Rethinking the customer journey in the age of AI written by John Jantsch read more at Duct Tape Marketing

Think about the last time you needed a service you’d never bought before. A new accountant. A contractor. Someone to fix your marketing.

Where did you start?

For more and more buyers, the honest answer is: they asked AI.

The research backs this up. 6sense found that 94% of buyers used a generative AI tool during their most recent purchase process. Adobe Digital Insights reports that AI referral traffic converts 42% better than traffic from other sources.

So right now, someone who needs exactly what you sell is asking AI who to hire. In seconds, they’ll get a short list of recommended names. You’ll never see it happen. And you may not be on the list.

The journey survived. The stages got rewired.

The path a buyer walks hasn’t changed. They find you, warm up to you, learn to trust you, take a small step, buy, come back, and send a friend. That part is human, and it isn’t going anywhere.

At Duct Tape Marketing we map that path with the Marketing Hourglass: Know, Like, Trust, Try, Buy, Repeat, Refer. A funnel pours people in the top and forgets them after the sale. The Hourglass widens again after the purchase, because repeat business and referrals are where the most profitable growth lives.

What AI changed is how buyers move through each stage. Walk through all seven with me and notice which ones feel thin in your business.

Know: how buyers find you

Buyers now run a full research phase inside AI before they ever reach your site. Someone asks “who helps a $5M HVAC company with marketing” and gets three names. “Best service near me” gets answered right in the AI summary while the map listings go unclicked.

And AI builds those answers from Reddit threads, directories, and reviews. Sources you don’t control. Forrester found twice as many buyers named AI their most useful research source than any other, ahead of websites, experts, and sales reps.

Whether you show up in the answer now matters more than where you rank.

What wins: pages that plainly state who you serve and what you do, structured so AI can read them. Presence on the sources AI pulls from. A mix of organic and paid.

On the paid side, we ran a small test of ChatGPT Ads for a client at $500 a month. Within a few weeks it landed an $18,000 client, larger than their average deal, with a shorter sales cycle. The buyer got educated inside AI before ever reaching out.

Like: how buyers warm up to you

AI made competent content free. So competent no longer stands out. Blog posts that could carry any company’s logo. LinkedIn posts written in the same AI cadence. Advice that repeats what every competitor is already saying.

The human voice is the one thing AI can’t copy. The founder on camera saying something a competitor wouldn’t. A story from inside your business that only you could tell. A public stance on where your industry gets it wrong. A voice people recognize in two lines.

The rule I gave the room: use AI to produce, never to think.

Trust: how buyers come to believe you

The web filled up with fakes, so buyers doubt everything by default. Reviews that all sound suspiciously alike. Testimonials with no name and no face. Stock-photo “teams” and AI-generated headshots.

When anything can be faked, real and verifiable wins. Your actual face and your actual team. Named clients a buyer can go check. A video testimonial from someone people can look up. Pricing and process shown in the open. Getting quoted or interviewed somewhere you don’t control.

Quick gut check: do you have a named, real proof asset you could point to right now?

Try: how buyers take the first step

Buyers want to try before they talk to anyone. They research and self-qualify long before they’ll book a call. Yet in most small businesses there’s no step between “interested” and “book a call.” A buyer who isn’t ready to talk has nowhere to go, so they leave.

That gap is a leak, and it’s the cheapest one to fix. AI made a real, low-risk first step cheap to build: a 2-minute diagnostic that hands back a personalized answer, a free teardown of their current setup, a calculator or scorecard they can run themselves, or a small paid engagement instead of a leap to a big commitment.

A buyer who takes your first step arrives more qualified and closes faster.

Buy: how buyers decide

Buyers show up already decided. They’ve read your pricing and your reviews. They’ve compared you to 2 or 3 competitors. They’ve fact-checked your claims against AI. The last question in their head is “why you, and can I trust you?”

Your job moved from convincing to confirming. Make the yes easy with clear pricing, a plain statement of who this is for, and one obvious next step. Answer the last doubt before it’s asked with a comparison, a guarantee, or a real result. And make sure what AI and reviews say about you backs up your own pitch.

