I've been an active user of ChatGPT, and it's been a huge help for coding, writing documents, emails, and Jira stories. Unfortunately, I can't use Copilot in my organization yet, as the legal team is still discussing the potential risks.
I'm eager to take things to the next level and initiate projects that can make a real impact on my business unit teams. So, I'm currently exploring ways to utilize Gen AI to enhance the cost-effectiveness of processes and tooling across my business unit.
I'm curious to learn from others about their internal initiatives using Gen AI that have had a significant impact in your organisation. Any insights or examples from your team would be greatly appreciated!
My prior place had a huge number of the staff (10k+/-) who were dedicated to customer support type roles in a call centre like environment.
I did make an MVP that was meant to be a digital customer in order to upskill new staff which was powered by a mixture of 11labs, whisper & gpt 3.5. This product was amazing to use but unfortunately didn't get any proper exec stakeholder buy-in so was left in the cupboard.
In my new place I have witnessed the alpha launch of an internal product used to summarise 3-5hr long meetings for executives to gain a brief understanding of what was potentially applicable specifically to their division.
I also use Claude to ask specific API/unfamiliar framework questions. (Claude isn't blocked 🥸)
Another way to get inspiration is to look through recent YC companies: https://www.ycombinator.com/companies?batch=W24.
Many of these are AI companies, and clearly they had enough conviction (and in some cases, enough traction) to build a startup around a problem with a genAI application. A few patterns I observe:
Some of these can likely be used internally with your business unit teams.
On Taro, we have TaroGPT, which provides the search functionality on our site. Unfortunately, usage of this has plateaued. It does add a lot of value for a core group of users, but it's not the mechanism for growth that we once thought it'd be.