Rethink, Rebuild, Reprice: The New B2B AI Stack

Agentic workflows, centaur teams, and value-based pricing — the operating system for modern marketing.

Oct 09, 2025

Dear Readers, in today’s edition:

OpenAI Enters SaaS — Quietly but Decisively
Your B2B SaaS Funnel Is Obsolete
Your Brand Voice Is Probably Slop. Here’s the Fix.
Why Admitting Your Flaws Is a Marketing Superpower
Stop Chasing Keywords. Start Building a Knowledge Graph
AI Research Isn’t a Shortcut. It’s a Trap.

OpenAI Enters SaaS — Quietly but Decisively

Last week, attention was on e-commerce features inside ChatGPT. Then came another wave: Apps SDK and Agent Kit. But almost unnoticed between these headlines was the most strategic move yet- OpenAI’s direct step into software applications targeting the marketing and sales stack.

On October 1, 2025, HubSpot (NYSE:HUBS) fell 7.2 %, joined by declines across the SaaS sector - Klaviyo - 12 %, Braze - 11 %, DocuSign - 12 %. Investors reacted to a simple fact: OpenAI, long seen as an infrastructure provider, is now competing in their category.

From Infrastructure to Applications

Until recently, OpenAI sold the tools others built on. Now, it is releasing its own. The first two are already operational inside the company:

Inbound Sales Assistant – an agent trained on OpenAI’s internal knowledge base that answers tens of thousands of inbound messages each month. After a few training loops with sales reps, accuracy improved from 60 % to 98 %, unlocking millions of dollars in annual recurring revenue.
GTM Assistant – a Slack-based copilot that generates briefs and Salesforce summaries ahead of meetings. Early internal data shows around 20 % productivity gains in sales teams.

For now, both remain internal proofs of concept, but they reveal the direction clearly: AI-native SaaS built on OpenAI’s own platform - sales, marketing, support, analytics.

Market Repricing

The October sell-off wasn’t about quarterly results; it was a repricing of competitive dynamics.
For years, SaaS companies sold the interface to business data. Now OpenAI is positioning itself as the intelligence layer that operates directly on top of it.

The company’s roadmap points to applications spanning:

Front office – CRM, inbound marketing, support
Middle office – analytics and pipeline management
Back office – finance and internal reporting

What’s emerging is a full-stack platform built around agents, not traditional users.

Four Pillars of the Strategy

1️⃣ SaaS Tools (The Bait)
Products like GTM Assistant and Sales Assistant demonstrate what “AI-native” can look like in practice. They serve as internal case studies ahead of wider release—proof that OpenAI can deliver finished software, not just APIs.

2️⃣ Apps SDK (The Ecosystem)
Embedding third-party apps inside ChatGPT positions it as a daily workspace. If CRM and analytics functions move inside that environment, the platform effectively becomes the operating system for digital work.

3️⃣ Agent Kit (The Engine)
A developer framework for building and managing custom agents. It lowers the barrier for enterprises to create in-house assistants while keeping them tied to OpenAI’s model infrastructure.

4️⃣ Commerce Layer (The Monetization Stack)
By facilitating transactions triggered by those agents, OpenAI can participate in downstream value creation rather than charging only for usage.

Together, these pillars form a blueprint for an AI-native enterprise layer that could sit above today’s software vendors.

A Structural Shift

OpenAI’s expansion represents more than product diversification; it marks the beginning of AI entering the application layer directly.
Each major technology wave has followed this pattern: infrastructure players climb the stack once the base matures. Amazon extended cloud infrastructure into retail operations, Microsoft evolved from operating systems to productivity clouds. OpenAI’s ascent follows the same logic—but on AI’s compressed timeline.

Its “AI Sales Agent,” quietly deployed as early as February 2025, suggests the company has been preparing this transition for months before public disclosure.

The Interface Becomes the Product

For two decades, SaaS competed for screen space inside browsers. Now the interface itself is changing. ChatGPT- with its SDKs and agent frameworks- may become the gateway through which most knowledge work begins.

In that world, competitive advantage shifts from owning the database to owning the dialogue. The decisive question isn’t which CRM users open first, but which assistant they talk to first.

Early, but Telling

This is still a proof-of-concept stage, not a new market equilibrium. Yet it points clearly to where momentum is moving: from individual SaaS products toward integrated, conversational systems that coordinate entire workflows.

