Salesforce is down nearly 40% from its 2025 highs. Workday is down 22%. The financial press has a name for it: "Software-mageddon."
This is not a valuation correction. It is a structural shift. Enterprises are cancelling per-seat SaaS licenses and replacing them with custom AI-driven workflows. The software they were paying per user per month for? They are building something better, faster, and cheaper with AI.
If you run a professional services firm in East Africa, this shift matters to you, even if Salesforce was never in your budget.
What's Actually Happening
For the past two decades, enterprise software ran on a simple model: buy seats. One seat per user, monthly fee, forever.
That model is breaking.
Anthropic's Claude Coworker and Salesforce's own Agentforce are making it possible for enterprises to build autonomous AI agents that handle work that previously required a licensed seat and a human behind it. Transaction reconciliation. Client onboarding. Data entry. Report generation. All being handed to agents.
The companies surviving this shift are running on consumption-based pricing. You pay for what the AI does, not for how many humans sit in front of a screen.
The seat-based model assumed value came from people using software. The new model assumes value comes from outcomes. That is a fundamentally different bet.
For vendors who built entire businesses around recurring seat counts, this is an extinction-level event. For the businesses paying those vendors, it is an opportunity.
- Legacy SaaS vendors with seat-based models are under structural pressure
- AI-native platforms with consumption pricing (Palantir AIP, for example) are gaining
- Enterprises are building custom workflows where they were once buying licences
The Goldman Signal
The most telling data point this week: Goldman Sachs is deploying Claude-based AI agents for transaction reconciliation, trade accounting, and client onboarding in banking operations.
This is Goldman Sachs. One of the most cautious, compliance-heavy organisations on the planet. They are handing core financial workflows to AI agents.
When Goldman moves, it signals that the technology is ready, the risk-reward calculation has shifted, and the question is no longer whether to adopt but how fast.
Kenyan and East African professional services firms are roughly 18 months behind this curve. That gap is closing faster than most people expect.
What This Means for East African Businesses
Here is the practical implication.
Right now, many firms in Nairobi are running operations on a combination of spreadsheets, WhatsApp threads, and whatever SaaS tool the MD discovered at a conference. Some have adopted Zoho, QuickBooks, or a vertical SaaS product built for a US market and adapted, badly, for Kenya.
The same forces disrupting Salesforce and Workday are about to hit those tools too.
The firms that get ahead of this are the ones building systems that fit their actual workflows, not adapting their workflows to fit someone else's software.
At Kaara.Works, we have seen this pattern directly. Petrus CRM, which we built for a Nairobi accounting firm managing 137-plus clients and KES 30M in annual profit, was built because no existing CRM mapped to how that firm actually worked. The custom system did not just replace a SaaS tool. It replaced five tools and a pile of spreadsheets, and gave the firm visibility it had never had.
A similar story is playing out at Nanyuki Active, where we are replacing an expensive third-party app with a custom system built around M-Pesa credits, class booking, and member milestones that actually match how a Kenyan gym operates.
The right system is not the most popular one. It is the one built around how your business actually works, in the market you actually operate in.
The Firms That Win
The transition from seat-based SaaS to AI-driven custom workflows does not happen automatically. It requires someone who can look at a complex business operation and know exactly what to automate, what to keep human, and how to connect the two.
That is the skill gap right now. Not AI. Not software development in isolation. The ability to understand a business process deeply enough to build around it.
The firms that win in the next 18 months will be the ones that get this right first in their sector.
If your firm is running on a stack of tools that were never quite right for how you work, this is the moment to fix that. As the barrier to building software has collapsed, the standard for what counts as 'good enough' has shifted permanently — [here's why mediocre software is no longer a viable position for any business](/blog/2026-02-27-world-class-or-irrelevant).
Building something like this? let's talk.
Want to discuss AI for your business?
Let's talk about how custom software can transform your operations.