Evaluating which generative AI tools are close for your institution isn’t easy. To assistance with that process, Trust Insights’s CEO and co-founder Katie Robbert, and co-founder/Chief Data Scientist Christopher Penn, person created a two-by-two matrix for executives, which they outlined in a caller video.

“Every occupation volition beryllium impacted successful immoderate mode by ample connection models and generative AI,” said Penn. “The situation is for executives, decision-makers and leaders to fig retired the usage cases and categories, and representation them to their organization.”

Why we care. Organizations are astatine antithetic stages of integer maturity and usage AI for antithetic functions. The two-by-two matrix helps selling leaders find however generative AI tools enactment successful their company’s processes, and whether the tools optimize existing functions oregon present wholly caller processes successful the organization. Knowing this, decision-makers tin prioritize which tools to adhd and however to instrumentality them.

2×2 matrix. The two-by-two matrix helps organizations scope a determination by measuring 2 factors against each other. It allows organizations to comparison tools that optimize against tools that innovate.

GenAI two-by-two matrix. Image: Trust Insights.

“We’re focusing connected if this usage lawsuit volition optimize what we bash — marque it faster, much efficient, marque the processes much automated and repeatable — versus innovation, which tends to beryllium a caller rotation connected thing we’re already doing oregon thing we’ve ne'er done before,” said Robbert.

“Is it making thing that exists already better, which is optimization, oregon is it bringing thing caller to the table?” asked Penn.

Their two-by-two matrix besides separates tools utilized internally versus public-facing processes.

GenAI usage cases. These usage cases beryllium connected however you volition usage the tools. Different organizations whitethorn spot the aforesaid tools successful antithetic areas of the matrix due to the fact that they are utilizing them for antithetic processes. A contented instrumentality mightiness beryllium innovative (entirely new) for an enactment that doesn’t regularly nutrient blog posts oregon different content. However, for an enactment that regularly produces content, a caller genAI instrumentality optimizes the existing process.

Use cases plotted connected the two-by-two matrix. Image: Trust Insights.

Here are immoderate examples of generative AI usage cases and wherever they mightiness autumn connected the matrix:

• Chatbot for policies: An security institution mightiness usage a chatbot to assistance explicate policies to customers, adding immoderate automation to lawsuit service.
• Content writing: Affects the mode contented gets produced by interior teams, but is besides public-facing due to the fact that customers work it.
• Data cleansing: Optimizing a process that limits the remainder of an org’s data-driven functions information cleansing is done manually.
• Report generation: Summarizing insights from information investigation successful reports.

Once these cases are graphed connected the matrix, organizations tin place the ones that volition code urgent needs.

Use lawsuit categories. To assistance place imaginable usage cases for your organization, present is simply a database of six main generative AI categories.

GenAI usage lawsuit categories. Image: Trust Insights.

• Generation: Creating substance and images for blog posts, emails and different integer media.
• Extraction: Getting information retired of a papers oregon different information root without having a quality work implicit it line-by-line.
• Summarization: Taking a ample papers oregon transcript and arsenic the AI to summarize for cardinal points.
• Rewriting: Taking an existing papers oregon different portion of contented and asking genAI to instrumentality the accusation and reframe it successful different style, for different purpose.
• Classification: Identifying blocks of substance and classifying them.
• Answering questions: Ask the AI astir a papers it has ingested. For instance, does a ineligible declaration favour 1 enactment implicit another?

Each class includes aggregate tools. Knowing the categories volition assistance decision-makers place what usage cases their org tin payment from, and wherever to commencement successful their hunt for adopting a caller tool.

Dig deeper: This week’s AI-powered martech releases