Your SaaS Stack Has a Hidden AI Tax. Find It.

Are your SaaS tools charging you for AI you never turned on?

Probably yes. Software vendors have figured out that bundling AI features into existing plans is an easier sell than launching a separate product. So they raise prices, repackage tiers, or quietly flip a default setting, and the charge hits your card before anyone on your team notices. This is what Forbes recently called the "AI tax": the incremental cost of AI features embedded in tools you already pay for, whether you use those features or not.

For a 20-person company running 15–25 SaaS tools, that adds up fast. Most operators we talk to have no idea what percentage of their monthly SaaS spend covers features their team has never opened.

How did this happen so fast?

The pattern is consistent across vendors. A tool you've used for years announces an "AI-powered" update. The new plan that includes it costs 15–30% more. Your existing plan gets quietly sunset or feature-capped. You either upgrade or lose functionality you depend on.

Microsoft 365 Copilot is the clearest example. The add-on runs $30 per user per month on top of existing licensing. A 25-person team that got nudged into it without a real evaluation is spending an extra $9,000 per year on an assistant most of them have never configured. Salesforce's Einstein GPT features are now baked into tiers that cost meaningfully more than what many SMBs were paying two years ago. HubSpot has restructured its pricing around AI capabilities at nearly every tier.

None of this is inherently wrong. AI features can be worth the money. The problem is that most teams are paying without deciding.

What does a 30-minute SaaS audit actually look like?

This is not complicated. It's a spreadsheet exercise. Here's how to run it:

Step 1: Pull every active subscription (10 minutes)

Check your credit card statements, your bank feed, and any software your IT or ops person manages centrally. Tools like Ramp or Brex can surface recurring charges automatically if you're using them. List every tool, the current monthly cost, and the plan tier.

Step 2: Flag the AI-tier upgrades (10 minutes)

For each tool, check the vendor's current pricing page. Compare what you're paying against their base tier. If there's a gap, read what's in the upgraded tier. If the delta is explained by AI features, flag it.

Step 3: Score usage honestly (10 minutes)

For each flagged tool, ask one question: is anyone on your team actively using the AI feature? Not "aware of it." Actually using it, in a workflow, in the last 30 days. If the answer is no, that's waste.

A simple scoring table helps:

| Tool | Monthly Cost | AI Feature Included | Team Using It? | Action | |---|---|---|---|---| | HubSpot | $800 | AI content tools | No | Downgrade or evaluate | | Zoom | $180 | AI Companion | No | Disable or downgrade | | Salesforce | $1,200 | Einstein features | Partially | Keep, train the team | | Microsoft 365 | $750 + Copilot | Copilot add-on | No | Remove add-on |

You're not trying to eliminate AI from your stack. You're trying to stop paying for features nobody is using.

Which tools are most likely hiding charges?

Based on what we see across SMB clients, these categories have the highest rate of AI-tier drift:

• CRMs: Salesforce, HubSpot, Zoho. All have restructured pricing around AI in the last 18 months.
• Productivity suites: Microsoft 365 Copilot is the biggest dollar exposure. Google Workspace Duet AI (now Gemini) is similar.
• Video and comms: Zoom AI Companion, Otter.ai, Loom. Often bundled into team plans.
• Design and content: Canva, Adobe Creative Cloud. Both have moved AI tools into standard plan pricing.
• Project management: Notion AI, Monday.com, Asana. Add-ons that frequently get auto-enabled during renewals.

This is not an exhaustive list. Any tool that sent you a "we've added exciting new AI features" email in the past year is worth checking.

Is any of this AI spend actually worth keeping?

Some of it, yes. The audit isn't about cutting AI. It's about making an actual decision instead of letting the vendor make it for you.

The right question for each flagged tool is: if we properly set this up and trained two or three people on it, would it save us more than it costs? That's a different question than "do we have access to it."

If you can't name one specific workflow where the AI feature would run every week, you're not ready to pay for it.

For tools where the answer is genuinely yes, the next step is implementation, not just keeping the subscription. An AI feature you're paying for but haven't configured is still waste. It's just more frustrating waste because the potential is real.

What we'd actually do

• Run the audit this week, not next quarter. Block 30 minutes, pull the statements, build the table. Most SMBs find $200–$800 in monthly waste on the first pass. That's real money.
• Set a simple policy before the next renewal cycle. Any AI feature add-on over $50/month requires a named owner, a stated use case, and a 30-day check-in. If no one owns it, it doesn't renew.
• If you find features worth keeping, actually deploy them. A decision to keep Copilot or HubSpot AI should come with a 2-hour team session to configure it for your actual workflows. Paying for AI and not using it is the worst outcome of this whole exercise.

If you want to work through this with other SMB operators who are doing the same thing, the community at skool.com/aiforbusiness is where those conversations happen.