Handling the “AI Discount” Question from Clients

TL; DR

• Price outcomes, not hours. AI collapses drafting time; your fee maps to judgment, QA, compliance, and results.
• Operationalize standards. Add an AI Accountability Addendum (provenance, data non-training, IP warranties, security, human signoff, remediation).
• Show your receipts. Use a Deliverable Map that labels AI acceleration vs. human decisions, with QA proofs and rights logs.
• Modernize pricing. Shift to fixed/outcome fees with SLAs; offer scope/service-level changes—not “AI discounts.”
• Use clear scripts. “Efficiency shows up as speed, scope, and reliability. If you want cheaper, we adjust scope or service levels—never the standards.”

Handling the “AI Discount” Question 

Oh, boy! Do I have a doozy for you today!

Scrolling through LinkedIn, I saw a post from a second connection.

It read, “I’ve recently been hearing lots about how agencies are finding efficiencies through AI.

“For example… overnight automation of a client brief that will process it into a draft pitch while the team sleeps. The human touch comes in for review, but what we used to pay for as clients is possibly operating very differently now. AI could be doing lots of heavy lifting.

“My question is… at what point as clients should we be asking for full transparency of AI involvement, so we can assess whether the price tag of the work is still correct?

“As someone who signs off invoices for research, strategy, and creative work… I’m starting to feel uneasy/uninformed.” 

We knew this day was coming, which is why I’ve been preaching charging for outcomes, not hours, for, well, forever. (I just looked—the earliest article I wrote on this topic was 2009!) 

If you’re not already doing that, now is the time to get your butt in gear! Because this LinkedIn post will not be the last time you hear of someone wanting to reduce their agency fees because they use AI. 

What’s Really Changing

For all of the tech bros calling for AI to take our jobs, there are the rest of us saying, “Nah, bro.” It’s not replacing us. Sure, it’s collapsing the time around our work, but we still have to be involved.

For all the hallucinations, inaccuracies, and downright false information, there is a massive need for humans. 

BUT! It has definitely saved us time (and for me personally, it has also helped with procrastination). 

We have a client who hosts a weekly webinar that is then transcribed into a podcast, YouTube channel, video snippets, newsletter copy, and social media posts.

It used to take us weeks to complete one webinar. Now it takes less than a day.

That’s the headline everyone sees. The time savings. That’s what this guy who posted on LinkedIn saw. 

Everyone overlooks the part where the real value lies: in the ability to be more strategic and accomplish even more.

Guess what happened when we started to use AI to deliver what he needed that much more quickly?

He was ecstatic! He didn’t ask to reduce his retainer. Not at all! In fact, we were able to shift some of our attention to other priorities. Another thing that made him ecstatic. 

When you’re explaining how AI is helping you, you can think about it this way:

1. AI acceleration: draft, summarize, cluster, suggest.
2. Human judgment: choose, shape, verify, make it make sense.
3. QA & legal: fact-check, source, rights, compliance.
4. Delivery & measurement: publish, pitch, amplify, learn.

There is still a lot of time and brain power in there. You’ve just shortened step one, so you can focus more on the output. Your fees should still align with the outcomes of steps two through four, as that’s where risk is reduced and value is created. 

If you’re still selling hours (and we can have a talk about why you should stop doing that), it becomes harder to make the case. When you do that, the question moves from, “Did you use AI and can you reduce your invoice because of it.” to “Did AI help you deliver faster, safer, and smarter—and can you show me where humans made the calls that protect my brand?”

When we reframe it this way, pricing makes sense again. A surgeon doesn’t discount their fees because the imaging and its results are faster. You celebrate the reduced time-to-treatment, and you still pay for the expertise that prevents catastrophic mistakes. You know, catastrophic mistakes like…your death.

Stop Itemizing Hours—Start Pricing Outcomes

If you price minutes, AI will eat your margin. If you price outcomes, AI becomes your advantage.

Clients don’t actually buy “12 hours of strategy” or “8 hours of copy.” They buy movement: qualified pipeline, high-quality coverage, faster cycles, visibility in human and AI search, and lower legal/reputation risk. 

That’s what you should be pricing. Not how long it takes you to do something. 

So, when they ask, “Did AI make this cheaper?”

You can respond with, “AI made the mechanics faster. What you’re paying for is the result—and the risk we remove to get you there.”

Anchor every one of your scopes of work to three to five measurable outcomes. Depending on goals, they might include:

• Pipeline / Demand: qualified opportunities, SQLs, demo requests, cost per opportunity.
• Earned Quality: tier of outlets, message pull-through, share of authoritative citations.
• Visibility Lift: branded/unbranded search and AI answer inclusion (are you recommended by assistants?).
• Cycle Time: time-to-first-draft, time-to-publish, time-to-placement.
• Risk Controls: zero retractions, rights secured, compliance passed on first review.

And, in your SOWs, include the following:

• Success criteria at the top (the scoreboard).
• Deliverable map showing where AI accelerates and where humans apply judgment.
• Measurement plan that ties outputs to business results.
• Change clause: if outcomes are exceeded early, scope flexes up (not price down).