Repeat: how customers come back

Anyone can generate Know, Like, and Trust content now. A delivered experience still has to be earned.

Here’s the good news for small teams. AI can personalize the relationship at a scale you never could by hand. Onboarding that adapts to exactly what the customer bought. Check-ins timed to how they actually use what they purchased, rather than a generic drip. Friendly reminders when it’s time to reorder or renew. Churn signals that catch an unhappy customer before they walk.

Refer: how customers send a friend

AI can prompt the ask, but the referral still runs on a person. Trigger the ask right after a win. Hand the customer a ready-to-forward intro. Give them something that makes them look good for sharing it. And sign the thank-you yourself.

One question worth sitting with: what happens in your business that’s worth talking about? Pick one moment.

Now score your own journey

Grade yourself 1 to 10 at each of the seven stages. Be honest.

When business owners do this exercise, a familiar picture shows up. A strong Know score sitting next to a Try score of 2. A journey that goes quiet after the first impression. A referral that never gets asked for.

Every one of those leaks traces back to the same place: no strategy holding the stages together.

AI is a multiplier. Point it at a real strategy and it pulls the right clients toward you faster. Point it at scattered tactics and it spends your money getting you lost quicker.

Strategy first, then AI

Strategy First is the layer that tells the AI where to aim. Who you serve. What makes you the obvious choice. The message and content that carry through all seven stages.

You have the diagnosis. Now get the plan. Book a Strategy First session and we’ll walk your journey together and find the strategy underneath it.

Did AI Make Referrals Your Most Important Marketing Channel written by John Jantsch read more at Duct Tape Marketing

Something odd is happening in marketing right now.

Output is way up. Results are down. Business owners tell me they’re publishing more content, sending more email, and running more outreach than ever, and the phone rings less than it did 2 years ago.

Here’s what I think is going on.

The great flattening

AI made competent marketing cheap. Any business can now produce a decent blog post, a polished cold email, and a professional-looking ad campaign in an afternoon.

So everyone did.

The result is a flood of marketing that all sounds the same. Buyers can’t tell who’s good anymore because everyone looks good on paper. And when everything looks professional, professional stops meaning much.

Two more shifts pile on. Ad costs keep climbing, and I don’t need to quote a stat you’ve already felt in your own budget. And buyers are doing more of their research inside AI assistants. They ask ChatGPT or Claude who to hire, get a summarized answer, and may never see your website at all.

Add it up and the channels most businesses depend on for new customers (content, ads, outbound) are getting louder, pricier, and less believable all at once.

The half of the journey nobody can automate

I’ve taught the Marketing Hourglass for 2 decades: Know, Like, Trust, Try, Buy, Repeat, Refer. Unlike a funnel, it widens back out after the sale, because the customers you keep and the fans you create are where the best growth comes from.

AI just made that second half a lot more valuable.

Anyone can generate Know, Like, and Trust content now. A delivered experience still has to be earned. So does a customer telling a friend, “hire these people, they took care of us.”

The bottom half of the Hourglass gets built one customer at a time, and that’s exactly why it’s about to become the highest-return part of your marketing.

I wrote this book 16 years ago

In 2010 I published The Referral Engine. The argument was simple: referrals are too important to leave to chance, so build a system that produces them.

Here’s a fact that surprises people. I’m known for Duct Tape Marketing, but The Referral Engine is my best-selling book. It still sells today, 16 years on. Owners have always understood at a gut level how much referrals matter, and the book keeps finding readers because the problem never went away.

What most readers didn’t have back then was urgency. Referrals stayed the nice-to-have channel, appreciated but never built for, because ads and SEO and email were working fine.

They’re working less fine now. And the ideas in that book turned out to be built for this exact moment. Here are the ones that matter most.

People are wired to refer

Making a good referral feels good. It’s social currency: introducing a friend to a great business makes you look smart and useful. That wiring didn’t change when the tools did.

