My new boss wants me to build an internal system that automates campaign setup across Google and possibly other advertising platforms. He believes it can be completed in about two months, partly because he expects AI to speed up the work.
I have spent the last 15 years building and maintaining client websites, ecommerce sites, CMS integrations, analytics, tag management, SEO, and tracking systems. I know PHP, JavaScript, SQL, APIs, and related tools, but I have never managed advertising campaigns or worked on this kind of campaign-management platform. I am also the only developer at our 15-person marketing agency, already responsible for roughly 60 websites and many analytics and tag-management accounts.
My concern is not simply whether I could eventually learn the technical pieces. A system like this would require understanding the media team's workflows, integrating with changing advertising APIs, handling sensitive budgets safely, testing thoroughly, and providing ongoing maintenance. We do not have anyone else who could support the system, and building it would take time away from existing client work. Similar software previously cost the company hundreds of thousands of dollars and took an outside development firm about a year to build.
How should I present this to my boss? I want to explain that the real issues are scope, risk, staffing, maintenance, and opportunity cost—not just whether I am willing to learn new programming skills.
3 Answers
Look for existing campaign-management tools before agreeing to build a custom one. If the company only needs a workflow around existing advertising platforms, purchasing or configuring something may be cheaper and safer than creating an application from scratch.
If your boss still wants a custom system, ask him to approve a written project plan with a budget, milestones, security and spending safeguards, ownership after launch, and dedicated maintenance time. Make it explicit that your current website, SEO, analytics, and client responsibilities will have to be reduced or reassigned. If he refuses to acknowledge those tradeoffs, then the issue is a management decision—not your inability to code.
Don’t frame this as “web development versus software development.” That distinction will probably turn into an argument, especially since a campaign tool could technically be built as a web application. Frame it around business risk and capacity instead.
Put together a short proposal showing what already exists, what a realistic first version would include, how long discovery and development would take, who would maintain it, and what happens if a bug spends the wrong amount of advertising money. Then give your boss clear options: buy and configure an existing product, hire an outside team, or formally reprioritize your current client work and fund the project properly. Also get the priorities and expectations in writing.
Exactly. The strongest argument is not “I can’t do this.” It’s “Here is the cost, timeline, operational risk, and work that will be delayed if I do this alone.”
You probably have more transferable skill than you think. Building ecommerce systems, working with APIs, databases, authentication, and integrations is software development, even if campaign management is new to you. A small proof of concept might be a reasonable learning opportunity.
However, a proof of concept is very different from a reliable internal product. You should not promise a production system in two months while continuing to support dozens of clients. Ask for a defined scope, dedicated development time, training from the media team, testing resources, and a maintenance plan. If those conditions are not available, the request is not realistically staffed.
Learning the domain is also part of the schedule. You would first need the media team to explain their process before you could safely automate it, and that effort should be included in the estimate.

You can also explain that AI may help write code, but it does not replace requirements gathering, API expertise, testing, monitoring, incident response, or someone accountable when an integration changes.