By this time we had already satisfied about 20 customers.
We were delivering solid results, but we were struggling with prospecting.
So, we decided to focus on upselling to our current clients.
Why? Because they already trusted us and had seen the value of our work.
We needed a way to deliver immediate, measurable value.
That’s when we've made 3 iterations that you can use:
- A Custom GPT for identifying automatable tasks
- Airtable x Zapier to analyze job descriptions and uncover process automation opportunities
- An Agent AI that observes real workflows and gives concrete automation reports
The goal? Help clients discover hidden tasks that could be automated and show them the exact ROI they’d gain by doing so.
Here’s how we used this system to close a $50K upsell.
And you can do the same for your agency right now!
Custom GPT
The power of GPT to save you headaches
The goal? Get clients to interact with the bot and return to us with clear, actionable automation ideas.
What we did
- We sent the custom GPT to our customers
- Clients were asked to chat with it about their most repetitive and time-consuming tasks.
- The GPT would analyze their inputs and provide automation suggestions based on the frequency and complexity of their tasks
What we learned
- Engagement: Clients enjoyed interacting with the GPT. It was simple and conversational, making it easy to get insights without a heavy technical process.
- Depth of understanding: The GPT could identify surface-level tasks like data entry or scheduling. However, it struggled to dive into more complex, multi-step workflows. It provided a starting point but lacked the ability to propose comprehensive automation solutions.
Benefits
- Ease of use: No onboarding or heavy training was required. Clients could immediately start using it.
- Scalability: We could roll it out to multiple clients quickly, with minimal adjustments.
- Quick ROI estimates: The bot provided fast, automated ROI projections for basic tasks, helping clients understand the potential savings.
Drawbacks
- Limited depth: It worked well for simple tasks, but for more complex processes, it didn’t provide enough actionable insights.
- Follow-up needed: Clients often came back to us with general ideas for automation but still needed our team to dig deeper and refine those ideas into real solutions.
Results
- Out of all the clients we tested this with, 6K in upsell was generated directly from the GPT’s suggestions.
- Of the 20 customers we sent it to, only 10 discussed it with GPT, and 6 came back to us with ideas for automation. Of the 6, we closed 2 for which we did the automations.
- It gave us an easy entry point, but the real impact on revenue was modest compared to more advanced tools.
<aside>
<img src="https://prod-files-secure.s3.us-west-2.amazonaws.com/8617aa0f-456a-4d2e-b591-b0dc6c3144b7/cf41af85-4efa-426a-bf40-bb6d0a6c0a9a/fast-forward-double-right-arrows-symbol.png" alt="https://prod-files-secure.s3.us-west-2.amazonaws.com/8617aa0f-456a-4d2e-b591-b0dc6c3144b7/cf41af85-4efa-426a-bf40-bb6d0a6c0a9a/fast-forward-double-right-arrows-symbol.png" width="40px" /> I want to test custom GPT with my customers
</aside>
Airtable x Zapier
No need for customer input
The goal? Automatically get real insights based on their job descriptions, which will give us real indications.
What we did
- Using Zapier, we built a workflow that would automatically scrape the job descriptions from job offers links.
- The descriptions were then analyzed to find repetitive, time-consuming tasks that could be automated.
What we learned
- Automation-ready data: By pulling data directly from job descriptions, we were able to generate real insights without needing clients to interact manually.
- Quick identification: The automation setup allowed us to quickly identify tasks without the need for their inputs
Benefits
- Fully automated process: Clients didn’t have to lift a finger. We handled the entire process of extracting and analyzing job descriptions automatically.
- Immediate insights: The automation generated ready-to-use reports that made it easy for us to approach clients.
Drawback
- Standardized approach: The task identification leaned heavily on standard roles and industry practices, which could sometimes overlook unique tasks specific to the client’s business.
Results
- This approach generated $4K in upsell across the clients we tested it with.
- Of the 20 customers we sent it to, we were lucky: 18 had current job offers, and Zap found automatable tasks for 15 of them. On the other hand, we only managed to close 1 contract with one of the companies.
- The fully automated nature of the process saved us time, but the results weren’t as tailored or comprehensive as we hoped.
<aside>
<img src="https://prod-files-secure.s3.us-west-2.amazonaws.com/8617aa0f-456a-4d2e-b591-b0dc6c3144b7/cf41af85-4efa-426a-bf40-bb6d0a6c0a9a/fast-forward-double-right-arrows-symbol.png" alt="https://prod-files-secure.s3.us-west-2.amazonaws.com/8617aa0f-456a-4d2e-b591-b0dc6c3144b7/cf41af85-4efa-426a-bf40-bb6d0a6c0a9a/fast-forward-double-right-arrows-symbol.png" width="40px" /> I want to use this Zap
</aside>
AI Agents
Real data, real ROI, all year round
The goal? Generate a continuous, detailed report of automatable tasks and provide us with a solid foundation for upselling automation services throughout the year.
What we did
- We asked our clients to install the Agent AI Chrome extension on their employees' workstations.
- The extension tracked clicks, copy-pasting, and other repetitive actions over time.
- The system sends us a notification as soon as a repetitive task has been detected.
- The system helps our customers to better monitor their processes
What we learned
- Real-time insights: The AI gave us precise, real-time data making the automation opportunities much more concrete and relevant.
- Constant upsell opportunities: With new tasks identified regularly, we could go back to clients throughout the year with fresh automation suggestions, maximizing the potential for upsells.
Benefits
- Deep task analysis: The AI captured exactly what employees were doing, from daily repetitive tasks to new workflows, offering the most detailed task analysis.
- High ROI accuracy: With real-time data, we were able to provide precise ROI calculations, showing clients exactly how much time they were wasting and how much they could save by automating.
- Continuous upsell potential: By monitoring tasks year-round, we always had new automation suggestions to present, increasing the chances of consistent upsells.
Drawbacks
- Initial setup: Installing the Chrome extension across multiple employees required some coordination and buy-in from the client.
- Privacy concerns: Some clients raised questions about monitoring their employees' activities, which we had to address by ensuring full transparency and data security.
Results
- This approach generated $40K in upsell from clients who have a real idea of how much we cost them, but above all of how much we save them in human costs thanks to detailed ROI
- Of the 20 customers to whom we sent the extension, 14 agreed to install it. On average, the extension found 4 automatable processes per customer, with details of ROI and task. We managed to upsell 10 customers, 6 of whom we re-signed over the year 2 additional dealas.
- It helped us create the automations much more easily because we already had the details of the processes and use cases.
<aside>
💸 If you're looking to drive more revenue without chasing new clients, this strategy is your answer! Here's why:
</aside>
<aside>
<img src="https://prod-files-secure.s3.us-west-2.amazonaws.com/8617aa0f-456a-4d2e-b591-b0dc6c3144b7/f4791443-1978-444e-b14c-365faa6718da/Rectangle_2.png" alt="https://prod-files-secure.s3.us-west-2.amazonaws.com/8617aa0f-456a-4d2e-b591-b0dc6c3144b7/f4791443-1978-444e-b14c-365faa6718da/Rectangle_2.png" width="40px" /> About the CEO of Mimetix.co
</aside>