The next frontier of enterprise software may be built inside the model, not around it.

Your B2B SaaS Funnel Is Obsolete. From Linear Stories to Compounding Loops

Why classic funnels break in B2B SaaS

The AIDA model—Awareness, Interest, Desire, Action—was built for straightforward, one-decision sales. SaaS doesn’t work that way.

“Buying SaaS isn’t a straight line—it’s a committee sport.”

The process is long, political, and cyclical: users feel pain, managers justify spend, executives sign. In this environment, AIDA’s neat progression collapses. The moment of purchase isn’t the end—it’s the midpoint. SaaS revenue is earned in usage, not checkout. The funnel must reflect that.

When AARRR isn’t enough

The Pirate Funnel (AARRR—Acquisition, Activation, Retention, Revenue, Referral) was meant to fix AIDA’s limitations by adding post-sale loops. It’s useful for metrics and growth experiments but still too linear for B2B SaaS.

“Acquisition” lumps every channel together. “Revenue” appears as a stage, when in practice it’s an outcome. And none of it accounts for multi-role buying committees or the long tail of customer success.

“No single funnel can explain SaaS. But some frameworks help you design it more intentionally.”

Enter the AEICIEA model

Impression Digital proposes a framework that better mirrors how SaaS buying really happens—Awareness, Education, Interest, Consideration, Intent, Evaluation, Action.

Each stage serves a distinct function:

Awareness surfaces the problem before the buyer names it. It’s educational, not promotional.
Education happens before any sales contact—buyers are learning the landscape and defining their terms.
Interest turns curiosity into structured search as teams start comparing approaches.
Consideration brings vendors into view: features, ROI evidence, customer stories.
Intent is when prospects give permission—requesting demos, signing up for trials, or sharing internal approval.
Evaluation digs into security, integration, pricing, and legal.
Action is the contract—but only the beginning of value realization.

“Every SaaS company loves to celebrate the sale, but that’s when your customer starts testing whether you can deliver what you promised.
”

Adding the missing stage: Adoption

The article’s most important addition comes after Action: Adoption. Turning customers into active users and internal advocates is the only true way SaaS compounds.

“In SaaS, the loop isn’t Awareness → Action. It’s Action → Advocacy.”

When Adoption feeds new Awareness and Education, your funnel becomes a flywheel. Customer success becomes marketing fuel.

Designing content that matches buyer intent

A major recommendation: stop mapping content to channels and start mapping it to intent. Instead of asking “What should we post on LinkedIn?”, ask “Which stage are we serving, and who’s the buyer at that moment?”

Awareness: tell stories that make the pain visible—industry data, relatable examples, or macro trends.
Education: teach frameworks and mental models—guides, webinars, or “how this works” explainers.
Interest & Consideration: offer comparisons—benchmark reports, case studies, or ROI breakdowns.
Evaluation: provide reassurance—security documentation, pricing transparency, integration walkthroughs.
Action & Adoption: design seamless onboarding, tutorials, community resources, and product updates that turn first users into advocates.

“Each piece of content should make someone inside your target company look smart in the next meeting.”

That’s the B2B equivalent of emotional appeal.

Measuring what matters

Metrics should follow intent, not hierarchy.

“If a blog post drives returning visits and longer reads, it’s working—even if no demo forms appear.”

Top-of-funnel assets succeed when they build recognition and revisit behavior. Mid-funnel pieces should be judged on engagement and sales-assist overlap, while late-stage content should accelerate deal velocity and improve close rates.

Alignment beats hand-offs

Ownership doesn’t belong to departments—it belongs to stages.

“A funnel isn’t a baton pass—it’s a relay where everyone runs overlapping legs.”

Product marketing informs Education. Customer success contributes to Evaluation. Marketing stays involved after the sale to capture stories of successful adoption. That overlap creates continuity—and credibility.

When to rebuild your funnel

All funnels expire. The signals are familiar: high traffic but few demos, poor MQL-to-SQL conversion, slow deal cycles, or inconsistent attribution.

“Most friction sits in stages your funnel doesn’t name.”

If your data looks broken, it may not be performance—it may be structure. The remedy starts with customer interviews, not landing page redesigns. Ask what content shaped their decisions and where they stalled. The answers often reveal blind spots between your existing stages.

From funnels to systems

“Funnels are useful fictions. What matters is the system you build around them.”

