When Operational Inefficiency Outpaces AI Risk: Rethinking the Bottom Tier of SaaS Service Delivery

58 days, 38 emails, 1 call, and 3 different account managers/executives trying to switch from annual billing to monthly (for accounting and cashflow recollection reasons) for a $780M public SaaS we’ve been using for over a decade.

We’re still at the upsell stage, despite the old-school “seat-based” model.

And one charging for 50 seats if you have 41 users. Annually.

A spectacular product I love and use daily, with some questionable management decisions, a bloated team in this department, and inefficient workflows (we had to follow up twice over 10 days for the lack of input).

It’s also one of my two worst individual investments in the market (next to Zoom, bought at the peak in 2021, now nearly 90% down).

The last email butchered out CTO’s name.

Oftentimes, as I argue against AI replacing humans, I stumble upon a public company with thousands of employees making embarrassing mistakes on a regular basis.

Whenever we think about agents hallucinating, we condemn them as unusable, comparing them with SUPERIOR humans as if everyone is an A-player with exquisite taste and impeccable skills.

Unfortunately, the bar has been declining year after year since around 2015, with quality far from perfect. And while perfection is both a myth and often an overkill (following the best practices from 80-20 or zero to one), some stories quickly turn to fables.

As we work closely with over a dozen Nasdaq-listed companies and different teams, we’ve met some incredible and inspiring people – often in the executive suite, senior director/VP and above, but also some respectable BDRs, account managers, marketing and product managers, SREs, and many other individuals. But as headcount grows, facing some of these uncomfortable conversations comes up more often, especially at the lower customer-facing tiers.

AI agents will wipe that specific “bottom 20 percent” layer faster than we can imagine. And quality will go up for that L1 tier that’s causing more harm than good.