5 ways we 10x our company with A.I. Assistants.

99% of the hype you see around A.I. is fueled by ChatGPT, but the real power for businesses comes from using OpenAI Assistants. Here is how our company 10x'd every employee.

Have you heard of ChatGPT? Of course you have. That was a test to see if you live under a rock. You passed. Here’s a bonus question. Did you know that ChatGPT was meant to just be an example of how to build apps on top of the OpenAI Model and they provide a even more powerful way to build custom integrations using OpenAI Assistants? If you didn't know, you are about to find out how we leveraged this to redesign our entire business.

The first decision we made was to become an A.I. First business. This means for any function, job or app we need we ask the question “Can AI help do this?”. Here are a few ways we executed this mission:

10X the value of Meetings

Every meeting should have A.I. attend and take notes. We started doing this in our business and it has redefined what a meeting is to our company. Before a meeting was only potentially valuable to anyone who attended. In many cases people only attended because they wanted to remain in the loop passively but didn't need to be active in the meeting. The famous “this meeting could have been an email” is more often then not. Once we launched our A.I. assistant “Prezi”, we have it attend all meetings and it summarizes each meeting, highlights key points and assigns next steps and tasks.

Once the meeting is over it sends out an email to everyone who attends and also can send to the entire company if it's relevant. People may not want to watch a 60 minutes video, but they will read a short email summary. And also, Prezi (the A.I. Assistant) can answer any questions they have about the meeting as well. It is also easy to grab past meeting tasks to reference to get updates as well.

This keeps everyone on the same page without having to read the whole book. Ever since we started using our A.I. assistant it has also made meetings more productive since we know they is a bigger impact to our discussions then just who is on the call.

How we did it:

• We used RAIA (raiabot.com) platform to easily launch and train A.I. Assistants on OpenAI

• We used TL;DR that integrates with Google Meeting.

• We used Zapier to do some light integration to send out emails.

10X your Teamwork

We Automated TL;DR (Too Long Dont Read)

One of the biggest time wasters in our business is having to summarize and share. Anytime a task is delegated it requires the transfer of knowledge so the new owner of the task k owns what happened and what needs to happen next. In most cases, the delegator may just forward a long string of emails or attached a bunch of docs. The famous “I forwarded you everything you may need, let me know if you have any questions” email should read “This is now your dumpster fire, good luck deciphering this chain of nonsense”.

So to solve this issue, we use another A.I. assistant to help summarize a series of documents or long strings of emails into a “brief” so the new owner of the task knows all the players and what is going on. Instead of the handoff taking time away from two people and potentially a lot of dropped balls and disconnect, it can be done in minutes.

Some examples where we use this the most is in support tickets, development requests, customer specs, contracts, training, internal updates and any operational policies and procedures.

How we did it:

• We used RAIA (raiabot.com) platform to easily launch and train A.I. Assistants on OpenAI

• We used RAIA to be able to email or SMS long form content to the Agent to Summarize.

• We used Zapier to do some light integration for emails and CRM integration.

10X your Support

Does your company have an Intranet? So do we. Do you use it? Neither do we. What is that? It many cases it's because it's not kept up to date, hard to navigate or just pages and pages of content that is too hard to dig thru for an answer. We had this problem, and so we launched another A.I. to answer questions anyone had internally about sales, support, operations any thing else they needed.

We trained the A.I. assistant on all the existing Internet web pages and documents and then also included ways for a time to easily “update the A.I.” on new information.

So we solved two problems. The first was one making it easier to get answers. Our people can text, email or live chat with an A.I. assistant a time they want. Second, anyone can update the A.I. Assistant with new information that may be relevant to others in the company.

Because of these two key changes, more people use it and get value. We quickly realized this could also be extended to our customers so we launched another A.I. Support Agent to handle incoming questions.

How we did it:

• We used RAIA (raiabot.com) platform to easily launch and train A.I. Assistants on OpenAI

• We used RAIA to launch a live chat agent to embed on our website and Internet.

Challenges to Overcome

With every business there are going to be challenges when deploying A.I. Here are some we faced and how we addressed each one.

1. Importance of Training. In order for your A.I. to provide value and strong responses you have to train it right. We leveraged RAIA to easily import all our documents and websites. We learned that some documents and information from different apps (CRM ,Support Tickets) may need to be cleaned up and refined before importing. RAIA also allowed us to launch a training Agent where our team could chat with the A.I. and train it real time.

2. Switching over from the “old” way can be a challenge without business buy-in. We were able to easily include our business users to engage in both training and testing (without needing a development team to assist) to make sure they gave the thumbs up before launching. This was critical in building a solution that works.

3. Fear, Uncertainty and Doubt. Anytime you deploy new technology it comes with a little F.U.D. Our team was able to overcome everyone's concerns across all teams by listing the potential issues and then building out a test plan to address each one. We took out time and tested using a small group of users and then eventually expanded to everyone.

Despite our progress so far, every day we see more potential to build more agents to automate our business. Even the smallest tasks can be removed from the day to day grind and automated with A.I.

It really begs the question: is this ultimately how A.I. will grow within organizations? The idea of having thousands of A.I. Agents running out business can be very scary (especially if your the IT department), but knowing that we are now going to benefit with every giant step A.I. takes into the future is worth it.

And if this isn't enough to get you to see the potential of A.I. - just close your eyes and imagine a world where we don't need to send each other memes about how much corporate life sucks - and continue to propagate our shared misery. Maybe, just maybe there is a gif-less future where we come together in a utopian workplace and the only thing we share is smiles and high-fives. 😊🙏