What Is This Episode About?
Welcome to AI and Automation in Action, where we show business leaders how to turn AI from a buzzword into measurable operational results. If you're serious about operational efficiency and responsible AI adoption, you're in the right place.
What Problem Was the Client Trying to Solve?
The team has been working with an electronics recycling company through one of their automation identification offerings. The business receives spreadsheets from customers listing hardware they're turning in, including make, model, device type, and specifications. Because pricing varies depending on specs, the company needed an accurate, item-by-item valuation before they could make a competitive offer.
Their old process involved a staff member manually searching each item on eBay, reviewing sales history, checking current pricing, and then compiling that research into an internal analysis. That manual workflow was consuming roughly 9 to 10 hours every week.
How Did the Automation Solution Work?
The team worked directly with stakeholders to fully map out the process. During that discovery phase, the client provided an internal electronics grading guide, which defines item condition categories similar to eBay's own classifications (new, unboxed, and so on). That grading logic was built directly into the platform.
Two API keys were created: one for eBay and one for ChatGPT, the AI model the client had already standardized on internally. Claude or Gemini could have been used instead, but the client's preference was honored. Those keys were integrated into a custom platform built for this workflow.
Now the client creates a bid list inside the platform. Items can be added manually for small batches, or the customer's spreadsheet can be imported directly. Once the list is loaded, the platform pulls historic pricing, current pricing, sales volume, and recent sales activity from eBay for every item. After that data is captured, the user triggers an AI analysis that applies the electronics grading guide and produces recommended next steps for each item.
How Much Time Did This Save?
The solution saves the client approximately 9 to 10 hours per week, all from a single automated process. The staff member who previously spent that time searching eBay, taking notes, and writing up a manual analysis now simply imports the equipment list and lets the platform run.
A potential future addition is an automatic posting step, where items that come in could be listed directly to eBay from within the platform, though that has not been formally requested yet.
What Other Industries Could Use This Type of Automation?
While the electronics recycling use case is very specific, the underlying pattern applies broadly. Any business that receives a list of items or equipment, manually researches pricing or data across multiple sources, and then compiles that research into a report or analysis could benefit from a similar approach. Examples include companies bidding on projects or RFPs that involve equipment lists. The key factor that made this project straightforward was that the client had already clearly defined the process, knew who was doing the work, and knew exactly how long it was taking.
The automation pulled listing data directly from eBay's API, but the conversation quickly turned to a broader point: API access alone is not enough. What matters is whether the specific endpoints you need are available. An endpoint is essentially a module or section within a tool, and even if a platform advertises API access, that access may only cover a limited set of those modules. If the endpoints required for your automation are not included, you either need to contact the platform's support team to get them enabled, or pay an additional fee to unlock that functionality.
A practical example: an API might return customer records but not customer orders, because orders sit under a separate endpoint. Beyond that, there are often rate limits to consider, controlling how frequently you can make requests and how much data you can pull at once. These constraints have to be factored into the design of any automation from the start.
What other business processes could this type of automation apply to?
The same approach used for eBay research maps directly to internal purchasing workflows. Rather than manually logging into five different distributor portals to look up a price, an automation could query all of them at once and return a comparison in real time. The AI layer could then factor in shipping costs and processing fees to surface the genuinely cheapest option, not just the one with the lowest sticker price. That removes a meaningful amount of labor time from a task that would otherwise require someone to do it by hand across every tool.
Who should be identifying automation opportunities inside a company?
This particular workflow was not flagged by management. It came out of the Executive Jump Start engagement, surfaced through a Microsoft Form sent to staff across the sales department. That detail reinforced a pattern that had become clear over time: leadership and ownership tend to be aware of a limited set of processes, but the highest-value automation opportunities are almost always identified by the people doing the work day in and day out. If a company is serious about finding ROI in AI and automation, it has to actively engage frontline employees and give them a structured way to surface what they are doing manually.
Another major benefit that surfaces during this process is internal documentation. Many processes are only living in one team member's head. Once uncovered, teams can actually step through, refine, and document those processes themselves before handing them over for automation. As Hunter and Mike both noted, this is a real vulnerability for most businesses. When the person who holds that knowledge retires, leaves, or is out sick, there is suddenly a gap with no written process to fall back on. The audit solves two problems at once: it creates the documentation, and then it automates the process so employees get time back to focus on higher-value work.
How Long Did It Take to Build an Automation That Saves 9 to 10 Hours a Week?
Hunter confirmed the platform is not fully complete yet, but the work so far has taken around 6 hours. That means the build time is already less than a single week of doing the process manually, and it delivers an automated solution that any number of employees can use, not just one. The time for the task itself drops from 9 to 10 hours down to minutes.
What made the result even more striking is that when the project was originally scoped, it was estimated to take around two to two and a half months. It ended up being delivered in under a day of actual work. Mike credited Hunter's efficiency and creativity, and noted that this kind of outcome is exactly what their Virtual AI Officer program is designed to produce: a prioritized list of automations ranked by ROI, starting with high-impact, lower-effort items before moving into more complex builds.
How Does Microsoft 365 Integration Factor Into This Automation?
Hunter also highlighted that the platform is integrated with the client's Microsoft 365 environment via single sign-on. To give an employee access, the client simply adds them to a group in Office 365. From there, all of their existing conditional access policies, password reset rules, and security configurations are automatically enforced within the new platform. No separate user management is needed.
What Are the Next Steps for Expanding This Automation?
Mike outlined several natural extensions of what has already been built. Because the client uses standard templates, the platform could be connected to automatically generate documents in Word or Excel. Microsoft Teams notifications could also be added. The broader principle, as both Mike and Hunter discussed, is that automations are building blocks. You build the first one, then stack additional capabilities on top, and reuse what is already in place. With this client's full list of planned automations still ahead, and the highest-ROI item already delivered well under budget, the team is optimistic about how much tangible value can be returned to the client going forward.
It really comes down to putting numbers behind who is doing a task, how often it is being done, and how much time it takes. You get two things out of that exercise. You get the documentation of the process, and you get the opportunity to ask whether you will see a return on the time and money it takes to build the solution. Having a conversation with Hunter, he will help you find the highest-value opportunities and likely come in under budget. Kudos to Hunter for walking us through all of that today.
Where Can You Continue the Conversation?
Thanks for tuning in to AI and Automation in Action. If today's conversation sparked ideas about how AI or automation could improve your operations, we would love to continue that discussion with you. Join us next time as we showcase another real-world implementation and break down how to deploy AI responsibly and effectively. If you want to brainstorm with our team, email us anytime at ideas@innovativeautomations.ai. Until next time.