Buy SaaS... AI Champions not AI Roadkill.

Looking at the broader market for potential dislocations and opportunities, there seems to be a great opportunity in SaaS, with many companies down almost 70%. We're all familiar with the doomsayers who set off the "SaaS-pocalypse," but in attempting to disaffirm their arguments — and with AI's breadth of ability growing daily — it's best to assess what AI can't and won't do, especially in the near term.

What AI Can't Do:

1. AI can't walk into a Chief Technology Officer's office with a demo and close a sale for a software package worth $1MM a year.

   
   

2. AI can't attend a trade conference and explain to operators the value it will bring.

   
   

3. When IT systems fail and corporate executives call to assess the magnitude of the problem, AI can't explain the problem and promise it will be fixed with any degree of certainty — certainty that's critical for the continuity of business and downstream stakeholders.

   
   

4. AI can't configure itself and integrate with a business without IT professionals tying in legacy systems, and without someone who truly understands the business informing the digitization of process. There are too many firewalls, existing security measures, and security risks created/destroyed when this is implemented blindly.

Truths About Most SaaS Customers:

1. AI and information systems aren't their area of expertise — or at least they've deemed other activities more pertinent to their operations and customers than developing IT systems in-house.

   
   

2. IT systems, despite being a non-core competency, are often paradoxically core to their businesses, and mistakes can't be afforded.

   
   

3. Most SaaS companies serve unique processes and trade secrets that differ greatly, even within the same industry.

What This Means:

1. AI lacks the ability to adequately market itself. An AI → Human sales pipeline will always be inferior to a Human → Human sales structure, especially when the decision results in large recurring costs and structural IT deficits or advantages far into a business's future.

   
   

2. In a crisis, AI infrastructure will be able to self-diagnose but won't be able to implement certain solutions in the physical domain (a bad chip, a flipped breaker, a short circuit), and will likely be unable to offer customers the degree of certainty they need in a crisis.

   
   

3. Existing SaaS companies are already well integrated with legacy systems — if they aren't the legacy system themselves — and don't have to worry about jumping the hurdles that a self-implementing AI might, hurdles that could cause untold security vulnerabilities.

   
   

4. Existing SaaS companies, especially vertical-market SaaS, have built industry competence over decades, as well as a reputation for honoring trade secrets. AI, by contrast, can absorb and integrate your processes, making them queryable by your competitors — especially when models aren't run on-premise.

How SaaS Will Have to Change:

Existing SaaS companies will have to adapt as quickly as possible, but since they already have industry-specific expertise and technological know-how, this shouldn't be that difficult. In the beginning, AI will be an additional expense that will have to be passed on to customers or absorbed through depressed margins. However, as administrative burden at SaaS companies drops and existing employees become more efficient, and as costs fall from tailoring models to specific industries (a golf course management software doesn't need to recite Homer), companies will likely return to a historically normal margin profile — swapping human costs for tokens. SaaS customers will likely pay similar rates, and maybe even see cuts if pricing pressure enters a vertical, while receiving a tailored AI workflow with a usage fee for consumption above a given limit. A usage fee that will approach nothing as models become more narrow and token prices collapse as token capacity enters the market.

"Traditional SaaS is the best channel to sell and implement AI as we know it today — tailoring it to customer needs and pushing it into the organization through normal sales channels, utilizing trusted, secure relationships and legacy systems."

Sifting through the ashes (What you want in SaaS investment):

1. Has a Strong unlevered balace sheet, this will allow them to float any disruption in free cashflow that might arise durning a bumpy transition.

2. Manager that is taking the threat seriously and implimenting improvements to legacy products.

3. Are critical to the businesses they serve (Systems of Record / Sales / POS / Ect.)

4. Serve non tech-savvy customers/businesses.

Companies to Look at:

1. Roper

2. Constellation Software

3. Salesforce

4. App Lovin

5. Service Now

6. Work Day

7. Daily Journal Corporation

8. Tyler Technologies