What is it FDE (Forward-Deployed Engineers)? How AI Moves From Demo to Profit

A few weeks ago I saw a teststore with an AI chatbot. It worked flawlessly: customers could ask questions which the AI answered, the AI recommended products, the AI guided the customers through the checkout process. Everything seemed to be ready for production.

It took a week for that shiny new AI powered chatbot to completely fall apart on a test store that was set up for testing. The chatbot was telling customers that new items had features that they did not, that things were in stock that were actually already sold out, and that customers could return items for reasons that were not in the return policy. In short, the AI powered chatbot was lying to customers. And to make matters worse, the system failed to properly integrate with the store’s payment processor. And when things started to fail, none of the store’s owners or employees knew how to fix it.

The gap between a demo that works and a product that makes money is huge. A lot of AI projects fail because they underestimate this gap.

There is a huge difference between a working demo and a system generating money for a business. To bridge this huge gap there’s a new role emerging: The Forward-Deployed Engineer (FDE).

Whether on-site or remote, Forward-Deployed Engineers are engineers on until their system begins to generate real revenue. Typically, Forward-Deployed Engineers are engineers first and are ‘forward deployed’ after a few months to a year as the organization discovers how hard it is to deploy AI-based systems to customers running their existing applications.

Step 1 – Find the problems for which the AI can provide a solution. Find the pain points in the workflow or in customer interactions for which an AI can provide an answer. A good FDE will understand that in many places there are work-arounds in place and that there are ‘hidden’ costs that are usually not recorded. By asking the right questions, the FDE can map out the current workflows and, based on that, identify where an AI can be of value.

This typically starts with connecting up an integration, testing it on the customers data, fixing a few of the edge cases that weren’t exposed in the demo. The best Forward-Deployed Engineers will also remain on site to be able to measure the real impact that the new system has on your sales.

Inevitably, there will come a point in time when you need to integrate your self-hosted shop (WooCommerce, OpenCart, PrestaShop or Magento) with an AI that allows customers to chat with your store in order to complete a purchase. The big question is: who can do this for you? A SaaS vendor, who will require you to rebuild your shop to suit their design specifications, or a Forward-Deployed Engineer, who will work within the existing store of a Self-Hosted Store Owner and create an integration that will work within the existing data structures and the current security model of the store?

A SaaS company would require you to switch to a store they have developed in order to integrate their AI-powered chatbot. However, a Forward-Deployed Engineer will work within your self-hosted store and try to get the AI-powered chatbot to work within your existing setup.

An FDE embeds with a business, working to set up an AI system for them within the confines of their existing business processes, store front, and data structures. Therefore, they work with a store owner’s existing store, existing payment processor, and existing database structures – all of which they know will work and those that won’t. This is particularly valuable for a self-hosted store owner, as they have a very good sense of the intricacies of their own site and made very deliberate technical decisions when setting up the store. An FDE works within the same bounds as a self-hosted store owner.

An FDE can also help identify any problems that the store owner may encounter when integrating their store with an AI service. This could include issues that are caused by the store’s payment processor, issues that are caused by a strange plugin, etc.

An FDE’s first week at a company will typically be spent learning about the inner workings of their current online store. They will need to work within the current setup and highlight where things are currently working well for the owner of the store as well as where there are areas for improvement that are being dealt with in less than ideal ways.

Week two: The FDE finds where the AI can add value to the business owner’s day-to-day. This is where the FDE can identify problems in the business owner’s process and help the business owner to implement AI to solve those problems. For example, the FDE may find that the same set of questions are asked 50 times a month. This could be a good opportunity for the business owner to train an AI to answer those questions for him/her. The FDE may also find that customers are dropping off at the checkout process or that certain products are being hidden from view and therefore unable to be purchased by customers. The FDE can then identify ways in which the business owner can use AI to solve these problems.

Week three: The integration is live and we test it. As expected, it starts to fail and we fix things up as they go. This is where we find the edge cases your demo didn’t cover.

By the end of month two, you will know if the integration and AI is increasing sales for your business. The FDE will continue to make changes and tweak the prompts and settings for AI throughout the month to get the best results for your business. At the end of the second month of work with the FDE, they will teach you and your team how to continue to make tweaks and changes to the AI as needed.

By the end of month three, you’ll probably find that you no longer need the services of a Freelance Data Expert. In all likelihood, they would have done their job so well and gone off to save another company’s bacon in the wee hours of the morning (say 2 am) during your store’s busiest sales week of the year.

Implementing AI systems to work with apps used to require a huge amount of resources and be very complex to implement, however now there are many tools available which make this very easy to set up and connect to current systems, often within 20 minutes. ChatGPT for example can be connected to any app within 20 minutes to create a new chatbot for instance. In addition to this, there are thousands of other AI system building tools available.

The hard part is everything after.

The demo of AI-powered shopping assistance was impressive in that conference room. However, there is a huge distance to the actual money making for a store. Most store owners are not aware of the huge gap between the prototype and production for an AI system. Deployment of the AI system fails and the store owner blames the AI.

They might work well on test data, but when it comes to be used on the data of a store owner’s store, it fails. They blame the AI for it not working, but the truth is that there is a huge gap between the demo in the conference room where the impressive presentation took place and a system that actually makes money for the store owner.

FDEs help fill this rather significant gap between what’s possible and what will actually work for your particular store. As I’ve stated many times before, an FDE is the hardest person to find when hiring for an e-commerce business (after software engineers, then AI experts), because on top of all that, the FDE needs to have a good understanding of customer communication (the lifeblood of a chatbot or conversational interface) and your particular e-commerce platform (most

Many developers and also so-called AI-specialists exist. But a FDE (Fullstack Deployment-Expert) with additional know-how in customer-communication, e-commerce and 24/7 faultfinding is difficult to find.

They also understand that there are many ways to solve a problem. That a simple script might be better than automating a whole process. That better product copy could solve the problem instead of implementing an FDE solution. That already documented processes could be the answer instead of having to automate.

The best FDEs are by nature skeptical of over engineering and will ask the question as to whether the simplest script is enough to solve the problem at hand or whether you are using AI to automate a process that could be done manually with a couple of scripts.

No FDE: a chatbot. Maybe clever, maybe not, but it won’t be increasing your sales.

An FDE will make an AI system for your online store (e-commerce web site) work perfectly for you. An FDE will get your AI system working for you in the best possible way with your real stock of products. An FDE will cause your online store’s AI system to handle all the questions that your customers have had about your products. An FDE will get an AI system to work perfectly for you even when things happen with a customer that an AI system was not programmed to handle and an FDE will get an AI system to keep on working for you even when you are not around (i.e. you are asleep).

The key to getting value from an AI system at a store is to get the AI system to work with the store’s real-world (offline) inventory and handle real customer inquiries and convert them into sales.

For self-hosted store owners, our FDE provides AI-driven customer engagement without them losing control of their data or having to change technology platforms.

Get started with a free 5 minute scan at dreamvention.com/scan to see the most common customer questions for your store, where your checkout fails, and which products go unmentioned.

If you’re interested in learning more about how to bridge the demo and production gap, how you can get your store’s data to work with your new AI powered chat system, and what kind of return on investment you can expect from your foray into customer engagement using AI powered chat then let’s chat. We can start here: dreamvention.com/contact.

The companies winning with AI today are winning not with the coolest technology, but by getting the best out of the technology they have by having the best out of the way deployment. Working with an FDE (or similar) is a great place to start.