Kipp Bodnar and Kieran Flanagan's new book, Loop: Outlearn. Outmarket. Outgrow., makes a compelling argument: the marketing teams that win will be the ones that learn fastest.
The book was introduced at UNBOUND 2026 and builds on the Loop Marketing framework HubSpot introduced at INBOUND 2025. Instead of thinking about marketing as a one-way funnel, Loop Marketing is built around four stages, Express, Tailor, Amplify and Evolve, that continually feed what you learn back into the next cycle.
I think they are right.
But applying Loop Marketing to B2B raises three practical questions:
What should we learn from?
How should we use what we learn to personalize the buyer experience?
And how do we feed that learning back into the next interaction?
Buyer questions help answer all three.
Most marketing systems are very good at capturing behavior.
We know which ad someone clicked, which pages they visited, whether they downloaded an asset and whether they opened an email.
What we usually don't know is what they wanted to know.
That distinction matters.
A prospect visiting several product pages is useful information. A prospect asking, "We're already spending heavily to drive traffic, but too few visitors convert. How would this help us get more pipeline without increasing media spend?" tells you much more about the buyer's pain, economics and success criteria.
We recently graded the websites of 130 B2B SaaS companies against the questions serious buyers typically ask during an evaluation. Of the 108 sites we could fully grade, the typical site answered only about half of those questions. Just 4% of answers explained how the product works, rather than simply describing the outcome. Only 6% made a meaningful case against a named competitor.
Chatbots didn't solve the problem either. Of 102 websites we reviewed, 41 had a chat widget. Only seven actually answered a buyer's question.
That creates an interesting problem for Loop Marketing.
If the goal is to learn faster, B2B marketers need better inputs into the Loop.
And one of the best inputs we have is hiding in plain sight: the questions buyers ask while they are deciding whether to buy.
Complex B2B purchases are rarely made by one person.
A champion may be evaluating the product. Security may be evaluating risk. Finance may be looking at economics. An executive may be deciding whether the investment is worth making.
Each asks different questions.
A champion might ask:
How would I use this every day?
Security might ask:
Where is our data stored?
Finance might ask:
What will this cost as we scale?
An executive might ask:
Why should we buy this instead of doing nothing?
Very few of those questions make their way back to marketing.
Your marketing system may record a pricing-page visit or a security-page view.
But the actual question disappears.
Clicks tell you what buyers looked at. Questions tell you what they were trying to understand.
That is a much stronger signal.
And it leads directly to the second question.
The Tailor stage is where I think AI changes the model most dramatically.
Most website personalization isn't particularly personal.
A visitor might see their company name in a headline, a different customer logo or a piece of content selected for their industry.
That is better than showing everyone the same thing, but the underlying experience is still largely static.
AI allows marketers to go much further.
A CFO, security reviewer and end user don't simply need different content. They ask different questions, and the right answer can depend on who they are, the company they work for, what they have already looked at and where they are in the buying process.
Consider a question like:
"What kind of ROI could we expect?"
A static page can publish an average ROI number or a customer case study.
But it cannot give a particularly useful answer to your company unless it knows something about your organization, use case and economics.
The same is true for questions like:
"How difficult would this be for us to implement?"
"Will this work with our existing technology stack?"
There is no single static page that can provide the best answer to every company and every buyer.
Conversational AI changes that model.
Instead of building a different webpage for every industry, company size, role, use case and buying stage, marketers can personalize every interaction with every contact and every account.
The answer itself can change.
With SlateCX, for example, a buyer can ask a question in their own words. The agent can combine that question with what is known about the contact and account, prior engagement and the context of the conversation to provide a more relevant answer.
That is fundamentally different from conventional website personalization.
Traditional personalization changes the page based on what you know about the visitor. Conversational personalization can change the answer based on what you know about the buyer and what they are asking right now.
That is a more useful definition of Tailor for B2B.
This is where buyer questions become more than another source of intent data.
They become the feedback mechanism that makes the Loop actually loop.
The process is straightforward:
Ask → capture → learn → improve → answer better next time.
And the learning can feed every stage of Loop Marketing.
Suppose thirty prospects ask:
How are you different from Competitor X?
That is a signal.
You may not have just a sales problem. You may have a positioning problem.
That comparison belongs in your messaging, website or sales materials.
Recurring questions reveal where your positioning is unclear and what buyers need you to explain better.
If you know which questions come from a CFO versus an IT buyer versus a champion, you start building a much better picture of what each person cares about.
And because that interaction becomes part of the contact and account context, the next answer doesn't have to start from zero.
The system learns not only what buyers ask, but who asks what.
