI tools, but layering them onto existing ways of working. Inside some of the most prominent companies we spoke with, there is no unified stack, no shared best practices. Individual employees are often paying for their own subscriptions.

The distance between these groups is growing because the builders’ systems compound while everyone else restarts from zero with every new tool.

Across all three tiers, one pattern held: AI-native tools are largely companion tools. They augment existing workflows rather than replace them.

So which companions earn a permanent place?

The Operator Stack

We asked operators at the leading edge of AI a simple question: what tools do you actually use?

Across interviews, the same structure emerged.

This list reflects the preferences of the operators we interviewed, not a comprehensive market map.

Every operator we spoke to had strong opinions about tools. But the deeper we pushed, the conversation shifted, from what they use to how they act.

Subscribe now

Today’s advantage isn’t artificial. It’s knowing where AI ends, and you begin.

AI has given everyone the tools to produce good work. But when good becomes free, great becomes more valuable. We found the difference between the two isn’t more AI, it’s knowing when to put the human back into the machine.

We call this The Imprint.

The moment where judgment overrides generation. Where the human hand re-enters the work and makes it unmistakably yours.

Every operator, team, and company has an Imprint. The question is where yours sits. Step in too early and you lose the leverage AI provides. Step in too late and the work becomes generic.

The people pulling ahead aren’t defined by their AI stack. They win by having the strongest sense of where their Imprint is.

Finding it isn’t intuitive, but across more than thirty interviews, four discoveries surfaced that reveal how the best operators have found theirs.

01

The Last 10% Is the Only 10% That Matters

AI compresses production, but expands judgment. The last 10% isn’t just the finishing touch. It’s becoming the most valuable work in the economy.

There’s a phrase in engineering called the ‘march of the nines’. It takes the same effort to go from 9% to 90% as it does to go from 90% to 99%. The closer you get to completion, the harder each improvement becomes.

AI has shifted work in a similar way. Until recently, the grind from blank page to draft was where most of the work lived. AI has collapsed that middle. A deck that took two weeks now takes two hours. Research that consumed an afternoon now takes minutes.

What AI hasn’t changed is the effort required to turn good work into great work. That effort lives at the end.

A creative agency founder, working with leading AI companies, uses Midjourney and ElevenLabs to get concepts in front of clients three times faster than the old model of scripts, stock imagery, and decks. But they won’t use AI for the finished work. “It doesn’t allow enough room for the human.” AI is “essentially a sketching tool, it gives us direction to where we’re going.” The final output? “The last 10% is always done in real life, with a human touch.”

An engineering lead at a major cloud platform described building infrastructure tools that reach 90% accuracy, then hitting a wall. “If you can’t trust it 100%, you’re not gonna give it full control. That last 10% is the hardest to bridge.”

But here’s what separated the builders from everyone else. They’re not guessing when to step in, they’ve engineered it. Confidence thresholds at every stage, back-testing on every output, and human override built into the system by design. The Imprint isn’t instinct for these operators; it’s infrastructure.

The CTO of a Series A company, whose entire product is generative AI, framed it most directly: “When execution becomes simple, judgment becomes essential.”

AI has made the first mile nearly free, but the last mile just got longer, and now it’s the only mile that differentiates you. One of the most advanced builders we interviewed put it plainly: “Are we going to crank it all out faster? Or are we going to keep the time we have and raise the quality instead?” That choice — speed or craft — is quietly becoming a dividing line.

The Human Signal

AI is very good at producing the right answer. But people rarely respond to the right answer. They respond to the human one, rougher at the edges, more raw, more emotional, harder to replicate. The part that can only come from the last 10%.

The evidence is everywhere. Vinyl hitting a billion dollars in U.S. revenue last year, driven by Gen Z. Film photography sales tripling in five years. Rosalía singing German opera at Berghain. Even the recent demise of the metaverse. In a frictionless world, friction becomes the unmistakable signal that a human was here.

Rosalia, a Spanish popstar, singing German Opera = Friction as a signal.

The same reflex showed up in our interviews. “AI waters down ideas.” “People are craving human touch.” “It breaks at the final layer.” One more round of prompt roulette rarely makes the output better. It makes it blander. That’s motion masquerading as progress.