This changes the conversation from “let’s discuss a lower retainer” to “how much more can you and your team do to help me reach more of these goals?”

#winning! 

How to Add AI to Your Contracts

I am going to preface what I’m about to tell you by saying I am not an attorney. I don’t even play one on TV. To ensure your AI policy is worded correctly in your contracts, call Sharon Toerek at Legal & Creative.

They work with agencies and know exactly what you need to have included.

They also have a nice AI toolkit you can buy if you don’t want their full services.  

With that caveat, there are a few things you can add to either your contract (with an attorney’s guidance) or to your statements of work.

First, in your paperwork, you can add the following phrase: “We may use AI tools to accelerate production. We remain fully accountable for quality, accuracy, rights, confidentiality, and outcomes. This addendum defines our standards.”

Here are 10 other things to consider:

1. “For each deliverable, we’ll note the stages and who/what did the work (AI-assisted draft → human edit → legal review → final).”
2. “A designated reviewer is accountable for factual accuracy, brand voice, and compliance. No AI output goes live without human approval.”
3. “We do not permit client data to train public models. We only put the bare minimum into prompts, and anything we enter is treated under the same strict privacy and security terms in our contract.” 
4. “We warrant that deliverables are original or properly licensed, free of undisclosed third-party claims. Client owns all final outputs upon payment.”
5. “We disclose categories of tools used (e.g., text generation, research synth, image gen, code assist) and hosting context (vendor/on-prem). We don’t disclose proprietary prompts.”
6. “Access is role-based; credentials stored securely; logs retained. We comply with applicable industry/regional restrictions and your security policies where provided.”
7. “Critical deliverables require named human signoff. We maintain a sign-off log available on request.”
8. “If an AI-assisted claim is incorrect, we will correct, retract if needed, and remediate at our cost within our signed agreement.”
9. “Efficiency gains are reflected in service levels, speed, and scope. Fees are based on outcomes and risk controls, not historical hours.”
10. “No live testing on the client brand without written consent. Experimental workflows are sandboxed and clearly labeled.”

You can also put these clauses in your paperwork so there isn’t any miscommunication.

Scripts You Can Steal

Now, if you’re like me when I’m learning something new or building up the confidence to stick up for our work, I need to do some role playing and practice. 

So here are some things you can practice.

When a client asks: “Did you use AI—and should this be cheaper?”

You can say, “Yep—we use AI to compress drafting time. Your fee aligns with outcomes and risk controls, including accuracy, brand voice, legal and compliance, and performance. Here’s our deliverable map showing where AI accelerated work and where humans made decisions and QA’d it. If you want a lower price, we can reduce scope or service levels, but we never lower the price on the standards.”

When procurement pushes for an AI discount, you can say, “Efficiency is already priced in as speed and reliability. You’re buying results plus indemnities, not hours. We can absolutely reprice if we change scope (fewer deliverables) or service level (slower turnarounds). Otherwise, we keep fees and deliver the higher-value package.”

When they demand your prompts or internal models, you can say, “We don’t share proprietary prompts. We do share tool classes, provenance logs, QA checklists, and data settings . That gives you transparency on risk without exposing our tradecraft.”

When you’re asked if it is safe, you can say, “Here are the three controls you care about: source logs and rights for every asset, named human QA, with an accuracy/rework warranty, and non-training settings. That’s why we price outcomes—we’re underwriting the risk.”

When a stakeholder says, “We just want hourly,”—I recently had this happen to me—you can say, “Hourly prices minutes, which AI collapses. You’ll either overpay or force bad behaviors. We focus on outcome pricing so we are aligned with your results. It also allows you predictability. If needed, we’ll provide time estimates, but we won’t peg value to a stopwatch.”

When they want to see the savings, you can say, “You already see them in cycle time and scope. AI shaved 40% off drafting; we redeployed that into deeper sourcing and message testing. Net: faster throughput, fewer revisions, better performance—without increasing risk. That’s the ROI.”

When legal or compliance is nervous, you can say, “No AI output goes live without human signoff. We maintain a signoff log, keep citations, run rights checks, and separate ‘lab’ from ‘production.’ You also have an approval process on your end to ensure nothing slips through the cracks.

When a CMO asks for a one-liner, you can say, “Don’t itemize the robot—operationalize the standards and price the outcomes.”

That last one is probably how I would have commented on the LinkedIn post. If you operationalize the standards and pay for the outcomes, it shouldn’t matter how your agency gets the results.

No AI Discount on Services

AI doesn’t give people a discount code for your services. Yes, it makes you more productive, but it still requires your oversight. More, in fact. 

If you can focus your time on creativity, judgment, QA, compliance, delivery, and learning, you’ll be able to protect your clients’ brands and move their organizations. 

That’s what your pricing should reflect.

If you take nothing else from today, take this: don’t you dare lower your fees because AI has made you more efficient. 

Instead, operationalize standards and price to the outcomes. 

When you anchor to that, the “AI discount” question stops being a tug-of-war and becomes a professional conversation about scope, service levels, and risk.

© 2025 Spin Sucks. All rights reserved. The PESO Model is a registered trademark of Spin Sucks.