Which means your customers want to refer you. Most just don’t, because nothing in your business makes it easy, obvious, or timely.

That’s a system problem. System problems have fixes.

A referral system beats a referral hope

Most owners treat referrals like weather. They happen or they don’t.

But a referral system has the same parts as any other marketing system: a clear picture of who you want referred, language your sources can borrow, an offer that makes the referrer look generous, a trigger for when to ask, and a way to track what came in.

Here’s the part that’s new. Follow-through used to kill referral programs. Somebody had to remember to ask, personalize the thank-you, keep partners warm. AI handles that kind of follow-through well now. The strategy still has to come first, but the old excuse about not having time is gone.

Referability comes before referrals

No tactic fixes an unremarkable business. Before any referral campaign, answer one question: what happens in your business that’s worth talking about?

Close the gap between what you promise and what customers experience. Then engineer one moment worth retelling. The surprise check-in. The problem you fixed before the client noticed it. The onboarding that felt like a welcome instead of paperwork.

Referrals are the byproduct of experience design.

Your best referral source may be another business

Direct referrals come from customers. Indirect referrals come from strategic partners: businesses that serve your exact customer right before or right after you do.

For most B2B companies, the partner engine is the bigger one, and the less built one.

Try this. List the 5 businesses your best customer hired in the 12 months before they hired you. That’s your partner shortlist. Teach those firms exactly who to send you, give them something worth giving away, and send business back on purpose.

I’ll give you a live example from our own shop.

My friend Mike Michalowicz runs a group of business advisors called Prosper. These are the people founders already trust with profit, operations, and the big picture. When one of their clients needs marketing leadership to hit a growth target, guess who gets asked for a name.

We’re working with 3 clients right now who came to us through that group. The engagements are going well, and going well is the whole engine: results for a referred client are what produce the next referral. Just recently, another advisor sent us a new client the same way.

The part worth studying is the order of events. Those first referrals showed up on their own. The system came second. Now we’re debriefing the advisors who send us business to learn what prompted each referral, defining exactly which founders we want sent our way, and building tools an advisor can hand a client the moment marketing comes up in a planning session.

Your business almost certainly has a pattern like this running right now. The work is spotting it, then formalizing it.

And a quick word if you’re a business advisor yourself. When a client needs marketing leadership and you want to hand them a name you can vouch for, that’s the exact gap our advisor partnership fills. See how the advisor partnership works.

The new part: AI is listening to your customers

Here’s what makes a 2010 book feel current in 2026.

When someone asks an AI assistant who to hire, the answer gets assembled from reviews, testimonials, case stories, and mentions scattered across the web. Your customers’ words are the raw material AI uses to describe and recommend you.

Word of mouth now travels machine to machine.

That turns review generation and testimonial capture into discovery infrastructure. Every satisfied customer is writing part of your AI sales pitch.

Try it yourself. Ask 3 different AI assistants who they’d recommend in your category and your market. Compare what they say to what your best customers say about you. The gap between those two answers is your work.

Where to start

Four moves for this quarter:

1. Track where your last 10 customers actually came from. Most owners are surprised by how much of their growth is already referral-driven.
2. Pick one moment in your customer experience and make it worth talking about.
3. Build your partner shortlist using the exercise above and set up 2 conversations.
4. Ask for reviews at high points (right after a win, a renewal, a thank-you email) and make it a 30-second task for the customer.

Do it in that order. A referral push with no strategy underneath it is just another random act of marketing.

The businesses that win the next few years will treat the second half of the Hourglass as their primary growth channel and build a system to run it. The good news is that your competitors are all busy generating more content.

A “Moore-ish” law for marketing written by John Jantsch read more at Duct Tape Marketing

Marketing is quietly crossing a threshold.

Not because we can “make more content” with AI.

Because every time the cost of thinking drops, the amount of experimentation and personalization you can afford explodes. And that changes the shape of marketing from campaigns you launch to systems you operate.

I’ve spent decades teaching small businesses to stop chasing shiny tactics and install a marketing system that creates clarity and consistent growth. That “system-first” mindset is exactly what this moment demands, except now the system can learn faster than your team can type.

