How Should an Ecommerce Agency Price AI-Powered Work?

An ecommerce agency should price AI-powered work on what ships and what it changes, not on hours. For most agencies the best fit is a hybrid: a base retainer for review, merging and strategy, a set monthly volume of shipped improvements, and an outcome bonus only where the client has clear goals and measurable KPIs.

This guide is part of the series on what an AI-native ecommerce agency is. It covers the practical question that follows: once AI agents do much of the finding and drafting, how should an ecommerce agency price AI services so the agency stays profitable and the client sees more value?

Why does AI break hourly pricing for ecommerce agencies?

Hourly billing ties an agency's revenue to time spent, and AI agents shrink the time. Kent Lewis of NextNW put the problem plainly in Agency Pricing Is Having Its Kodak Moment (June 18, 2026): "When a task takes fewer hours, revenue declines even though the client receives the same or better outcome."

Clients feel the same shift from the other side. Rick Watson, founder of RMW Commerce, said on BigCommerce's Keeping Commerce Weird podcast that "what AI is doing is reducing everyone's expectations of what they should pay." An agency that keeps selling hours ends up discounting its own speed.

The volume of work does not shrink, though. John Duncan, CEO of the Salesforce Commerce Cloud agency 64labs, argued in Managed Agents, Not Managed Services (August 31, 2026) that "the cost per unit of work is collapsing, so the volume of commissioned work goes up, not down." A pricing model for AI-powered work has to capture that larger volume instead of the shorter hours.

What pricing models can an ecommerce agency use for AI-powered work?

Ecommerce agencies pricing AI-powered work generally choose between four models, and most end up combining two of them. The comparison below summarizes how each model bills, when it fits and what to watch.

Hourly or time and materials still suits uncertain scope and advisory work, such as a replatforming assessment. For delivery work that agents speed up, hourly billing turns every efficiency into lost revenue.

Productized packages sell a fixed deliverable at a fixed price, such as a quarterly accessibility review of top pages or a Core Web Vitals review of key templates. Agents make these packages cheaper to deliver, so the margin improves as long as the deliverable is defined tightly.

A retainer of shipped improvements keeps the predictability clients like but changes what the retainer buys: review, merging, strategy and an agreed volume of shipped changes each month, instead of a bank of hours.

Outcome or hybrid pricing ties part of the fee to measured results. Common Thread Collective, an ecommerce growth agency, describes the hybrid in Ecommerce Agency Pricing in 2026 (March 18, 2026) as a base retainer plus performance bonuses that "gives the agency a floor while tying upside to results."

When does outcome-based pricing work for an agency?

Outcome-based pricing works when the client can say exactly what success looks like. Rick Watson said on the Keeping Commerce Weird podcast that "the limitation of the outcome model is usually client driven," because many clients "can't articulate goals, KPIs, or deliverables beyond a number in their head."

The metric matters as much as the model. Common Thread Collective warns that fees based on platform-reported ROAS "can incentivize inflated numbers," and argues that "the pricing model matters less than what metric the agency optimizes for." Outcome fees are safer when lift is measured in orders, revenue or conversion on the site itself.

Outcome pricing is already showing up in Noibu deals. In one, an agency paid on incremental growth took over a merchant's CRO and web work for 2026. For that kind of agency, anything that finds and drafts more improvements raises what the agency earns.

How do you set a price for an AI-powered agency retainer?

Use these six steps to move a client from hours to a retainer built on shipped work. They apply whether the agency serves solo Shopify merchants or enterprise teams on Salesforce Commerce Cloud.

