Which AI initiatives have delivered the most measurable business value in the past 12-24 months, and how did you measure that value?

IT Strategy & RoadmapMachine Learning (ML)
1.3k viewscircle icon3 Comments
Sort by:
Information Technology Recruiter in IT Services20 days ago

Customer Support Automation

Lightbulb on1
22 days ago

We have recently implemented ChatGPT Enterprise for the firm. This has given us the ability to start collecting better information about usage and what people are asking. It allows us to build a repository or capability around helping people get from point A to point B. All of the basic use cases—summarization, drafting, getting smart on a particular topic—have been good use cases for us. Now, we want to see how much people are making use of these tools and what we can start to replicate across the business.

CIO23 days ago

We have had good success on a few fronts. One of the first use cases we introduced AI for was in our security tools, such as buying platforms that already had AI built into them—Abnormal Security, Insider, and several other platforms. These tools make it much less manual to inspect patterns and detect phishing attempts, tag emails that are high risk, or catch exfiltration of data when people leave. That was our initial introduction to AI as a company, and it was very successful. We have improved our security posture significantly as a result.
In addition, like many other companies, we have focused on productivity and generative AI use cases. For example, recording meetings and generating voiceovers for training videos: you provide a script to our AI tool, and it generates the audio based on samples from our training staff, creating full-length training videos for internal use. These initiatives, which we started about 12 to 18 months ago, have delivered very high ROI. We have been able to save money while delivering high-value content.
Another significant area has been automation. We have leveraged AI for AP automation, such as reading PO numbers and processing invoices without being fixed to a particular format. Using AI as a supplemental tool for some of our automation techniques has yielded measurable results, including a reduction in error counts and a decrease in manual hours spent on processing. For content creation and training, the reduction in manual labor has also been significant.

Content you might like

What UX elements of AI-driven web messaging do you find most useful for customer service use cases?

GIFs7%

Option buttons30%

Carousels & swiping16%

Videos11%

Images22%

Maps3%

Text5%

Date & time pickers1%

List pickers1%

View Results

Were you blind sided by any unexpected costs related to AI in the first half of this year? If you could go back to the start of the year, what would you do differently?

Choose your preferable search engine and mention the reason below:

Solr15%

Elasticsearch66%

Algolia5%

Other (Please specify)12%

View Results

As part of our NYSE IPO prep, we’re debating how to communicate our system hardening efforts in regulatory disclosures (e.g., SEC Form 20-F, SFC).

Would you recommend sharing % compliance (e.g., “85% CIS Tier 2”) or sticking to qualitative descriptions of how we identify and mitigate risks? Also, do SFC/ISO 27001 expectations require full ISMS integration, or is a % model acceptable?

We have 1000s of forms. Is anyone else using Azure AI to classify and report on large collections of forms by type and version? I’d like to connect with others doing this.