Most people building AI right now have the same problem. They can build something cool. They just can't sell it.
We've watched over 20,000 people build AI agents and apps on Chipp. Some charge nothing. Some charge $5,000 a month. The difference isn't the technology. It's how they position it.
So we ran a workshop on exactly this. Here's everything we covered — and everything you need to know to actually get paid for AI work.
Here's the full replay video and slides here with highlights shared throughout. Subscribe on YouTube for future event replays!
The Big Shift Nobody's Talking About
Service is becoming a product. Products are becoming services.
If you're an agency or consultant selling AI services, you're probably already building repeatable deliverables that look a lot like products. And if you're building an AI product, you're probably getting pulled into consulting — onboarding calls, custom prompts, integration help.
The line between products and services is dissolving. And that's actually a good thing, because it means you can start in either camp and borrow from the other.
- Products give you scale and a larger market. But they're lower-priced and harder to customize.
- Services give you higher prices and deep customization. But they eat your time and are harder to scale.
AI blurs these together. A well-built agent is a product that delivers a service. That's the unlock.
The Three Eras of AI (And Why They Change How You Charge)
Ben Thompson recently shared three clear eras of AI:
Chatbots (2022): You ask, it answers. One pass. Done. Cost per interaction: fractions of a penny.
Reasoning (2024): You ask, it thinks. It drafts, reviews, considers tone, rewrites weak parts, then gives a final version. Cost: a few cents.
Agents (2025): You say “handle my customer support inbox.” It reads, reasons, drafts, checks policy, verifies facts, sends, logs, and repeats. Cost: a few dollars.
Each era is more powerful and uses more compute. The impact scales. So does cost. If you price it right, so does revenue.
You can imagine that previously a human was needed to talk to a chatbot and to invoke a reasoning model. Now, there is no bottleneck with humans. An agent can spin up unlimited numbers of reasoning chatbots - the scale grows exponentially. So does the cost and hopefully impact.
Stop Comparing AI to Software. Start Comparing It to Labor.
This is the single biggest framing shift when selling AI.
- Customer support rep: roughly $4,000/month fully loaded, maybe ~50 conversations/day
- AI agent: roughly $150–300/month in token usage for similar workload
- Typical savings: 90–95%
The usage bill isn’t the cost — it’s the savings receipt.
The Old Way vs. The New Way of Pricing
Old way: seat-based. Pay per user whether used or not.
New way: usage-based. Pay for work done. Slow month = low bill. High-output month = higher bill tied to value delivered.
Four Pricing Models and When to Use Each
In his book Scaling Innovation, Madhavan Ramanujam and Eddie Hartman lay out the different billing models in the Age of AI.
Seat-based: Best when value scales with human users. Risk: caps upside if AI replaces labor.
Usage-based: Best when usage tracks value/cost (support, voice, API). Risk: unpredictable bills can scare finance.
Outcome-based: Best alignment with ROI (per resolved ticket, qualified lead, successful workflow). Risk: attribution trust.
Hybrid (base + usage): Often best for AI. Predictability for buyers, upside for builders.
How to Price AI Services
Charge for value and what the customer will pay.
Here's a range of what we have seen Chipp users charging:
- Strategy: $5,000–$20,000
- Integration + Training: $2,000–$5,000
- Build (per agent): $1,000–$5,000+
- Maintenance: $500+/month
- Analytics: $500+/month
Stacking strategy + build + monthly maintenance commonly creates $10K+ engagements with recurring revenue.
How to Price AI Products
Chipp makes it easy to build and sell AI in whichever form makes sense for you: from a single app to an entire platform.
Here's a few ways you may sell AI products:
Single app: one use case, usage or flat plan.
App bundle: multiple coordinated agents, higher ACV.
Platform / white label: client self-serve under your brand.
Client billing with markup: clients pay usage, you keep margin.
The Pricing Conversation Cheat Sheet
“Seems expensive.” Compared to what? The human doing it costs much more.
“What if usage spikes?” Spikes usually mean more work completed and more value delivered.
“I don’t understand tokens.” Think electricity: pay for what you use.
“Why not flat monthly?” Flat fees disconnect price from delivered value.
How to Actually Get the Sale
- Ship early and charge on day one
- Use escalating pricing cohorts
- Offer early-access discounts for structured feedback
- Run POCs as business-case creation, not technical demos
- Start where no human currently handles the task (e.g., after-hours)
- Don’t spend cycles converting hard skeptics
The Model Pricing Trap
Model costs fall fast over time. Build with the best models now for customer experience, but anchor pricing on customer value — not your current token costs.
The Bottom Line
- AI is getting exponentially more capable.
- Usage-aligned pricing maps cost to value delivered.
- The adoption window is now.
You’re not paying for AI by the month. You’re paying for outcomes by the task.
Now go sell something. Chipp is here to help.
Build with Chipp
Whatever you're building, there's a path to get started with Chipp.
Chippster. Sign up for Chipp, build your first agent, and start selling.
Every Chipp user gets access to the platform, the models, and the ability to charge clients directly. This is where most people start — and where some people build entire businesses. Free to get started at build.chipp.ai.
Scout. You know an industry. You know the people, the pain points, the workflows. You just don't want to be the one building.
A Scout goes deep into a vertical — medical clinics, agencies, nonprofits, trades — and connects businesses with AI solutions. You bring the relationships. We help you deliver. Think of it as a channel partner for the AI era.
Architect. You're the builder. An Architect doesn't just know the industry — they build the solution.
Custom agents, white-label platforms, full AI stacks for a specific niche. Architects create the playbook that Scouts can sell into. You build once, deploy many times, and earn on every deployment.
If you're interested in becoming a Scout or Architect, reach out — this is brand new ground and we want your input on how to shape it.
Upcoming Learning Events
Join us for our upcoming cohorts: Technical Agents of Change and Business Agents of Change.
You'll learn to build agents that will impact your business and create products and services you can sell.
Join fellow Chippsters in-person for Chipp Con: Fargo. This two-day event is an opportunity to learn, share, and connect with fellow AI builders in your verticals.
Come join us in your favorite town you've never visited.
