I built a Claude-powered receipt bot on a 15 min demo call last month and completely forgot about it.
The customer reached out to tell us it's handled 1.2k invoices across their business since.
It auto-detects any invoice email:
- uploads attachments to Google Drive + renames the files accordingly
- categorizes the expense + extracts all info to Google Sheets
- texts them a notification
My prediction is that most vertical SaaS is going to be replaced by custom AI tools like this in the next few years. If a tool you need doesn't exist, you can now build it. If it does exist, you can probably build it for cheaper.
I built a Claude-powered receipt bot on a 15 min demo call last month and completely forgot about it.
The customer reached out to tell us it's handled 1.2k invoices across their business since.
It auto-detects any invoice email:
- uploads attachments to Google Drive + renames the files accordingly
- categorizes the expense + extracts all info to Google Sheets
- texts them a notification
My prediction is that most vertical SaaS is going to be replaced by custom AI tools like this in the next few years. If a tool you need doesn't exist, you can now build it. If it does exist, you can probably build it for cheaper.
graph TD
A[Claude-Powered Receipt Bot Created]
A --> B[Forgotten After Demo Call]
B --> C[Customer Follow-up]
C --> D[1.2k Invoices Handled Across Business]
D --> E[Invoice Email Auto-Detection]
E --> F[Uploads Attachments to Google Drive]
F --> G[Renames Files in Google Drive]
E --> H[Categorizes Expenses]
E --> I[Extracts Information to Google Sheets]
E --> J[Sends Text Notifications]
D --> K[Prediction]
K --> L[Most Vertical SaaS Will Be Replaced by Custom AI Tools]
K --> M[Build Your Own Tool or Make it Cheaper]
I built a Claude-powered receipt bot on a 15 min demo call last month and completely forgot about it.
The customer reached out to tell us it's handled 1.2k invoices across their business since.
It auto-detects any invoice email:
- uploads attachments to Google Drive + renames the files accordingly
- categorizes the expense + extracts all info to Google Sheets
- texts them a notification
My prediction is that most vertical SaaS is going to be replaced by custom AI tools like this in the next few years. If a tool you need doesn't exist, you can now build it. If it does exist, you can probably build it for cheaper.
I built a Claude-powered receipt bot on a 15 min demo call last month and completely forgot about it.
The customer reached out to tell us it's handled 1.2k invoices across their business since.
It auto-detects any invoice email:
- uploads attachments to Google Drive + renames the files accordingly
- categorizes the expense + extracts all info to Google Sheets
- texts them a notification
My prediction is that most vertical SaaS is going to be replaced by custom AI tools like this in the next few years. If a tool you need doesn't exist, you can now build it. If it does exist, you can probably build it for cheaper.
I wish I only had to pay for software when it delivered results.
But the reality is that success-based pricing won't work for 90% of *today's* products. Here's why ⤵
Intercom's AI pricing went viral last year. They were charging $0.99 per AI-driven resolution.
This model changes the relationship between customer and vendor:
👉 There's zero risk to trying out a new product
👉 The more you use the product, the more ROI you generate
👉 The vendor becomes 100% focused on delivering outcomes
👉 The vendor can get big $$$ since the upside is uncapped
The issue is attribution.
You want the customer to get a fantastic outcome -- and you want them to recognize that your product powered that outcome.
As soon as you start charging for success, the customer begins to rethink the results.
-- Did your product really drive the outcome?
-- Or did they drive the outcome (with a small assist from the product)?
If your product relies on people to change their behavior in order to generate ROI, then success-based pricing could set you (way) back.
If your product does the work itself, owning the service end-to-end, that opens up new monetization possibilities.
My eyes are on this next wave of SaaS: "Services as a Software" 🍿
—
🎁 I share the latest insights into SaaS growth & pricing via my newsletter, Growth Unhinged. Follow along: https://lnkd.in/exTbjKaM#saas#ai#monetization#pricing
Building a pricing page that converts isn't easy.
But Notion's nails it. 🎯
I just tore down Notion's pricing page, and here's why I rate it an 8/10:
• Smart packaging: I can identify where I belong immediately
• "Free for teams to try": Excellent risk reversal 👍
• Positive friction: Guest limits drive upgrades
• Real testimonials: Names, titles, AND company names
• Detailed FAQ: Builds confidence like a champ
A few tweaks could make it even better:
• Clarify the value prop: "One tool for your whole company" - but for what?