The conventional explanation is that AI lacks taste. AI has taste, but its default taste is bad. The chatbot’s voice is earnest. The image generator’s output is emotionally vacant. People clock it immediately. This matters because when someone senses AI in the output, their trust decreases. Not because the information is wrong, but because the presence of a machine implies the absence of care. The top operators know this and treat AI’s default taste the way a director treats a first rehearsal - raw material, not the performance.

The last 10% isn’t just where the human adds value, it’s where the human removes the residue of the machine. The time AI frees up is being reinvested here.

AI closed the gap between a blank page and a competent draft. It did not close the gap between good and great.

02

The Discipline of Less

Your edge doesn’t come from what you add, it comes from what you remove. The best operators are using constraint as a forcing function.

Norway is the most decorated Winter Olympic nation in history, despite having 60x fewer people than the USA. What explains this? The philosophy is counterintuitive. Athletes spend the majority of their time at low effort, only pushing when it counts. They don’t chase new methods or fads. They only add something when they’re certain it improves the system.

The Norwegian Method in action: Klæbo, 6x golds in ‘26, most in Winter Olympic history.

The same pattern showed up in our interviews. In a market flooded with new tools, the most disciplined operators rejected most of them. They don’t reach for new, they reach for useful.

Constraint as Strategy

An AI-native founder described his philosophy as “delete and eliminate as many unnecessary things as possible.” He’s deliberately focused on constraints to force sharper thinking. His company runs what would traditionally take over a hundred people with fewer than thirty. Not because he couldn’t hire, because he chose not to. This deliberate friction didn’t slow him down; it made him innovate.

A founder building enterprise AI tools took the same approach, “to use the least amount of tools possible,” so that everything the company learns compounds in one place. Notion for documentation, Linear for product, Cursor for code, Claude Projects for everything from sales coaching to recruiting. Fragmentation kills the knowledge that makes a company smarter over time.

They chose Claude over ChatGPT because of editorial sensibility. They chose Granola because it disappears into the workflow. The selection criteria wasn’t “what can this do?” It was “does this earn a place in my day?”

This is what separates builders from adopters. Adopters chase the new tool. Builders go deeper on the system they already have.

The best operators are treating tool selection the way the top investors treat their portfolio: conviction about what to hold, discipline about what to cut.

03

Control is the Currency

Why is it that in every major technology wave, the early leaders rarely win?

From PCs to the internet to mobile, the pattern repeats. Google didn’t win the web by building websites. It won by controlling discovery. Apple didn’t win mobile by inventing the smartphone. It won by controlling the ecosystem. The technology commoditizes. The tools converge. And value shifts to the systems built on top, to the people who control how the technology is used.

The same shift is now underway in AI. The main LLMs can do roughly the same things at the same level. The tools built on top are converging too. If your AI stack looks like everyone else’s, it gives you the same advantage as everyone else. Which is no advantage at all.

It’s no longer about which tool to use. It’s which systems to build.

A founder scaling a global AI marketplace built multiple custom voice agents using ElevenLabs. One onboards over a thousand customers a month, 65% of all interactions. A second makes one thousand outbound sales calls daily. Within two months, they outperformed every team member except the founder. These aren’t off-the-shelf products, they’re proprietary systems trained on his business, his data, his customers. Every interaction makes them smarter.

“We tear up ideas, poke them,” said a veteran at one of the world’s leading AI platforms, describing how his agent teams work. Confidence scoring at every stage. Humans plugged in at every threshold. The system isn’t autonomous; it’s adversarial by design.

A creative technologist took it further: three different LLMs critiquing each other’s code before any human sees the output. “By themselves, they’re not good enough.” He built a cultural monitoring engine on the same principle, scraping sport, social media, and internet culture, delivering a daily Slack digest so analysts spend their time thinking, not searching.

Then there’s the CEO who turned Claude Projects into a hiring screener, a sales coach, and a deal copilot. “A second brain that is not biased.” She makes every final call. The system doesn’t replace her judgment, it sharpens it. And competitors can’t reverse-engineer it by buying the same tools.

The gap isn’t between companies that use AI and those that don’t. It’s between those that rent AI tools and those that own AI systems. The first group gets efficiency. The second gets the moat.

04

The Missing Layer

Every operator we spoke with described the same friction. Not a missing tool. A missing architecture. Right now, what’s filling that gap is you.