The “Moore-ish” Law of Marketing: When the Cost of Thinking Drops, Experimentation Explodes

Table of contents

1. A “Moore-ish” law for marketing
2. Three forces changing marketing right now
3. What this means in the near term
   1. Content shifts from assets to streams
   2. Personalization shifts from segments to situations
   3. Agents start eating coordination work
4. The role shift: from makers to operators of systems
5. What to do right now: a near-term playbook
6. How this fits the Duct Tape Marketing system
7. FAQs

A “Moore-ish” law for marketing

Here’s the framing:

When the cost per marketing experiment keeps falling, the number of experiments you can afford rises dramatically.

If it gets 2x cheaper, or 2x faster, to generate, quality-check, and deploy a new variation, you do more of them. Over time, that pushes marketing from campaign-centric (big launches, big bets) to system-centric (continuous learning, continuous improvement).

The unlock is not “AI content.” It’s collapsed time-to-learning.

When time-to-learning collapses, the limiting factor shifts:

• Not “can we produce it?”
• But “can we measure what’s true and decide well?”

That is the big idea in one sentence:

Execution moves from human throughput to machine throughput, while humans move up the stack to judgment, strategy, constraints, narrative, offers, positioning, and ethics.

Three forces changing marketing right now

1) Cost per experiment is falling

The ability to create variations is no longer scarce. What is scarce is the ability to run clean tests, protect the brand, and decide.

2) Time-to-learning is collapsing

Shorter loops mean you can improve messaging, creative, and offers continuously instead of waiting for a quarterly campaign post-mortem.

3) Coordination work is becoming automatable

Most marketing teams spend more time coordinating work than creating leverage. As tools integrate into work apps, AI can draft, route, repurpose, tag, schedule, and execute multi-step workflows under human supervision.

What this means in the near term

1) Content shifts from assets to streams

The near-term change is simple to say and hard to implement:

Your team stops “shipping one landing page” or “one email sequence” or “one ad set.”

You ship:

1. A message system
   Positioning, proof, objections, tone rules, prohibited claims, and brand constraints.
2. A modular content library
   Claims, examples, stories, CTAs, offers, proof points, and objection-handling modules.
3. A generation and QA pipeline
   A workflow that continuously produces variants, checks them, deploys them, measures performance, and feeds learnings back.

Role impact

• Copywriter becomes editor-in-chief plus conversion strategist
  Owns voice, truth, persuasion, compliance, and performance feedback.
• Designer becomes system designer
  Builds templates, components, motion rules, and brand constraints.
• Content lead becomes content operations lead
  Owns workflow, governance, QA, and measurement loops.

2) Personalization shifts from segments to situations

Segmentation is still useful, but the economics are changing.

When personalization gets cheaper, you stop asking only:

• “Which segment is this?”

And start asking:

• “What situation are they in right now?”
• “What job are they hiring us for?”
• “What objection is active?”
• “What constraint is binding: budget, time, risk, internal approval?”
• “What is the next best step that fits their reality?”

Role impact

• Campaign manager becomes journey architect
  Owns triggers, decisioning, orchestration, and next-best-action paths.
• Marketing ops becomes decision ops
  Owns data quality, identity, measurement, guardrails, and evaluation standards.

If you want one practical takeaway: the “unit of personalization” is shifting from a persona to a moment.

3) Agents start eating coordination work

Most marketing teams spend more time coordinating work than creating leverage:

• Creating briefs
• Routing approvals
• Repurposing content
• Tagging and organizing assets
• Scheduling and posting
• Producing “version 14” of a variation
• Summarizing results and sharing updates

As AI integrates with workplace tools, these coordination tasks can be automated or semi-automated with human checkpoints.

Role impact

A new role emerges, especially in teams that want scale without chaos:

Marketing agent wrangler
The person who builds repeatable agent workflows, monitors outputs, tunes prompts, sets permissions, and makes sure “automated” never means “unaccountable.”