1. Split the judgment from the drafting. List what agents now do for each client, such as finding problems and drafting changes, and what only your people do: choosing what fits the brand, reviewing code, merging and owning the result. The second list is what the client is really paying for.
2. Define the unit of value. Make the unit a shipped improvement with a before and after measurement, not a ticket or an hour. A unit might be a merged pull request that removes a checkout error, a rewritten product page or a completed experiment.
3. Price the base retainer on judgment. Set a monthly base that covers review, merging, strategy, client calls and reporting. This is the floor that keeps the agency whole even in a slow month.
4. Set volume tiers for shipped work. Offer two or three tiers by how many improvements ship each month and how many stores or page groups are covered. Clients who want more work move up a tier instead of buying more hours.
5. Add outcome pricing only with agreed measurement. If the client can name goals, KPIs and a baseline, add a bonus tied to measured lift in orders, revenue or conversion. Measure on the site, not on ad platform dashboards.
6. Report what shipped and what moved. Each month, list every change that went live, who approved it and what it changed. Reporting in outcomes is what lets the agency defend or raise the price at renewal.

A retainer built this way can be raised at renewal on evidence. The Blue Badger customer story describes how the Montreal agency increased client retainers by 20% after adding Noibu to how it serves clients.

Should an agency charge clients for the AI tools it uses?

One participant quoted in the NextNW article put it simply: "You don't bill for AI. You bill for outcomes and value." Clients rarely want a line item for a model or an agent; they want to know what shipped and what it changed.

When a tool runs on the client's own store data, two approaches keep the relationship clean. Either pass the tool cost through at cost with a clear label, or bundle it into the retainer tier and price the tier on the work it produces. Both avoid marking up software the client could buy directly.

Pricing also decides whether agents get used at all. In one Noibu trial, an agency that bills only for work the client requests chose not to run scheduled agents from its own account, because proactive work had no way to be paid for. A retainer of shipped improvements removes that friction.

How does AI pricing change the agency's relationship with the merchant?

Pricing on shipped work keeps the agency at the center of the relationship. The agency stays the reviewer and merger of record: agents find and draft, a named person at the agency reviews and merges, and the merchant approves what goes live. For the review side, see how to keep AI-generated pull requests reviewable.

Merchants also get more for the same spend. Over 90% of website errors are never reported, so a large share of useful work never reaches an agency's queue as a ticket. When agents surface that work from store data, the agency can ship more each month. In one Noibu trial, a merchant moved its development agency from pay-as-you-go to a retainer once agent findings created a steady flow of work.

Outcome reporting needs a fair test of whether a change helped. How to know if a website change actually improved conversion covers the measurement side, and how to split a budget between a CRO agency, an in-house team and AI agents covers how merchants think about the spend.

Where does Noibu fit in AI-powered agency pricing?

Noibu gives agencies the finding, drafting and measurement that a shipped-improvements retainer depends on. Noibu's six AI agents work in the Noibu console and in AI tools teams already use, such as Claude, ChatGPT, Microsoft Copilot, GitHub Copilot and OpenAI Codex.

The CRO Agent finds underperforming product pages and rewrites them. The Bug Resolution Agent traces top errors to the codebase and opens a pull request for review. The Performance Agent measures Core Web Vitals from real shopper traffic and stages changes. The ADA Agent audits high-traffic pages against WCAG and reports gaps. The A/B Testing Agent runs experiments, measures lift and reports the winning variant, and the ROAS Agent scores paid traffic on orders rather than clicks. Together they give an agency a steady supply of work to price and the measured results to report.

Every agent runs the same loop: find, build, approve, ship, measure. Nothing ships without approval, so the agency keeps the review and merge role that its pricing is built on. For agencies that hold the client's repository, see how to connect AI agents to GitHub when you do not own the repo.

Noibu is AI for ecommerce operations, the always-on optimization system that keeps online stores running at their best. A team of supervised AI agents finds what's broken, slow, or leaking revenue, does the work, and hands results back for human approval. Trusted by retailers like Free People, Guess, Liquid I.V., and Ariat.

Frequently asked questions

Related reading

• What is an AI-native ecommerce agency?
• How Blue Badger increased client retainers by 20% with Noibu
• How to keep AI-generated pull requests reviewable

See the work agents find on a client store: run a free website audit →