That system includes shared metrics, unified messaging, and a single source of customer insight across marketing, sales, and success. The strongest B2B SaaS teams treat the funnel as choreography—each stage feeding, learning, and improving the next.

The inspiration layer

The mindset shift is simple but profound: marketing isn’t persuasion; it’s acceleration. Every asset should shorten the distance between curiosity and conviction.

“Stop trying to move people through your funnel. Start designing experiences they want to move through.”

In SaaS, growth doesn’t come from pushing harder. It comes from aligning better—one stage, one story, one buyer at a time.

Your Brand Voice Is Probably Slop. Here’s the Fix.

Everyone wants a memorable brand voice. But most companies end up with generic “slop”—personality without a purpose, tone without a thesis.

The problem isn’t your style. It’s your lack of a real point of view.

Think your “human-centric” voice is cutting through the noise? It’s likely just more noise. Most brand voice guides are recipes for slop, focusing on quirky adjectives instead of substance.

Slop is brand voice without a point of view. It’s personality without a purpose. It’s the brand equivalent of a conversation with a person who won’t stop talking about themselves but never actually says anything.

Voice is the vehicle, not the cargo.

The fatal mistake is confusing how you speak with what you say. Your voice is the delivery mechanism. Your Point of View (POV) is the actual message.

Without a strong POV, your voice is an empty vessel. It’s forgettable because it has nothing to anchor to.

Your point of view is what you say. Your voice is how you say it. Most brands get this backwards. They start with voice and never find their point of view.

Here’s the hard part: A real POV is uncomfortable. It means having an opinion and standing against something.

The Anti-Slop Playbook

Ready to stop sounding like everyone else? The fix is a three-step process.

1. Find Your POV First. Before you write a single word, define what you believe. What’s your controversial take on your industry? What hill will you die on? This is your non-negotiable foundation.
2. Build Voice on Top of POV. Once your message is locked, the “how” becomes an amplifier. Your tone exists to serve your POV, not substitute for it.
3. Create a “Slop Filter.” Ruthlessly ban the fluff. Forbid empty adjectives (”passionate,” “innovative”), corporate jargon, and self-congratulatory intros. If it says nothing, cut it.

The goal isn’t to sound different. It’s to have something different to say.

Why Admitting Your Flaws Is a Marketing Superpower

The default marketing playbook is a lie. It tells you to project perfection, hide every weakness, and scream your strengths from the rooftops.

It’s a strategy built on insecurity. And your customers can smell it a mile away.

The most powerful way to build trust isn’t to pretend you’re flawless. It’s to admit you’re not.

This isn’t theory. Look at Volkswagen’s legendary 1960 “Lemon” ad. They found a tiny scratch on a new car and publicly branded it a “lemon,” unfit for sale. The ad openly discussed their obsessive quality control.

The result? It became one of the most famous campaigns in history and cemented the VW Beetle as a symbol of honesty and reliability.

This strategy has a name: the “blemishing effect.” It’s a psychological judo move that uses your audience’s skepticism against them.

When you lead with a minor negative admission, you non-consciously signal to people that you’re honest and trustworthy. As a result, they’re more likely to believe the positive things you say next.

By pointing out a small flaw yourself, you disarm the cynical part of the buyer’s brain. You’ve proven you’re not hiding anything. Now, when you talk about your strengths, they actually listen.

But there are rules.

Here’s the hard part. This isn’t a free pass for a bad product. Self-deprecation is a scalpel, not a wrecking ball. Get it wrong, and you just convince people you suck.

Use this playbook:

1. Your core offering must be strong. The VW Beetle was a fundamentally great car. Self-deprecation only works as a frame for genuine quality.
2. The flaw must be minor. Volkswagen highlighted a cosmetic scratch, not a faulty engine. Listerine admitted its taste was bad, not that it failed to kill germs.
3. The timing must be right. Reveal the flaw after the customer has already registered your key value propositions. Let the good news sink in first.

The Hans Brinker Budget Hotel in Amsterdam built its entire brand on this. They proudly call themselves “The Worst Hotel in the World.”

Their ads promise terrible service, no amenities, and general misery. It’s hilarious, honest, and it perfectly filters for their target audience: young travelers who just need a cheap bed. It sets expectations so low they can only be exceeded.

Stop chasing the myth of perfection. Your audience already knows it doesn’t exist.

Find your “lemon”–that small, forgivable flaw. Own it. Then pivot to what actually makes you great. It’s the fastest path to earning the one thing that matters: trust.

