Satisfaction Optimizer

A Smart AI Assistant for IT Pros to make them look like heroes.

What I have done with this project is bridge SpiceWorks with an Analytics Databases, then design a machine learning structure to learn from all the interactions pouring through SpiceWorks. The sole purpose is it to make the IT Pro's life easier, while maximizing the end-user's satisfaction (in a measurable way). Here are some of the features:

  1. The AI can handle requests and problems automatically. The AI combines machine learning with Expert System AI via the knowledge base, computing scores of tuples of issues reported by users, with the score of user ratings, to find the best solutions to issues. This will allow the AI to get smarter as time goes on, and become more expert in solving problems. Note that this doesn't have to necessarily be pre-programmed. The system is designed to learn as you go. So it learns to recognize key phrases and connect them to solutions. I would like to work with SpiceWorks to data-mine your built-in knowledge base, and match it to issues in a machine learning system.

  2. The AI can assist the IT Pro to handle problems semi-automatically. The AI can make suggestions, and actually compose and write comments to help end-users, then offer it as a suggestion to the IT Pro.

For example: An end-user complains their "printer isn't printing".

The AI can ask the end-user, "Did you see any error messages or codes?"

When the end-user responds, for instance "I saw error code 13.1C", the AI knows what this error code belongs to, and will compose a message for the IT Pro to use: "There is a jam inside the fuser. Don't try to clear it. I'll do it, so it doesn't damage the printer."

The AI will offer that message to the IT Pro, then if the IT Pro approves, it will be sent to the end-user. So what's happened is, the AI has reduced the diagnosis process by asking all the right questions (all without the IT Pro lifting a finger).

  1. User Feedback/Ratings Feed The AI With Knowledge Every closed ticket automatically gets a Satisfaction Survey. The results get stored in the database, and help the AI learn what makes end-users more satisfied. This helps it learn how to solve problems satisfactorily faster.

This system demonstrates two approaches to automated problem solving. One is to have the AI automatically ask the user for more probing information to identify the issue, and the other approach is to have the IT Pro confirm solutions and suggestions before sending them. With the second approach, the IT Pro will also be training the machine learning component of the AI with their choices via GuidedSuggestions.

  1. Tracking & Predicting The system tracks all tickets and inventory events in an analysis database. Then the AI can show trends, and times when issues tend to get worse. For instance in the chart screenshot, you can see that printer issues peak in January and really heavy in March. This might be because of year end reports, and maybe tax season. By exposing these patterns, the IT Pro can prepare the office and staff for those heavy times (refresh toner cartridges, schedule maintenance just before those heavy times, etc.)

  2. Emergency SMS Bridging System End-users need to feel that IT Staff is looking after them. By bridging SpiceWorks tickets through Text Messaging, users with urgent needs FEEL more taken care of when they can text the IT Staff to feel that they aren't alone. If the end user has an urgent issue, it can be routed to the IT Pro that is on call. Just have the end user request SMS help, and the AI will then bridge with the IT Pros cell phone via SMS. There is smart routing, so the on-call IT Pro can handle multiple tickets at once, and never reveal his actual cell number. Two way privacy so no one knows the others cell number. All interactions are recorded and tracked fully in their appropriate SpiceWorks tickets.

  3. Customer Service --> Social Media Bridging --> Social Media Satisfaction As I explored Spiceworks, the ticket and inventory system is so powerful, we decided to use it as our main customer service portal (instead of Zendesk). It's much cleaner and more powerful. I feel Spiceworks should reach out and market to the customer service and brand management space. To further that, I also included Social Media integration. This way any complaints or issues about the brand or company posted to social media, instantly and automatically become SpiceWorks tickets, to be handled, and resolved.

If an issue pops up over social media, the AI will spot it, and automatically enter it into the system as a new SpiceWorks ticket with full tracking back to the social media where it appeared. As the IT Pro responds to the ticket, the AI knows contextually that this belongs in social media, and response back to the social media (posts response to twitter, etc.) This alone, combined with knowledge-bases can be a product all by itself, and SpiceWorks provides a powerful solution to a significant problem for companies to manage.

  1. Gamification of Satisfaction All the end-user/customer ratings are tracked in an analytics database. So the system can show the IT Pro/CustomerService what their average rating is, and more importantly the AI can learn what generates more satisfaction. Businesses optimize everything nowadays. Optimize conversion rates, optimize shopping cart orders, optimize web design. Finally there is a system to optimize customer/end-user happiness.

  2. Testing & Optimizing Responses For Satisfaction via A/B Testing More than one response can be entered as a response to an issue. So the system can do A/B Testing to see which responses generate the highest satisfaction scores. If SpiceWorks does expand to the customer service space, then this becomes even more powerful.

Consider this. Imagine if SpiceWorks did expand into the customer service space. Imagine if SpiceWorks could solve satisfaction problems for businesses and their customers. For instance, if a customer is requesting a refund, even if in the end, you refund the customer, the wording you use may have that customer leave being happy, and or possibly lead to additional sales. You could say, we are sorry to see you go, here is your refund. But if you want to try this other product instead which might serve you better, I've included a 50% off coupon code, and here is a link to that product. By SpiceWorks being able to A/B test, and learn what works, the possibilities are endless.

Share this project: