Faster Horses vs. Teleporters

Why you need both for successful AI adoption and how to manage them

Last week, I had 3 AI strategy calls with execs from different industries (media, finance, and manufacturing). While each went quite different, one thing commonly stood out:

By now, everyone knows AI adoption comes in two flavors.

First, there's the "AI for personal productivity" category. It's what I call giving people faster horses. Everyone is still doing the same work, just faster, cheaper, maybe better.

Second, there's the "System performance" level. It's not necessarily improving existing workflows, but redefining how work is done. Some work may even become obsolete altogether. I call this giving people Teleporters instead of Faster Horses.

The key to get really good (and measurable) results from AI lies in doing both – rolling out Faster Horses and Teleporters where it makes sense.

Today, I'll share more examples of what these are, plus some tactics I've learned for managing both.

Let's dive in!

The Problem

Walk into any company rolling out AI these days and you'll see two camps:

Camp 1: "AI isn't transformational enough. We're just making people work a little faster (at best)."

Camp 2: "We're flooded with AI use cases, but can’t handle them all. Every department wants something different."

None of these camps sees results that make executives happy. Because both are mixing up Horses and Teleporters.

Camp 1 is frustrated because they're trying to achieve Teleporter success with Faster Horses. They want transformation but they're only building productivity tools. Camp 2 is overwhelmed because they can't tell which use cases are Faster Horses (productivity quick wins) versus Teleporters (system redesigns that need more effort). Everything looks equally important.

The solution is treating both with the right approaches, timelines, and success metrics.

Faster Horses

Faster horses are the obvious AI wins. ChatGPT for email drafts. Copilot for code suggestions. AI note-taking in meetings. Document summarization.

But they don't necessarily drive ROI. We already knew before ChatGPT that real business productivity is less about your people, and more about your systems.

Examples:

  • Your sales team writes faster emails, but still sends the same number.

  • Your developers code better, but follow the same dev cycle.

  • Your analysts create reports cheaper, but don't deliver more insights.

In all of these cases, adopting AI for "personal productivity" actually leads to decreased business performance – because you still have the same output, plus AI cost on top.

Don't get me wrong – I'm not arguing against personal productivity use cases. In fact, I think every business should do this. Right. NOW! Giving people easy access to general-purpose AI tools is mandatory these days (if you won't, they'll do it themselves) – and it unlocks bigger opportunities (I'll get to those in a second).

But don't deploy productivity use cases expecting huge P&L impact. You probably won't get it.

Most likely, everyone's "more productive" but the business metrics look exactly the same.

That's the natural limit of faster horses.

Teleporters

Teleporters don't necessarily enhance workflows. They redesign them. Instead of making A → B → C → D faster, they take you from A → D, maybe introducing some post-hoc checkpoints at C.

While these sound tempting, these cases don't just happen. Leaders have to make them happen, especially in larger organizations. In contrast to Faster Horses, Teleporter cases aren't discovered bottom-up, but designed top-down. Besides centrally supported technical enablement, this also requires top-management support to predict and prescribe how work should be done.

What it actually takes:

  • Executive commitment to process redesign, not just tool adoption

  • Cross-functional collaboration to map and rebuild workflows

  • Change management to get people comfortable with fundamentally different work patterns

  • Risk tolerance for experiments that might not work the first time

Most companies underestimate this. They think teleporters are just "better AI tools" when they're actually organizational design projects that happen to use AI.

Example:

Take the role of a Product Manager. In the old days, their job meant writing down detailed product specifications, translating those into user stories, spending hours in meetings and updating Jira boards. Using AI tools they can now write these specs faster, get automated transcriptions of their meetings or have AI pre-fill the Jira board. These are Faster Horses.

Vibe coding a prototype is a Teleporter. Instead of doing all of the above, the PM uses AI to get a simple demo of the envisioned feature up and running so they can show the desired experience to the developers (who then build it out. With the help of AI, of course). Google added vibe coding to PM interviews. Shopify is doing the same.

