I subscribe to ChatGPT Pro (~₹10K/mo). My use-case is to build tools that let me and my team create ads and emails at scale. We also use it for campaign ops through Codex's computer-use, creating and modifying dashboards, reviewing CRM setups, and setting up one-time or recurring automations.
The recurring work alone adds up to roughly ₹1.3L a month, about 13× the subscription. That leaves out the one-off tools, integrations, code fixes and side projects I haven’t priced.
When I look at everything together, I put the role AI fills for us closer to a ₹30 LPA marketing hire. I still provide direction and review, but it handles work I’d otherwise do myself or need marketing and engineering teams for. We haven’t needed another full-time hire to do this work.
The monthly maths
Zomunk shares flight deals. One of the ad formats we have is creating ads from the live deals on the website. We produce a static and a video for each of 2–5 deals a day, five days a week. This translates to ~65 deal ads a month. Separately, we create about 65 Gemini static ads with similar calculations. We also design + code about ten emails a month.
I've assumed certain salaries: ₹50K for a graphic designer, ₹20K for an ops intern, ₹30K for a junior marketer, and ₹1 lakh for an engineer with 2–3 YoE.
| Task | Assumed manual effort | Monthly work value |
|---|---|---|
| Static deal cards | 65 deals × 20 min | ₹6K |
| Video deal cards | 65 video deal ads × 30 minutes | ₹9K |
| Separate Gemini static ads | 65 creatives × 45 minutes of designer time | ₹15-20K |
| Deal preparation | 65 deals × 10 minutes of intern time; working assumption | ₹2K |
| Mixpanel and Axiom dashboards | 10 hours of engineering work | ₹6K |
| Reporting and campaign ops | Daily manual ops by marketer | ₹24K |
| Creator post reviews (UGC) | 4-6 hrs | ₹2K |
| Customer.io audit | One 8-hour day by the same marketer | ₹2K |
| Email design | 10 emails × ₹4,000 design effort | ₹40K |
| Email HTML | 10 emails × ₹2,000 dev effort | ₹20K |
| Total | ₹1.3L |
Ads and videos
Earlier, we'd send a deal to our designer, wait for the creative, then upload it through Meta Ads Manager. A few hours across handoffs, with a chance the deal could expire before the ad went live. If a deal expired late at night, the ad could keep running until someone paused it.
Now our webhook-fed tool creates static and video deal ads, and pauses ads when the deal expires.
For separate static creatives, we built a Gemini workflow where my team can upload reference ads and get versions tailored for Zomunk. A designer would otherwise spend around 30 minutes creating each variation.
Gemini costs us roughly ₹3–5K a month, separate from ChatGPT and excluded from this calculation.
Designing emails
Email design needs a different skillset from updating a deal-card template. A high-quality design costs around ₹3–5K per email, or ₹30–50K for ten emails, before the HTML development.
We now use ChatGPT to turn references from Really Good Emails into emails tailored for Zomunk. It writes the HTML and reviews layouts across desktop and mobile, covering both design and coding within the subscription.
Campaigns and dashboards
I use Codex's computer use to review campaigns, pull reports and check Customer.io automations. I also create and update Mixpanel and Axiom dashboards, and trace broken events through our repos. These tasks would otherwise involve a marketer for daily ops and engineering for dashboard and event changes.
I've also used ChatGPT for Meta audience syncs, monitor email performance and delivery, and analytics deep-dives.
Conclusion
Many of these tasks used to sit on my own list. With the context I give it, AI often feels almost like a clone of me handling the execution.