Stop Chasing Keywords. Start Building a Knowledge Graph.

The one-keyword, one-page SEO strategy is a relic. AI Overviews and chatbots don’t care about your carefully optimized H1s. They care about synthesis.

They build answers by pulling from multiple sources. The only question left is whether your content is structured to be one of them. If you’re still playing the old game, you’re already losing.

AI search engines are becoming answer engines, synthesizing information from various sources to provide a single, comprehensive answer.

This isn’t just another algorithm update. It’s a fundamental shift from information retrieval to knowledge synthesis. Your site is no longer a collection of documents. It’s a dataset waiting to be queried.

The old model is dead. You need to think like a database.

Forget about isolated articles. Start thinking in interconnected concepts. Your job is to create a logical, hierarchical structure that an AI can parse instantly.

Your website must become its own knowledge graph.

This means organizing content so an AI understands the relationships between ideas. It needs to see your domain as an authoritative, single source of truth.

The goal is to structure your content in a way that an AI can easily parse, understand the relationships between different pieces of information, and see your website as an authoritative source on the topic.

How do you do that? With a framework built for this new reality.

Introducing the TSEA Framework: Topic, Subtopic, Entity, Attribute.

This isn’t theory. It’s a content architecture. It organizes your expertise into a machine-readable hierarchy.

Topic: The universe. Your broadest subject area (e.g., “Cloud Computing”).
Subtopic: The galaxy. A major pillar within that universe (e.g., “Serverless Architecture”).
Entity: The star system. A specific, definable noun within the subtopic (e.g., “AWS Lambda”).
Attribute: The planet. A specific property, feature, or question about the entity (e.g., “AWS Lambda pricing models”).

This model forces you to map out your entire domain. It moves you from creating random posts to building a comprehensive knowledge base, piece by piece.

Your Action Plan: From Chaos to Clarity.

This framework requires discipline, not just more content. It’s about being strategic, comprehensive, and interconnected.

Map Your Universe. Define your core Topic. Break it down into Subtopics, Entities, and Attributes. This is your content blueprint.
Build Foundational Pillars. Create comprehensive, long-form guides for your main Topics and Subtopics. These are the hubs.
Create Atomic Content. Develop specific pages or sections for each Entity and Attribute. Answer one question, perfectly. These are the spokes.
Link with Intent. Use internal links to connect the hierarchy. Attributes link to Entities. Entities link to Subtopics. Subtopics link back to the main Topic. You are literally building the web for the web crawler.

By organizing your content this way, you’re essentially creating a knowledge base that AI can easily query. This structured approach... positions you as a comprehensive resource, making it more likely that you’ll be cited in AI-generated answers.

Here’s the brutal truth.

This is not a quick fix. It’s a fundamental re-architecture of your content strategy. It requires a full audit, ruthless pruning of thin content, and a long-term commitment.

Most won’t do the work. They’ll keep chasing keyword vanity metrics off a cliff.

The reward for doing it right? Becoming the definitive source AI turns to for answers. The alternative is becoming invisible.

AI Research Isn’t a Shortcut. It’s a Trap.

Everyone thinks AI will supercharge their research. The opposite is true. It’s a productivity sinkhole disguised as an efficiency tool, turning simple fact-finding into a frustrating maze.

The myth of the AI research assistant.

The sales pitch was simple: instant, accurate sources. But the reality is a mess of digital ghosts. AI confidently hallucinates academic papers, invents quotes, and serves up dead links. It’s not just unreliable; it’s actively misleading.

“I’d been using AI to help me track down sources for articles. It was supposed to be a fast track. But it turned into a funhouse of mirrors.”

Each citation is a new rabbit hole.

What does this mean for your workflow? Instead of writing, you’re forced to become a digital detective, hunting down phantom sources. Every AI-generated reference demands a manual audit.

This isn’t an upgrade. It’s a time-suck that kills momentum and introduces a dangerous layer of potential error. You’re not saving time; you’re just shifting it to tedious, low-value verification.

The hard truth: Trust nothing.

This isn’t a call to abandon AI. It’s a reality check on its current limitations. The machine is a powerful tool for brainstorming and summarizing known information. It is not a librarian.

“For now, the human fact-checker is not just valuable; they are indispensable.”

Treat every AI citation as a clue, not a fact. The final authority is still you.

B2B Marketing Round Up

Discussion about this post