Three AI models reviewed our service menu and told us where we were priced wrong. The strategy gap was 40%. The production gap went the other direction. Here is every number.
I put our full service menu into ChatGPT-4o in January 2026 and asked it to tell me what we should charge. The results were uncomfortable. GPT thought we were underpriced on strategy work by about 40%, overpriced on production work by about 15%, and roughly accurate on web development packages. When I ran the same exercise with current and past clients through an anonymous survey, buyers came in lower than both AI and our actual rates on everything. Here is the full side-by-side and what it means for how we price.
I took our current service menu, 11 items ranging from a $1,500 content audit to a $28,000 enterprise website package, and ran it through three AI models: ChatGPT-4o, Claude Sonnet, and Gemini Advanced. I gave each model our full service descriptions, three client testimonials that referenced specific outcomes, and a brief description of our niche: conversion-focused web and marketing for B2B companies between $2M and $20M ARR.
I asked each model the same question: "Based on this service description and the client outcomes described, what price range do you think is appropriate for each service, and how does it compare to market rates for comparable services?"
I also sent an anonymous survey to 22 current and past clients with the same service descriptions and the question: "What would you expect to pay for each of these services?"
Here is what I found.
| Service | What we charge | AI consensus | Buyer median | AI vs. us |
|---|---|---|---|---|
| Brand and positioning audit | $3,500 | $5,200 | $2,800 | +49% |
| Website conversion audit | $1,500 | $1,800 | $1,100 | +20% |
| Starter web package | $8,500 | $7,900 | $6,200 | -7% |
| Growth web package | $18,500 | $18,000 | $14,500 | -3% |
| Enterprise web package | $28,000 | $31,500 | $22,000 | +13% |
| SEO retainer, starter | $2,400/mo | $2,200/mo | $1,800/mo | -8% |
| SEO retainer, growth | $4,800/mo | $5,400/mo | $3,900/mo | +13% |
| Paid media management | $3,200/mo | $2,900/mo | $2,500/mo | -9% |
| Content production | $2,800/mo | $2,600/mo | $2,100/mo | -7% |
| Marketing strategy retainer | $6,500/mo | $9,200/mo | $5,100/mo | +42% |
| Fractional CMO | $8,500/mo | $12,000/mo | $7,200/mo | +41% |
The pattern is not random. AI models consistently priced strategy and advisory work significantly above our rates and priced execution work in line with or below our rates. Buyers priced everything lower than both.
All three models gave similar reasoning on the strategy gap. They assessed our strategy and fractional CMO work as high-expertise, outcome-attributable services and compared them against public data on fractional executive compensation and strategy consulting fees.
The public benchmark for experienced fractional CMOs runs $10,000 to $15,000 a month, based on what fractional executive marketplace data and professional services research shows (per Harvard Business Review analysis of fractional leadership compensation, 2024). We are charging $8,500 a month. GPT's analysis: "If your case studies support 30% or more pipeline growth attributable to your strategy work, you are leaving significant money on the table relative to comparable service providers."
The AI models are treating our strategy pricing like a consultant would: looking at outcome data and comparing against market rates for equivalent outcomes. Our buyer survey treated it like a product: comparing against the perceived cost of doing it themselves or hiring someone junior in-house.
That gap is the core pricing psychology problem for every service business. Buyers anchor on inputs, time and effort, not on outputs, the revenue impact of the outcome. AI anchors on market rates for the outcome category. The AI is probably right. McKinsey research on B2B pricing consistently shows that buyers accept higher prices when the ROI logic is made explicit and tied to specific evidence (per McKinsey, "Rethinking B2B Pricing," 2023).
On execution services, including content production, paid media management, and basic SEO, the AI consensus came in slightly below our current rates. The reasoning from all three models was consistent: "These services are increasingly commoditized by AI-assisted production tools. Market rates for execution-heavy work are compressing."
That matches what I hear from buyers. They are anchoring on what AI tools can do in-house now, and execution-only work is repricing downward. A content production retainer that faced no pushback at $3,000 a month in 2023 faces skepticism at $2,800 a month in 2026 because buyers know they can produce content faster with internal AI tools.
The services that are not compressing: anything that requires strategic judgment, client relationship management, and cross-channel integration. Strategy retainers and fractional CMO work are going up in value as execution commoditizes. The "I can do this myself with AI" effect applies to production, not to strategy that produces attribution-clear revenue results.
Being transparent: Claude Sonnet, made by Anthropic, was one of the three models I used. Its analysis of our enterprise web package was the most specific of the three. It flagged that we were pricing the package against commodity web development, not against business outcome. Its recommendation: "Consider reframing the enterprise package around the expected conversion rate improvement and qualified lead increase attributable to the rebuild. If your case studies show a $180K ARR increase from the website project, $28,000 is not a large expense. It is roughly 6 weeks of the additional ARR. Price and present it accordingly."
That was useful, practical feedback. It is also consistent with what outcome-based pricing research shows: buyers accept price anchored to ROI more readily than price anchored to deliverable scope (per McKinsey, 2023). If you can show the math, you can price to the outcome.
The buyer survey came in lower than both AI estimates and our actual rates on every single service line. That is not surprising. Buyers have price anchors and preferences for paying less. What matters is the magnitude of the gap and what it reveals about buyer psychology.
