GenAI Use Cases: Where to Start?

One of the frequent topics of discussions we are having with AuxoAI clients are "GenAI sounds great. But how do we think of the use cases? Where do we start?

And underlying this question are three implicit considerations:

1) Will this technology work as promised? Should I put my neck out for this?

2) Are their enough believers in the company? How much work needs to be done to evangelize?

3) Who is going to be the champion within the enterprise that is ready to experiment?

Depending on the responses to the above, two paths emerge:

Path 1: Incremental Value Mining Approach:

Mostly observed in companies where either top down leadership buy in does not yet exist. There are other strategic priorities that take precedence. They see AI as important but not something that will drive structural competitive advantage or disadvantage. At least not yet! However, They want to experiment with GenAI and continue learning:

The use cases they pick are -

- Internal and lower risk areas

- focused on efficiency gains, treating the effort as automation ++

- mostly pointed use cases with less upstream and downstream impacts

examples include: various kinds of internal chatbots (HR, product support), basic templatized content generation like emails/general presentations/ marketing materials

Path 2: Transformative Approach:

Mostly observed in companies where there is a top down belief that GenAI is a big deal to their competitive position. Its a huge opportunity or risk. CEO is normally the champion and asking his/her team to push the agenda. They want to learn and do it fast and iterate.

And most of them do not think use cases. They think workflows:

- The focus us on metrics. They ask "what is the metric that I want to impact and by how much?" - new hire training time, first call resolution rate, lead to close rate etc.

- the workflows are in critical business functions (R&D, Sales, service delivery)

- Cross functional teams are set up that are empowered to make decisions and change processes

- What process and talent changes need to happen is a question even from the POC phase (they are preparing to make it work vs. wait and watch)

It is early days for Enterprise AI at Scale. More clients choose Path 1 over Path 2 for now. And good news is that we are starting to see the confidence gained from Path 1 approach being translated into conversations to Path 2.

And as we see more Path 2 choices being made, there will be a convergence of digital and AI transformations.

Satish Tammineni and Amaresh Tripathy collaborated on this post to summarize the AuxoAI experience and learnings over last year.