The tools work, they just don’t talk to each other. There is no shared context between apps, no continuity between sessions, and no awareness of what you were doing five minutes earlier in a different tool.

So you became the integration layer. Every time you copy context from one tool to another or re-brief AI on a decision already made, you’re doing the work the architecture should be doing. The most capable people we interviewed are spending real cognitive load on this infrastructure.

The founder of one of the fastest-growing AI companies is using a custom platform as “duct tape to hold over ten tools together”. A technical lead at a major cloud platform put it simply: the goal is an AI-native operating system, “because we’re still using Mac and Windows from the nineties.”

What operators are asking for is remarkably consistent: “A single layer that sits across everything”, “holds context”, “understands priorities”, and “acts proactively without being prompted”.

But the builders in our interviews aren’t waiting for it to arrive, they’re creating it themselves. Custom orchestration layers, tools stitched into proprietary workflows, context held together by force of will and engineering. The solutions aren’t elegant, but they compound, and that’s the point. The operators who control their own systems aren’t just moving faster on today’s work; they’re building the connective tissue that makes tomorrow’s work better.

The market has started to respond. Several of the operators we interviewed are already using products like Highlight to build context layers across their desktop, and open protocols like MCP are creating infrastructure that lets AI models reach into external tools and data for the first time. But these are early moves. What operators are actually describing is more fundamental: one surface, one context, one point of control that makes the architecture invisible. It doesn’t fully exist yet.

One quiet signal points to where it’s heading. Several operators told us that voice has become their default input mode. They don’t type prompts, they talk. Removing the friction of writing allows ideas to move closer to the speed of thought. This is a signal of what the missing layer actually looks like: not another app, but persistent intelligence you interact with as naturally as a conversation.

The companies that crack this won’t win by building a better point solution. They’ll win by making every point solution work together, with the human at the centre.

The Advantage Isn't Artificial

We set out to understand what the best operators are actually doing with AI. The pattern was clear.

The operators ahead of the curve know exactly where AI ends and they begin. They’ve condensed their stacks, built proprietary systems that compound with every interaction, and refused to automate the last 10% that only a human can do. That’s the Imprint.

As AI makes competent output ubiquitous, the advantage comes from what only a human can add: judgment, taste, conviction.

Every operator in this report is building something that didn’t exist two years ago. So are we. Atmosphere partners with founders scaling AI, health, and climate companies. If you’re navigating the same territory, we’d like to hear from you.

hello@atmospherepartners.co

This report was researched, drafted, and produced using a custom AI system built across capture, transcription, analysis, and editing, with human judgment at every stage.

Created by Atmosphere

Atmosphere is a growth partner for companies at inflection points, from initial go-to-market to enterprise transformation. We’ve spent the last two decades at companies large and small. Google, Pinterest, Saatchi & Saatchi, our own startups, and over a dozen as growth partners. We exist so the most ambitious ideas in AI, health and climate scale the fastest.

atmospherepartners.co

Subscribe now

Views from Atmosphere

We’ve spent the last two decades building and scaling companies at Google, Pinterest, Saatchi & Saatchi, Character.AI, our own startups, and over a dozen more as growth partners. Exits. IPOs. Product launches across 50+ countries. Platforms scaled from early traction to hundreds of millions of users.

That experience gave us a specific vantage point.

We sit at the intersection of technology, brand, and commercial strategy, close enough to the frontier to see what’s coming, deep enough in the work to know what’s real.

We also spend our time talking to the founders and operators building what’s next. Views from Atmosphere is where we share that perspective.

What you'll get

Bi-weekly essays and analysis drawn from our work with founders across AI, health, and climate. What's scaling, what's stalling, and why. Plus longer research reports when a topic demands it.

Who we are

Scott Coleman led growth at Google, Pinterest, and Character.AI before its acquisition. He advised the New York Times on AI and has scaled platforms from zero to hundreds of millions of users across 50+ countries.

Henry Simonds is a three-time founder with two exits who started in advertising at Saatchi & Saatchi before moving to the operator side, leading go-to-market and growth across AI, climate-tech, and consumer brands.

Views from Atmosphere is published by Atmosphere, a growth operating partner for companies at inflection points. If you found this interesting, share it with someone navigating the same territory.

Get in touch: hello@atmospherepartners.co