Buyers were closest to our rates on website packages, with a median gap of about $2,300 per project. They were furthest from our rates on strategy and advisory work, with a median gap of $1,400 a month. The strategy gap reveals that buyers in our survey were treating strategy work as an add-on to execution, not as the primary deliverable with independent value.
I asked three buyers directly: "How much more would you pay for strategy work if we could guarantee a specific revenue outcome?" Two of three said they would go 30 to 50% higher with a performance guarantee attached. One said he would not pay more regardless of the guarantee because his board does not approve consulting engagements above $5K a month without an established relationship.
That second constraint is real and not easily solved by pricing changes alone. It goes away with case study evidence and relationship tenure. Buyers who have seen evidence of outcomes pay for outcomes. Buyers who have not seen evidence are anchoring on their prior experience with agencies that overpromised and underdelivered.
First: raised our marketing strategy retainer floor to $7,500 a month. The AI consensus and market data supported it. We signed two new retainers at that rate within three months of making the change. One prospect walked away. The math is fine.
Second: started separating strategy from execution in every proposal. Previously, our proposals bundled strategy thinking into execution retainers as an included component. Now strategy and execution are separate line items with separate scopes and separate justifications. This makes the strategy value visible as a standalone investment rather than invisible overhead folded into production costs.
Third: built a dedicated case study deck for strategy-track engagements. Two case studies showing specific revenue outcomes attributable to our strategy work, with before and after metrics. We use it in every enterprise sales conversation for engagements above $7K a month. Conversion rate on proposals in that range has improved meaningfully since we added the deck.
Check how we have structured the Striveloom service tiers to reflect the strategy-execution separation.
Here is the actual takeaway from this experiment: AI pricing analysis is not perfect, but it is better than founder intuition for one specific thing. It has no emotional attachment to your current rates. Founders undercharge on strategy because raising rates on existing clients feels risky and uncomfortable. AI has no such attachment. It looks at the market data and tells you where you are relative to it. That is valuable input.
Run this experiment with your own service menu. Give a capable AI model your service descriptions, your three best client testimonials with outcome data, and your current prices. Ask it to compare against market rates for comparable services. Listen carefully to where it says you are below market. That is almost certainly where you are being timid.
The AI will not replace your pricing judgment. But it will tell you when your judgment is colored by risk aversion rather than market reality. That is worth running the experiment.
Yes, with caveats. AI models are useful for comparing your prices against market rates for comparable services, identifying where you are systematically above or below market, and pressure-testing your pricing rationale. They are not useful as the sole decision-maker because they lack context about your specific client relationships, sales pipeline, and cost structure. The best approach: use AI analysis as one input alongside direct buyer feedback and industry benchmark data. The gaps between all three perspectives are usually where the most useful pricing insights live.
AI models benchmark prices against public data on comparable services. Strategy and advisory work is benchmarked against fractional executive rates and management consulting fees, which run significantly higher than execution-service rates. Execution services are benchmarked against a market that is compressing as AI tools commoditize production work. The models are correctly identifying that strategy requires scarce judgment that AI cannot yet replicate, while execution increasingly uses AI assistance that buyers can access independently.
Buyers anchor on perceived input cost, not on outcome value. They estimate what it would cost them to do the work themselves or hire someone junior in-house, then apply a modest premium for expertise. This systematically undervalues outcomes that produce attribution-clear revenue results. The fix is presenting outcomes with evidence before discussing price. If a buyer sees a case study showing $180K ARR attributable to your strategy work before seeing the $9K monthly retainer number, their anchor shifts from input cost to ROI. Present evidence first, price second.
Strategy, advisory, and fractional leadership services are commanding higher rates as execution commoditizes through AI. Services that require judgment, stakeholder management, and cross-channel integration cannot be replaced by AI tools, which is making them more valuable relative to execution. Web development, particularly with unique UX and conversion architecture, is holding rates well. Pure content production, social media posting, and basic SEO services are facing the most pricing pressure as buyers can replicate parts of these with internal AI tools.
Performance-based components can be valuable additions to retainers, not replacements for base retainers. In our buyer survey, two of three buyers said they would pay 30 to 50% more for strategy work with a performance guarantee attached. The structure that works: a base retainer that covers your labor cost plus a modest margin, plus a performance bonus if specific agreed metrics are hit. The base protects your downside. The performance component aligns incentives and justifies a higher total fee. Avoid pure performance-only arrangements: they put all the risk on the agency and create adversarial client dynamics around attribution.
Founder & CEO of Striveloom. Software engineer and Harvard graduate student researching software engineering, e-commerce platforms, and customer experience. Builds the agency that ships like software — one team, one pipeline, one platform. Writes on AI agencies, web development, paid advertising, and conversion optimization.
Most agencies sell hours and activity. This is how to find the rare agency that sells outcomes and how to write a contract that actually enforces it.
Median agency hourly rate in 2026: $185. Our 87-founder survey breaks down rates by service type, team size, and niche versus generalist positioning.
Most agencies bill hours so they make more money when projects take longer. Here is the exact pattern, the four contract clauses to delete, and a checklist that has saved buyers $40K on average.
Book a free 30-minute call to scope your project. Fixed pricing, transparent timelines.