That makes the next interaction more relevant.
Buyer questions also give you a practical answer to one of the hardest questions in content marketing:
What should we publish next?
Every important question you can't answer is a piece of content you probably need to create.
If buyers repeatedly ask about implementation, publish a detailed implementation guide.
If they ask about a competitor, build the comparison.
If security questions keep appearing, improve your security documentation.
The unanswered questions become a content calendar written by your buyers.
This matters even more as buyers increasingly use ChatGPT, Claude, Perplexity and other AI services to research products. Those systems can't provide specific answers if the underlying information has never been published.
For many B2B marketers, the question behind Evolve is simple:
What should we fix next?
Imagine seeing:
That is already a prioritized improvement list.
And because those questions can be connected to contacts and accounts, marketers can go further.
Maybe SSO questions primarily come from enterprise accounts. Maybe implementation questions spike at a particular deal stage. Maybe security questions consistently come from one member of the buying committee.
Now buyer questions aren't merely content ideas.
They become data you can use to improve messaging, personalization, sales enablement and product strategy.
That is the feedback loop: real questions change what you say, what you publish and how you answer the next buyer.
You don't need new technology to start.
Begin with information you already have: sales-call transcripts, security questionnaires, onboarding conversations and win/loss interviews.
But to make the process continuous, you need three things:
That is one of the problems we built SlateCX to solve.
A Slate gives buyers a persistent workspace where they can ask questions, get answers based on approved company content, collect useful materials and return later.
Once a buyer saves a Slate, their questions and engagement can be connected to their HubSpot contact record. If the AI agent can't answer something, the question can be routed to someone on the team.
The goal isn't simply to answer more questions.
It is to create a self-improving system:
Buyer asks → answer is personalized → signal is captured → gap or pattern is identified → marketing improves → next buyer gets a better experience.
Then the cycle starts again.
One of the ideas in Loop I particularly like is Buyer Sim: using AI as a stand-in buyer to pressure-test messaging before launch.
I'd use it.
But simulated buyers and real buyers aren't the same thing.
A simulated buyer can identify obvious weaknesses in positioning or content before you launch.
A real procurement team may ask whether you support the identity provider used by its German subsidiary or what happens to its data after the contract ends.
Those are exactly the questions you may not predict.
So, use Buyer Sim before launch. In fact, we developed a set of Buyer Sim questions based on frameworks like Gartner's six B2B buying jobs and used them to evaluate how well many B2B websites answer them.
Capture real buyer questions after launch.
Then feed those questions back into Buyer Sim so the next round of simulation starts from real buyer evidence.
That is another way the Loop improves itself.
You don't have to wait until you've collected months of buyer conversations to start.
Start by seeing how well your website already answers the questions serious B2B buyers ask.
We built a free AI Visibility Tool that tests your site against buyer questions based on established B2B buying frameworks. It looks at whether the answer is available on your website, not simply whether the right keywords appear, and tests what AI can tell a buyer based on the information you've published.
You can run it free at slatecx.com/ai-visibility-tool.
The result gives you a starting point: Which important buyer questions can you answer today, and where are the gaps?
The grader gives you a baseline. Your own buyers make it specific to your business.
Look at recent sales calls, security questionnaires, implementation conversations, win/loss interviews and customer emails. Capture the questions in the buyer's own words.
Then compare them with your grader results.
You'll probably find two types of gaps:
Content gaps: Questions you should be able to answer publicly, but your website doesn't answer well today.
Personalization gaps: Questions where there isn't one universal answer because the response depends on the buyer, account, role or situation.
That distinction matters. The first requires better content. The second requires a better way to use buyer context when answering.
Don't just collect questions. Decide what each one should change.
Does it expose:
Then make the improvement and measure what happens next.
The objective is not simply to answer more questions. It is to create a continuous learning cycle:
Buyer asks → answer → capture the signal → identify the gap → improve → answer better next time.
That's when buyer questions stop being isolated conversations and start becoming a learning system.
Kipp and Kieran's central argument is that the fastest learners win.
I think that is particularly true in B2B.
But learning faster depends on having the right signal.
Your buyers are already telling you what they care about, what they don't understand and what may be preventing them from moving forward.
They're asking salespeople.
They're asking colleagues.
They're asking security teams.
And increasingly, they're asking AI.
The opportunity is to capture those questions, understand who is asking them, provide a better answer and feed what you learn back into the next interaction.
Buyer questions are the signal. Personalized answers are the response. Continuous improvement is the Loop.
If the fastest learners win, the first step is making sure you can hear what buyers are asking.