The role shift: from makers to operators of systems

If change keeps accelerating, the safest career position is not “the fastest maker.”

It is:

The person who can design the system that produces outcomes repeatedly.

Here’s a simple mapping.

Roles that shrink (execution throughput)

• “Write 20 posts”
• “Make 30 ad variations”
• “Draft 10 nurture emails”
• “Create first-pass briefs”

These become machine-default, especially for first drafts and variant generation.

Roles that grow (judgment, leverage, trust)

• Positioning and offer design
  What to say, to whom, and why it’s true.
• Creative direction
  Taste, narrative, cohesion across channels.
• Performance strategy
  What to test, what to stop, what to double down on.
• Marketing operations and governance
  Permissions, QA, brand safety, compliance, evaluation.
• Customer research synthesis
  Turning messy reality into usable direction.

What to do right now: a near-term playbook

If you want a practical playbook that fits this Moore-ish acceleration, focus on four builds.

1) Build a truth layer

Your team needs a single source of truth that answers:

• What claims can we make?
• What proof supports each claim?
• What is disallowed legally, ethically, or brand-wise?
• What language do we never use?
• What industries, customer types, or outcomes require extra care?

This is how you prevent fast nonsense.

AI without a truth layer produces confident randomness. AI with a truth layer produces scalable clarity.

2) Standardize a production pipeline

A healthy pipeline looks like:

Brief → generate → fact-check → brand-check → legal-check (if needed) → deploy → measure → feed learnings back

Notice what’s missing: polish endlessly.

If the system is meant to stream variants, your job is not perfection. Your job is controlled learning.

3) Create an evaluation habit

The question is not “did AI write it?”

The question is:

Did it move the KPI while protecting the brand?

This is where many teams will break. If you cannot evaluate, you cannot scale.

At minimum, define:

• Your primary KPI by channel
• Your guardrail metrics (complaints, unsubscribes, refund rate, brand sentiment indicators)
• Your decision cadence (daily for ads, weekly for emails, monthly for site and SEO)
• Your stopping rules (when to kill a test quickly)
• Your doubling rules (when to scale a winner)

4) Reskill around leverage

Train marketers to do the work that scales:

• Design experiments
• Write constraints
• Critique outputs
• Interpret results
• Orchestrate tools and workflows
• Document learnings so the system improves over time

Many teams will run more tests, but fail to compound the learning. The habit of documenting what worked and why becomes a strategic advantage.

How this fits the Duct Tape Marketing system

This moment does not replace strategy. It punishes teams who try to skip it.

Duct Tape Marketing has always been rooted in the idea that marketing is a system, not a pile of tactics, and that clarity beats chaos.

AI acceleration rewards that approach because:

• A system gives you the message constraints that prevent garbage at scale.
• A system gives you the customer journey structure that makes personalization meaningful.
• A system gives you a measurement discipline so “more output” becomes “more learning,” not “more noise.”

Or said another way:

AI makes tactics cheaper. It also makes strategy more valuable.

If you want to future-proof, build the machine, but lead it with principles:

• Strategy before tactics
• Truth over hype
• Consistency over novelty
• Learning over launching

FAQs

1) Is this just about creating more content faster?

No. The advantage is not volume. The advantage is iteration, testing loops, and faster time-to-learning. Volume without evaluation just creates more waste.

2) What is the biggest risk as experimentation gets cheaper?

Scaling bad assumptions. If your truth layer is weak, you will publish confident errors, drift off-brand, and damage trust faster than ever.

3) What should small businesses do if they do not have a data science team?

Start simpler. Use AI to increase iteration on high-leverage assets where measurement is clear, like ads, landing pages, and email subject lines. Keep the loops tight and focus on one KPI per test.

4) How do we prevent brand inconsistency when AI is generating variants?

Operationalize brand constraints, not just guidelines. Build templates, component rules, disallowed language lists, and a review checklist that enforces your standards.

5) Do we need “agents” right now?

Not to start. Begin with a standardized pipeline and a truth layer. Agents become valuable when you have repeatable workflows worth automating, and clear checkpoints for approvals and measurement.

