🎯 Intercom's risk-free purchase strategy

A playbook for pricing AI in your products

Read time: 3 minutes 56 seconds

I realised we crossed a fun milestone a few weeks ago:
1 million reads of Strategy Breakdowns articles.

You reading this email puts us at something like 1,226,989 reads.

The internet is such a quirky thing.

Hope you enjoy - thanks for being here.

– Tom

Today’s breakdown is brought to you by Kinde and Mixpanel:

What do Atlassian, Dovetail, Datadog, and Rippling all have in common?

They’re among the 38,000 developer teams that use (and adore) Kinde: user identity and access management for modern applications.

Kinde represents a new era for auth; incredible DX, beautiful screens, and real people supporting the world’s fastest-growing software companies.

Applications with more than 100,000 MAU switching to Kinde can access 6 months free, with migration support. Chat to the team.

Picture this: People are hitting your website, clicking elements, signing up, reading content, downloading resources, bouncing, returning, and purchasing.

But all of this activity is invisible to you.

Wouldn’t it be nice to visualise, analyse, and leverage all of these insights, without a technical skillset?

If you’re building a website or product, Mixpanel is the plug-and-play tool for visualising user activity, and making better decisions.

Thank you for supporting our sponsors, who keep this newsletter free

Intercom's risk-free purchase strategy

Chess Move

The what: A TLDR explanation of the strategy

Not all software should be SaaS.

Intercom, a leading customer support solution, recently adopted an interesting pricing strategy for their AI chatbot: “Fin AI”.

Rather than a fixed monthly recurring fee, Fin uses a “pay-per-resolution” structure, charging customers $0.99 for each customer inquiry it successfully resolves.

This means that Fin customers only pay when they receive the outcome they (and their own customers) want the most - a resolved conversation.


Strategy Playbook: Only charge your customers if you make them happy.

This move follows CEO Eoghan McCabe’s viral Twitter thread, announcing their $94m bet on AI customer service through several screenshots of a private internal strategy email.

Jackpot for strategy nerds

Here are some mental models you can use to think about Fin’s pricing strategy.


The how: The strategic playbook boiled down to 3x key takeaways

1.  Make it an obvious buy with pricing that starts at $0/mo

The “cost-per-resolution” pricing structure eliminates upfront financial risk for customers.

Like “cost-per-click” Google Ads, or “comission-per-successful campaign” on Kickstarter, customers are guaranteed to pay nothing if the service doesn’t deliver the expected value, which minimises psychological barrier to adoption.

Rather than passing exact API costs directly to the customer (a common AI app pricing model, often involving a “Bring Your Own Key” mechanic), Intercom only charges if their solution actually does its job. Otherwise, they eat the cost themselves.

This demonstrates the confidence they have in their own product, which subsequently improves customer trust. Customers know that Intercom doesn’t generate revenue if Fin doesn’t deliver, so the ‘perceived value’ of Fin increases.

Intercom’s trust in themself → Customer trust in Intercom.

Plus, a fixed cost-per-resolution allows companies to easily model their cost-savings when using Fin.

Products with pricing models that allow for clear cost-benefit analysis make for easier purchase decisions.

2. Align your financial incentives with customer value

By making revenue-generation dependant on customer success, Intercom’s interests are aligned with their customers.

Intercom is financially incentivised to continuously improve Fin, since the better Fin is at resolving customer issues (i.e. improving “Resolution Rate”), the more revenue Fin generates per customer.

Adding features like multilingual support or debugging tools benefits both Intercom and their customers.

This virtuous cycle means customers have high confidence they will enjoy sustained value creation.

Beyond just product improvements, Intercom is incentivised to help customers uplift the overall maturity of their Customer Support organisation. Since Fin is only as helpful as the content its trained on, Intercom provides extensive self-serve docs, best practices, and optimisation tutorials for customers to benefit from.

As a customer’s internal support content gets better, so does Fin - along with Fins resolution rates.

3. Let customers use it exactly when they need it (and not when they don’t)

For some customer segments, a change like this introduces some downsides. For example, small startups with high ticket volume might experience a necessary price increase.

But Intercom doesn’t make Fin mandatory to use. Nor do they make it a one-size-fits-all solution.

Customers who opt-in enjoy a suite of flexible customisation tools to ensure Fin works in a way that suits their specific needs.

Here are a few customisation options available:

Audience Targeting Rules - Define which customer types, customer locations, products, services, or sub-brands will trigger a conversation with Fin.

Omnichannel support - Decide where to use Fin, including Intercom Messenger (web, iOS, and Android), as well as WhatsApp, Facebook, Instagram, SMS, and Email.

Custom Answers - Build bespoke answers to the most important questions for increased accuracy, consistency, and specificity.

Workflow Embedding - Define when to trigger Fin within your existing workflows, and when to handoff to a human support agent.

Usage Limits - Set alerts and hard limits to prevent Fin from triggering beyond your allocated budget each month.

Performance Reports - Constantly monitor and assess Fin’s performance, with filtering for any parameter, to ensure you are getting what you need from it.

Fin doesn’t replace your current toolbox (if you don’t want it to).

It’s an electronic multi-tool that’s been added to your toolbox, with no upfront cost, that you only pay for each time you get tangible value from it.

Rabbit Hole

The where: 3x high-signal resources to learn more

We have a growing audience of 45,000+ readers from top companies like Google, Meta, Atlassian, Stripe, and Snapchat. If you’d like to sponsor Strategy Breakdowns, see more information here

What did you think of today's edition?

Login or Subscribe to participate in polls.

That’s all for today’s issue, folks!

Thanks again to our sponsors Kinde and Mixpanel - both tools worth checking out for anyone building products in 2024.

ICYMI, last week Strategy Breakdowns turned 365 days old, so we published a retrospective with transparent behind-the-scenes details.

Click here to read about our weird and wonderful first year.


or to participate.