And it makes sense. Because that's not just faster PM work. That's different PM work.

Same outcome. Different system. Different performance.

Some more examples:

Customer support Horses: Support agents use AI to draft responses faster. Knowledge base search gets smarter.
Teleporter: Redesign around AI-first triage. Instead of ticket → human routing → research → response, you get: issue → AI classification → auto-resolution or expert escalation. Most issues never need human routing.

Finance: Horses: Accountants get AI assistance with reconciliations - AI helps match transactions faster, suggests likely matches. Teleporter: Build continuous reconciliation. Instead of month-end → gather data → manual matching → exception handling → sign-off, you get: real-time matching → instant alerts → continuous validation. Month-end becomes confirmation, not exploration.

How to get there

So can we now skip the Faster Horses and just shortcut to the Teleporters? No. Because like in the non-AI world, if Teleporters were so easily accessible, everyone would use them. But often, they just don't exist (yet). Which means either the tech, your organization, or your data isn't ready for them.

How do you know if a Teleporter exists? There are two ways, depending on your AI adoption approach.

Option 1: You spend years (and a whole lot of R&D budget) to find it out.

Option 2: You leverage existing AI initiatives (horses) across multiple departments to collect evidence. This works especially well for the Divide and Conquer approach.

In all of the companies I've consulted for, Faster Horses were necessary to discover Teleporter opportunities, and ultimately build literacy and quick wins.

  • Your sales team gets comfortable with AI email tools, so they're ready when you redesign the entire lead qualification process.

  • Your developers build confidence with AI coding assistance, so they can help architect the new PM prototyping workflow.

  • Your accountants know the pitfalls of AI, so they're aware of the gotcha moments for the monthly check-ins.

Making sure that these don't exist in isolation, but actually unlock Teleporters over time is the goal of your AI roadmap.

How to manage each

Horses are typically developed bottom-up. That means broad enablement, workshops, courses and knowledge sharing in tools like ChatGPT, Copilot & co. The rest is "just" culture and making sure everyone understands the "reason why".

Teleporter opportunities are identified top-down. They're typically hard to spot from the day-to-day hustle because they require the "bigger picture". The tool of choice for me is an AI Opportunity Scan which I run in multiple short workshop formats where we look strategically for major problems and bottlenecks that stand in the way of Teleporter-like process improvements (and how AI can potentially help).

Practical tips:

Rolling out the combo requires different thinking for Horses vs. Teleporters.

  • For Horses: Focus on adoption and immediate value. Measure time saved, quality improved. Celebrate quick wins. (Even if they don’t affect the P&L)

  • For Teleporters: Focus on redesign + system integration. Measure throughput, cycle time, elimination of steps. Expect longer implementation. Identify enablers. Celebrate milestones.

In terms of AI solution design, Horses typically match the left side of my Integration-Automation Framework (Assistants and Copilots), while Teleporters are delivered as Autopilot and Agentic solutions. In other words: Don't rely on MS Copilot Chat to transform your organization. And likewise don't build custom n8n agents just for personal productivity gains.

Conclusion

Henry Ford once (allegedly) said: "If I asked people what they want, they'd say a faster horse". And there's nothing wrong with a Faster Horse in the short term.

But Henry Ford couldn't have dreamed that besides a car, people could even use a Teleporter instead.

Your product team needs AI prototyping tools (horses) AND the entire feature development cycle collapsed (teleporter).

Action item for today: Audit your current AI initiatives. Which are horses? Which are teleporters? How are they connected?

Start with horses. Discover teleporters. Build both.

See you next Friday,
Tobias

P.S. If you want to see Horses → Teleporters in action, join my upcoming AI inside spreadsheets workshop next Thursday. You'll learn how AI can help you turn faster data cleaning (horse) into the foundation for analysis workflows that collapse entire research processes (teleporters). Details here →

Reply

or to participate.