Decision Intelligence

Optimizing Outcomes with AI Powered Decisions: Webinar Q&A Part 2

Decision intelligence is about actions leading to outcomes. We highlight how to get started, the success factors, and how to calculate ROI.



In a recent webinar hosted by about "Optimizing Outcomes with Decisions Powered by AI," Ashok Kartham, Founder & CEO of, joined Dr. Lorien Pratt to answer a wide range of questions.

Here is part 2, created based on the webinar transcript. Click here for part 1. You can download the presentation, or watch the webinar recording here.    

Decision Intelligence: What role does Generative AI play in decision intelligence? 

Ashok Kartham: GenAI has a critical role to play in Decision Intelligence. If we take the decision framework, there are at least four areas where GenAI can help.  

GenAI-enabled conversational experience allows users to ask questions, seek recommendations, and access any information to support their decisions.  

Gen AI makes all knowledge, including structured and unstructured content, easily accessible to all stakeholders. The Retriever-Augmented Generation (RAG) enables users to fetch relevant information from domain-specific knowledge bases. The semantic search enabled by Gen AI provides more meaningful answers than keyword searches.

The decisions and recommendations can now be explained more descriptively using the text generation capabilities of Gen AI.  

Gen AI can also help automate actions by generating the code the AI agents need to execute the process steps in a workflow.

It is incredible what GenAI can do now with text generation and how fast the capabilities are advancing to generate pictures, audio, and video. It is also important to remember that Decision Intelligence requires other AI/ML models and optimization algorithms in addition to GenAI to improve the decision quality and optimize the outcomes.  


DI: What are some of the success factors to implement DI? 

AK: The most crucial success factor is to select the right use case with high business value and feasibility. Companies can leverage the Enterprise AI as a Service model to reduce the time, cost, and risk with Di projects. Having the right technology partner based on your industry, domain, and use cases is also critical. A phased approach to incrementally realize the value and continuously improve will help companies succeed with their Decision Intelligence initiatives.  

If you read some of the reports after Apple dropped their Car project recently, one factor for failure mentioned is that they were initially too ambitious about autonomous cars that didn't have a steering wheel or front-facing seats. After investing billions of dollars, even the best company in the world couldn't realize its objective. In comparison, Tesla delivered autonomous cars years ago and has kept improving towards full self-drive. Even if FSD was delayed by a few years, the company and drivers have been realizing the value of autonomous cars.


We can learn from this rare failure to start the DI objectives at a feasible level and then have a business model to continuously improve and build on success.  

 How to calculate and realize ROI from Decision Intelligence?   

AK: Decision Intelligence creates value through increased revenue, cost reduction, improved efficiency, and risk mitigation.  

Data-driven decisions during marketing, sales, and service stages help companies grow revenues by increasing customer conversion and enhancing customer experience.  

Decision Intelligence helps businesses save costs by optimizing operations, improving supply chain efficiency, reducing inventory levels, and allocating resources more efficiently.  

Companies can also improve productivity and efficiency by leveraging decision intelligence to augment and automate time-consuming decisions.  

They can mitigate risk through predictive analytics and improved decision-making processes.

In addition to these quantitative returns, Decision Intelligence provides many qualitative benefits such as improved employee satisfaction, better brand reputation, and faster adoption to changing business conditions.  

DI: How can companies get started with DI solutions?

AK: The best place to start with a Decision Intelligence solution is by defining and modeling an impactful decision. You can then identify the data and knowledge needed to support the decision. Once the use case and required data are identified, you can implement the solution using the right decision intelligence platform and application that supports that use case well.  

A successful Proof of Concept (POC) or a pilot project can validate the decision model and the outcomes. You can also gain stakeholder acceptance by approaching the DI solution in three incremental stages: analysis, augmentation, and automation of various decisions.  

DI: Companies have the decision logic embedded in their current systems. How can they apply AI to these applications?

AK: It is better to augment these applications first with DI instead of ripping and replacing them, which can be costly. You can infuse intelligence into the existing processes by moving the critical decisions into the DI layer. The current process may have implemented decision logic using business rules, workflows, and business process management. DI can provide real-time decision-making through API, augment and automate select steps in the workflow, and generate actionable insights using predictive analytics.  

Having the decisions in the DI layer will also improve transparency and reuse across various business processes.  

 What are the risks with DI, and how can companies mitigate them? 

AK: Despite tremendous economic potential, industry studies show that 85% of AI projects fail to go beyond the POC stage. Key contributing factors to the success of DI projects are realizing the expected business value, availability, and quality of the data, having the team with the required skills, selecting the right technology platform and partners, stakeholder acceptance, setting reasonable expectations about AI capabilities, and optimizing the operation of an AI system in production.

risk mitigation in decision intelligence

By taking a decision-centric approach, you can focus on the specific data required to support the decision, which could be only 10% of the data. In contrast, a data-centric approach to consolidate and cleanse all the data would be time-consuming and costly.  

Companies can address project execution risk by leveraging the "Enterprise AI as a Service" (AIaasS) model, the Decision Intelligence platform, and relevant intelligent business applications. This approach will reduce the upfront investment and skills required to deploy and sustain the DI applications.

 How can companies monitor decisions to continue to optimize the outcomes?  What metrics can companies use to assess the effectiveness or quality of decisions? 

AK: To continuously optimize outcomes, companies can implement relevant Key Performance Indicators (KPIs), leverage Decision Intelligence (DI) for ongoing monitoring, gather user feedback for decision refinement, and apply AI for optimizing decisions.  

One of the primary advantages of decision intelligence is its ability to track the results of decisions. Continuous monitoring enables companies to evaluate the impact of their decisions and adjust strategies as necessary to optimize outcomes.

To assess the quality and effectiveness of decisions, companies can use metrics such as performance against KPIs, user satisfaction, cost-benefit analysis, decision cycle time, and adaptability to changing conditions.  

Decision intelligence is not a one-time implementation. The real benefit lies in its ability to foster ongoing learning and improvement. By continuously analyzing outcomes and adjusting strategies, companies can enhance the quality and effectiveness of their decisions over time.

 How do you see decision intelligence changing in the next few years?

AK: In the coming years, Decision Intelligence (DI) will help accelerate the journey towards fully autonomous business operations. In the long term, DI will transform systems to make complex decisions with minimal human intervention.

With the rise of Edge AI, DI will also increasingly operate at the edge, closer to the decision point or where data is generated. This shift will enable faster, more efficient decision-making processes across various locations and contexts.

While Generative AI (GenAI) currently excels in creating human-like text, audio, and video, its integration with DI will extend its capabilities to decision-making. Future DI systems will harness GenAI, alongside other AI/ML technologies, to generate decisions mirroring human judgment and reasoning.

Achieving broader adoption of DI involves enhancing the intelligence of systems, processes, and workflows. I expect DI to become as widespread and integral to business operations as Software as a Service (SaaS) and Business Intelligence (BI) have been over the past decade.

I believe these developments will lead to a future where DI becomes deeply integrated into the fabric of business operations, driving innovation, efficiency, and more autonomous decision-making across industries.

Click to access the on-demand webinar


Similar posts

Get notified on new Decision Intelligence insights

Be the first to know about new decision intelligence insights to understand and engineer how AI-powered decisions are made and how outcomes are evaluated, managed, and improved.