I know that most of you are looking forward to the long Family Day weekend. Pam and I plan to get out on the sleds as much as we can but still leave time for push-ups. Yes, yours truly signed up for CMHA’s 2000 push-up challenge that runs from February 5th to the 22nd.  At the time of writing, I am over 25% of the way there and thankfully not too sore to scribe this week’s newsletter.

Prior to Thursday, if you had only looked at the main stock indices this week, you may have thought that everything was going along swimmingly. New records were being set – a feat that was not lost on the US Attorney General. Yet, while "old economy" stocks were hitting fresh highs, a specific and violent divergence was continuing to brew just beneath the surface. For two years, the AI story has been about who was building the tools and who was going to use these tools to become more efficient and more profitable.

Since late last year, the narrative has shifted to who is being disintermediated by them, from workers to companies that were supposedly going to reap the benefits of integrating AI into their products. The belief was that those that incorporated AI the fastest would effectively create a “moat” in their respective industry, from accounting to law, from wealth management to market data services and lots of others in between. It now looks like that moat is leaking. Welcome to SaaSpocalypse.

This term was coined last year and refers to “software as a service” in an apparent apocalyptic state. What is an example of software as a service? There are plenty. Think Salesforce, Adobe, Google Workspace, and Microsoft 365. If you remember my discussions of last year, when we talked about the need for companies to successfully monetize AI in order to keep feeding the AI infrastructure spending binge. The problem is that AI has evolved to the level of “agency” or what is referred to as agentic AI, meaning that instead of asking Google Gemini something (prompting) and getting an answer or even analysis of something, AI agents can act independently on set tasks. In other words, agentic AI is the system and AI agents are the tools.

If you have been using specific software in your business, but now have the ability to build your own AI agents to perform the same tasks and create the same solutions, why do you need to buy a software license for each of your employees? In November, International Data Corporation (IDC) and IT servicing and consulting firm, Gartner, released reports predicting that by this year, 40% of job roles would involve collaboration with AI agents (in other words, 40% less demand for software licenses). Note, some of these software firms have been investing billions into AI – something ironically the market started to demand from each earnings call. The belief was that there were going to be plenty of dollars coming in from software sales to pay for the build.

Unfortunately, the very AI companies that many of these companies had been partnering with (and investing in) started offering solutions that the end user could take advantage of. Think of OpenAI (ChatGPT) and Anthropic (Claude). At the end of January, Anthropic released Claude Cowork which was a platform that could log in to a company’s enterprise software and do things like send emails and design workflows. Literally days later, Anthropic came out with specialized “plugins” and one of these was the legal plugin. It showed that it could automate things like contract reviews. It got worse last week when Gartner itself lowered its 2026 guidance for revenues, indicating that its clients were deferring purchases of its software platforms. At the same time, we had a bit of a “model war” going on as Anthropic unveiled Claude Opus 4.6 and OpenAI followed with the release of GPT-5.3-Codex. Fast forward to this week and what were individual instances of poor company performance turned into indiscriminate selling by investors across a broad cross section of software companies.

This has also shown up in the financial services industry, from wealth management to exchanges. The above chart shows the S&P500 index mapped against the NASDAQ, as well as two S&P sub-groups – investment banking (which includes wealth management firms like Charles Schwab and Raymond James) and financial exchanges and data (like S&P Global, TMX Group and LSEG). In terms of wealth management, the sell-off was triggered by the debut of agentic AI platforms that can scan 1040s, pay stubs, and meeting notes to draft personalized tax strategies in minutes.

By Thursday, the scare over an AI takeover in the services sector had spread to even the trucking and logistics sector – find examples of stocks. The trigger? New AI-powered telematics and freight-matching platforms (like the SemiCab deployment by Algorhythm Holdings) demonstrated a 300-400% increase in volume scaling without adding a single human head. Other examples included C.H. Robinson, the biggest "human middleman" in the freight world, fell sharply on Wednesday. The Russell 3000 Trucking Index experienced its worst decline in years. Many of you will have heard of Geotab, a private Canadian tech company based in Oakville, that is now a global leader in telematics or connecting fleet vehicles to the internet to help businesses manage those fleets more efficiently. The company’s CEO, Neil Cawse, was recently quoted as saying “AI is going to run operations, not just conversations”.

To say that this transition in AI technology is going to be disruptive is an understatement. Valuations among much of the tech sector were stretched last year and the reality check created by the unveiling of agentic AI solutions could play out a while longer, at least to the point where companies learn to adapt.  I believe we are in the beginning phase of a replacement effect on the labour force. Case in point, while this week’s January US jobs report came in with a better-than-expect payrolls gain of 130K, the trend in monthly growth has definitely been downward over the year. As the chart below shows, this is also showing up in consumer spending, where the monthly change in retail sales peaked at 1.5% early last year and was flat in January.

This brings me back to my core thesis from 2025. AI might increase corporate margins, but there is a fundamental paradox at play. Robots don't buy houses and they don't shop in downtown stores. If we continue to see "AI-washing" of job cuts—where firms use the tech as an excuse to boost short-term margins—we eventually hit a wall. When high-earning service professionals lose their "alpha," their spending power disappears. You cannot have a consumer-driven GDP (which accounts for 70% of the economy) if you replace the high-earning consumer with a high-efficiency algorithm. And if economic growth fades or turns negative, we have another corporate revenue headache to deal with.

Is it all gloomy? No. This might be the necessary “augmentation" that brings an out-of-balance system back into equilibrium. Let’s take the wealth management industry. Consulting company, McKinsey & Company and other groups have estimated that there will not be enough advisors over the next decade to handle the massive generational wealth transfer. True, AI might eliminate the “commoditized advisor”—those that just sell products. But for a full-service wealth manager, AI creates efficiencies around investment management and customer service that allows for more time to spend on the higher value holistic needs of the next generation. Indeed, the shortage in manpower to service that wealth transition could be more than made up for in efficiency.

Even in this example though, it is important to note that AI cannot replace the human touch. Yes, it can generate a tax strategy, but it cannot understand legacy. It cannot walk you through the emotional complexity of a family business transition or provide the calm judgment required when the markets inevitably experience higher volatility. In a world of automated expertise, the only true alpha left is human trust. We'll keep the machines in the basement and the humans at the table.

On behalf of the Pyle Wealth Advisory team, have a wonderful weekend.     

Andrew Pyle