6) Which roles should we hire or promote for this shift?

Look for people who can design systems, run experiments, and make decisions with incomplete information. “Taste plus rigor” becomes a premium combination.

7) How does personalization change first for most teams?

You move from broad segments to situational messaging on the same core journey. Think objection-based variants, industry-context variants, and stage-of-awareness variants, all measured and refined continuously.

8) How do we know we are using AI in a way that drives growth, not just efficiency?

If your AI program only measures time saved, you are still in productivity mode. The shift is tying AI-enabled workflows to business outcomes, with clear accountability for impact.

Next step: If you share your context (agency serving SMBs, in-house B2B, local service business, SaaS, ecommerce), I’ll translate this into the three highest-leverage workflows to automate first, the roles to redesign, and the metrics that keep the machine honest.

Are Algorithms Making Us Stupid? written by John Jantsch read more at Duct Tape Marketing

Spotify Wrapped is a brilliant marketing play. Every year, millions of people gleefully share their top songs, favorite artists, and most-listened-to genres, essentially turning their personal data into free advertising for the streaming giant. But while it’s fun and feels personalized, it also sheds light on something deeper—and a little unsettling—about the world we live in today.

This year, my #1 song was “Boulder to Birmingham” by Emmylou Harris. It’s a beautiful, haunting tribute to Gram Parsons, her mentor, who died of a drug overdose. My wife and I sing along every time it plays. The thing is, I never asked Spotify to play that song. Not once. And yet, it kept showing up in my mix, again and again. Along with it were several John Prine tracks I didn’t seek out either. In fact, I didn’t actively choose any of the top songs Spotify says I loved this year.

This might sound like a minor quirk in an otherwise delightful digital experience, but it’s actually symbolic of a much larger issue. Increasingly, the world around us is being curated not by us, but for us, by algorithms that interpret our past behavior and then decide what we should see, hear, and engage with next.

Sure, this applies to music. But it also affects our news feeds, our product recommendations, our search results, our social media content, the ads we’re exposed to, and even the lies we’re told. The more we click, the more the algorithm “learns” about us. And the more it learns, the narrower our world becomes.

This is the filter bubble in action. Over time, our exposure to new ideas, unfamiliar perspectives, or even just different kinds of content diminishes. We’re not discovering anymore; we’re being fed what the machine thinks we already like, or worse, what will keep us clicking.

On the surface, this kind of personalization feels convenient. But in practice, it can be dangerously limiting. It traps us in echo chambers, reinforces existing biases, and makes it harder to challenge our assumptions or grow intellectually and emotionally. When the only ideas we hear are the ones we already agree with, how do we grow?

This isn’t just a tech problem. It’s a human one. And it has consequences that reach far beyond our Spotify playlists. It’s affecting how we think, how we relate to others, and how we understand the world. It’s part of what’s driving polarization, misinformation, and a culture of outrage. We’re not just being shaped by what we consume; we’re being shaped by what we’re allowed to consume.

It doesn’t have to be this way.

The good news is that we still have agency. We can choose to seek out opposing views. We can read books from outside our usual genres. (I’m not an architect, but Christopher Alexander is one of my favorite authors). We can listen to podcasts that challenge our thinking. We can actively resist the passivity that algorithms encourage. But we must do it intentionally.

Because left to their own devices, the algorithms will not feed us what we need—they will feed us what keeps us comfortable, entertained, and clicking.

I didn’t always see it this way. For a long time, I appreciated the convenience of curated content. But now I’m convinced: algorithms, as they are currently used, are making us intellectually lazy. They’re dulling our curiosity. They’re making us stupid—not in terms of raw intelligence, but in terms of awareness, perspective, and growth.

So here’s my challenge to myself—and maybe to you too: go out and discover something. Don’t wait for it to be served up by a machine. Curate your own experience. Choose what you want to see, hear, and learn next.

Because if you don’t, someone, or something else, will do it for you. And you might just find yourself singing along to a song you never chose in the first place.