• Explain what "blocks" are (I'm a bit confused here)
• Move that giant scrolling comparison table to a lightbox
• Fix the weirdly aligned AI pricing toggle 🤔
Want more B2B SaaS pricing insights? I'm tearing down a new pricing page every week.
If you liked this follow me for more teardowns! 👀
Such a simple review, but it makes our team happy.
While we strive to bring the best AI products to our customers, we never forgot about our customer support.
Most SaaS companies forget about the last (s) in SaaS.
For those who don't know, SaaS stands for "software as a service".
Most tech companies think they can create a product and never talk to their customers.
In order to be successful, you can't neglect the service part.
We love showing that we have 100s of online reviews. Makes the sales process much easier.
#ai#aivoice#testimonial
Outcome based pricing is beginning to surface. Intercom and how it prices the Fin AI Agent is a popular example.
For outcome based pricing to really take off we will need to solve for attribution. The techniques and data to do this are increasingly available.
#pricing#SaaS#outcomebasedpricing#AI
I wish I only had to pay for software when it delivered results.
But the reality is that success-based pricing won't work for 90% of *today's* products. Here's why ⤵
Intercom's AI pricing went viral last year. They were charging $0.99 per AI-driven resolution.
This model changes the relationship between customer and vendor:
👉 There's zero risk to trying out a new product
👉 The more you use the product, the more ROI you generate
👉 The vendor becomes 100% focused on delivering outcomes
👉 The vendor can get big $$$ since the upside is uncapped
The issue is attribution.
You want the customer to get a fantastic outcome -- and you want them to recognize that your product powered that outcome.
As soon as you start charging for success, the customer begins to rethink the results.
-- Did your product really drive the outcome?
-- Or did they drive the outcome (with a small assist from the product)?
If your product relies on people to change their behavior in order to generate ROI, then success-based pricing could set you (way) back.
If your product does the work itself, owning the service end-to-end, that opens up new monetization possibilities.
My eyes are on this next wave of SaaS: "Services as a Software" 🍿
—
🎁 I share the latest insights into SaaS growth & pricing via my newsletter, Growth Unhinged. Follow along: https://lnkd.in/exTbjKaM#saas#ai#monetization#pricing
Vertical market software will become even more niche.
AI means that dev and maintenance costs will continue to decrease.
Combine that with influencer / owned distribution channels and you get self-sustaining software businesses at much smaller scale than is possible today.
Think SaaS or Agents designed only for 100-200 potential customers.
As an AI guy, I recognize the irony in ragging* on agents and AI all the time. However, I can appreciate YC's recent interview with Jake Heller @ Casetext for its exploration of horizontal vs. vertical AI applications.**
A year or two ago, it was common practice to dump all over AI tools as being a "ChatGPT" wrapper. But, as Jake points out, when you start getting down into specific use cases, you end up building a ton of infrastructure necessary to support even the most basic workflows. At what point does it stop being a wrapper and start looking more like a full-fledged product?
To apply this to marketing analytics, let's say I want to run some quick-and-dirty analysis across my leads and opportunities objects in my CRM. As Andrea 🤓 Lechner-Becker pointed out last week, the "right answer" looks like a data warehouse (likely with a data extraction and loading tool) with a business intelligence tool on top. Maybe an MTA tool. Maaaybe some data management capabilities in there, too.
Whoops, I just recreated a mini modern data stack, and now I'm on the hook for maintaining all this infrastructure.
As one of those aforementioned "ChatGPT wrappers" (and yes, I have no issues calling us that 6 months ago), we've found ourselves having to build:
▶ Data connectivity, credential management, and data extraction
▶ Analytical data warehousing
▶ Automated data wrangling, transformation, and cleaning
▶ Automated data modeling
▶ Query, visualization, and interpretation support (with a bespoke UX on top)
At some point along the way, we crossed the line from ChatGPT wrapper to "mini data stack in a box." 🤷♂️
We are going to see this trend continue across nearly every vertical AI application across every industry. The ones that don't are going to get stomped on by OpenAI, Google, and Anthropic.
---------------
*Pun intended
** https://lnkd.in/dGpUFSTn
Healthcare in SEA
1moAgree. We've never used any off the shelf ERP and CRM nor do we plan to.