My Agency Only Bills for Work I Request. How Do I Get Proactive Site Improvements?
If your agency only bills for work you request, proactive improvements have to start on your side. Get a ranked list of what is broken, slow or underperforming on your store, approve the items worth doing, and send each one to the agency as a clear, scoped ticket. The agency keeps shipping; you decide what it ships.
This guide is for merchants whose agency, developer or dev shop works on a request-only basis: it builds what you ask for and bills for it, and it does not go looking for problems on its own. That model is common and reasonable. It just means proactive improvement needs a different starting point.
- A request-only agency does exactly what the contract pays for: the work you ask for.
- Proactive improvements start with a ranked list of what is broken, slow or underperforming on your store.
- Each approved item becomes a scoped ticket, which is ordinary billable work for the agency.
- AI agents can find and rank the work and stage changes for review; the agency still reviews and ships.
Why does a request-only agency never do proactive work?
A request-only agency is paid for the hours or tickets a client asks for. Time spent combing through analytics, session replays or speed reports to find problems nobody mentioned is time the agency usually cannot bill. So it does not happen, even when the agency is skilled and the client relationship is good.
The gap matters because the most expensive problems are often the ones nobody reports. Over 90% of website errors are never reported by the shoppers who hit them, so a broken promo code field or a slow mobile product page can run for weeks without a single ticket being filed.
None of this means the agency is the problem. It means the job of finding the work has no owner. Once someone owns that job, the agency can do what it already does well: build and ship. For a wider view of how agencies are adapting, see what an AI-native ecommerce agency is.
How do I get proactive site improvements from a request-only agency?
Proactive improvements from a request-only agency come from a simple weekly loop that you own and the agency executes. The five steps below work under most existing contracts, because every item reaches the agency as a normal request.
- Get a ranked list of what needs work. Use a tool that watches real shopper sessions and ranks what it finds by revenue: errors, slow page groups, underperforming product pages and checkout drop-offs.
- Decide what is worth doing. Pick the few items with the clearest revenue impact. Skip anything you would not pay the agency to do.
- Turn each item into a scoped ticket. Give the agency the page, the evidence, the expected outcome and how you will check it. A clear ticket is billable work the agency can quote and ship.
- Let the agency review and ship. The agency keeps ownership of the code it maintains. Where an AI agent has staged a change or a pull request, the agency reviews it like any other contribution.
- Measure what moved. Compare conversion, speed or error rates on the changed pages before and after. Feed the result into next week's list.
Step three is where most proactive work stalls. A vague request such as "the checkout feels slow" turns into an hourly investigation; a scoped request with the page, the evidence and the expected outcome turns into a quote. For a template, see how to write a dev ticket from a site finding that an agency will act on.
What should the weekly list of improvements include?
A useful weekly list is short, ranked and specific enough that the agency can quote each item without an investigation. The table below is the checklist to hold each item to before it becomes a ticket.
| Item | Include | Not enough |
|---|---|---|
| Where | The exact page or page group, device and browser | "Checkout" with no device or step |
| Evidence | Session replays, error details or a speed measurement from real shoppers | A screenshot from one office laptop |
| Revenue impact | Sessions affected and the conversion or revenue at stake | "It looks bad" |
| Expected outcome | What should change, and the number you will check | "Make it better" |
| Owner and review | Who reviews the change and who approves it going live | No named approver |
If you want the list delivered to you rather than built by hand, a scheduled report can do the first pass. See how to get a daily AI report on your store’s errors, speed and conversion.
Should I change my agency contract to get proactive work?
Changing the contract is an option, not a requirement. If you bring the agency a ranked list each week, every item is already requested work, so the current terms still apply. A contract change makes sense when you want the agency itself to own discovery, for example through a proactive retainer, a fixed monthly scope of improvements or a shared target such as checkout conversion.
Agencies are rethinking these models as AI changes how much finding and drafting work costs. For both sides of that conversation, see how AI agents are changing ecommerce agency retainers and whether you should worry if your ecommerce agency uses AI.
Whichever route you choose, keep approval with you. Proactive does not mean unreviewed: nothing should go live on the store without the merchant signing off.
Where do Noibu's AI agents fit with a request-only agency?
Noibu's agents take on the part of the loop that has no owner today: finding and ranking the work. The CRO agent finds underperforming product pages and rewrites them to convert better, with changes you can review and edit before publishing. The Performance agent measures Core Web Vitals from real shopper traffic, ranks the worst-performing page groups and stages changes for review. The Bug Resolution agent identifies the top fixable issues, traces the root cause to the codebase and creates a pull request for review.
Every agent follows the same loop: find, build, your approval, ship, measure. Nothing ships without approval. The agents work in the Noibu console and in the AI tools teams already use, such as Claude, ChatGPT and Microsoft Copilot, through Noibu MCP, so the merchant and the agency can both see the same findings.
For a request-only agency, a staged pull request or a drafted page change is a head start, not a takeover. The agency reviews it like any other contribution and ships it under the work it already bills for. To see how the CRO agent works on your own store, read Noibu’s CRO agent explainer.
Frequently asked questions
Why won't my agency suggest improvements on its own?
Most agency contracts pay for hours or tickets the client requests, so time spent hunting for problems nobody asked about is unbilled. The agency is following the contract. If you want proactive work, someone has to find the opportunities and turn them into requests, and that can be you, a tool, or a separate retainer.
Do I have to pay my agency more to get proactive site improvements?
Not necessarily. If you bring the agency a ranked, scoped list of changes, each item is ordinary requested work under your current terms. You only need a new arrangement if you want the agency itself to do the discovery, which usually means a proactive retainer or a fixed monthly scope.
Can AI agents replace my ecommerce agency?
AI agents are best treated as infrastructure the agency builds on, not a replacement. Noibu's agents find issues and opportunities, rank them by revenue and stage changes or pull requests for review. Your agency still reviews, ships and owns the code it maintains, and you approve what goes live.
How often should I send my agency a list of improvements?
A weekly list works for most stores: short enough to act on inside one sprint, frequent enough that a new problem does not run for a month. Send urgent items, such as a checkout error affecting revenue, as soon as they are found rather than waiting for the weekly list.
What is a free way to start finding proactive improvements?
Start with Noibu's free website audit, which lets the AI agents run on your store's own data and rank the biggest drop-offs with the cause behind each one. You can hand the top findings to your agency as tickets the same week.
The short version
A request-only agency is not going to find your store's problems for you, and it is not paid to. Own the finding and ranking yourself, send the agency scoped tickets for what you approve, and measure what moved. That keeps the agency doing what it does best and gives you proactive improvements without tearing up the contract.
The free starting point is Noibu’s free website audit: the AI agents run on your store's own data and rank the biggest drop-offs with the cause behind each one, so your first week of tickets is ready to send.
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.
