Friday, September 23, 2022

Business Goal Optimized AI

A key goal of AI Engineering is to drive value from AI solutions. When Gartner says “By 2025, the 10% of enterprises who establish AI Engineering best practices will generate at least three times more value with their AI efforts than the 90% of enterprises who do not,” they mean that organizations with robust AI Engineering capabilities will have tools, processes and people empowered to meet more ambitious business goals from their AI applications.

CognitveScale’s Cortex Enterprise AI platform includes capabilities specifically designed to help AI Engineering teams to deliver business goal optimized AI solutions. 

How does CognitiveScale’s Cortex Enterprise AI platform drive business goal driven AI?

Goal-oriented AI Campaigns

Campaigns are a key component of CognitiveScale’s Cortex Enterprise AI platform. Cortex Campaigns empower businesses to make data-driven decisions that result in the best possible outcomes. Campaigns provide a framework for defining the goals, KPIs, measurements, data sources, and missions to generate executable models for tracking business goals and key performance indicators (KPIs) through continuous learning and feedback.


  • Define Goals: The CognitiveScale Cortex AI Platform clearly defines business objectives using observable data-driven KPIs from connected data, models and actions

  • Target Cohorts: Defined subsets of populations for multiple "entities" (customer, client, patient, member, agent, etc.) based on Profile-of-One
  • Create Missions: Goal-driven AI-Powered Campaigns have multiple Missions with many interventions that drive value
  • Run Simulations: Enable reinforcement learning by running simulated Missions that accelerate speed to value

  • Refine: Refine goals, cohorts, mission plans and interventions based on continuous learning and feedback

For example:

  • A health care provider organization creates a Campaign to get more patients to participate in preventative medical services.  Models identify at-risk patients, but goal-driven Campaigns impact the patient journey to meet specific goals like higher quality care and lower costs.
  • A financial services company creates a Campaign to engage with customers and deliver personalized recommendations so that they can learn about new offerings that will drive revenue (up-selling or cross-selling). Personalized customer engagement insights are delivered through client-optimized interventions: the right message through the right channel (app notifications, email, etc.).

Value from Goal Driven AI Campaigns is a Team Sport

We have written previously that AI is a team sport (here) requiring the input of numerous personas across the enterprise: Data Engineers, Data Scientists, Software Developers, Architects, Subject Matter Experts (SMEs) and Business Line Owners, and Compliance, Risk Management and Auditors. CognitiveScale’s Cortex Enterprise AI platform enables all of these personas to participate in the development of AI applications and ultimately, in the optimization of value from these solutions.

While Data Ops and Model Ops are fairly well understood to get models developed that can provide various insights (risk scores, predictions, etc.), it takes many more people to make insights actionable as part of solutions, to get applications into production, and to then build out Campaigns with Mission Plans and interventions that drive results.  Feedback loops and training data then allow application development teams to optimize the performance and value of models and solutions.

Components of a Cortex Campaign



Target of Campaign Goals. Defined using a query/filter applied to a Profile (e.g. subsets of customer types like “customers of a specific type of product”).


A Goal is the desired result of a Campaign. Goals reflect how the business objectives for a Campaign will be met using data-driven objective values that can be observed. Success of a Goal is tracked by KPIs and measures that are associated with the Goal. Each KPI targets a specific Cohort or Cohort Group.


Missions are meant to achieve a predefined goal state for a cohort or a cohort group (success criteria for an individual member of the Cohort).  This is accomplished using a set of interventions.


Interventions are actionable components designed to meet a mission goal. Mission goals ultimately contribute to the overall success of a campaign.


Simulations are a set of optimization steps that help Cortex arrive at a set of plans or groups of interventions for the target cohort. Simulation uses reinforcement learning techniques to determine the best possible intervention for a given subject within the cohort that will lead to the goal being accomplished.  Prior to the deployment of a Campaign, Simulations are run multiple times using synthetic or simulated data without needing to write lines of code. The output of a simulation run are Mission Plans (groups of Interventions) and are reviewed and refined to identify the best plan(s) that best meet the Mission Goal.

A screenshot of a computer

Description automatically generated with medium confidence


Simulations Allow you to Test and Refine you Mission Plan

Cortex Campaign Deployment

Once Missions are marked ready for deployment, Campaigns are validated for any errors and checks prior to marking it for deployment.  Deployment of Campaigns deploy all the relevant artifacts such as Agents and Skills along with executable containers to Cortex Runtime.  Deployment of Campaigns follow best practices of version control and build process using standard CI/CD process.

Mission Execution

Missions are the executable that leverages Cortex Runtime to provide the next set of interventions for cohorts. Each intervention is implemented as a Skill within Cortex.  An example of a Skill for an intervention can be a notification action to inform the customer about the nearest pharmacy to get flu shots via a mobile app. 

Customer Engagement

Interventions track the engagement of the cohort population and allows for capturing of feedback signals from real users within the cohort. These feedback signals can be tracked against each intervention. Cortex allows for a retraining of models within a Mission using a technique known as Online Learning.  Captured feedback is then processed and aggregated at Mission as well as Campaign levels totoo track and measure Campaign KPIs.

Data, Model & Business Goal Observability

Historically, Data Engineers, Data Scientists and Application Developers have focused on data and model observability, but goal-optimized AI that delivers value requires business goal observability.

Summary: CognitiveScale’s Cortex Enterprise AI Platform Enables Business Goal Optimized AI

CognitiveScale’s Cortex Enterprise AI platform empowers businesses to build and deploy goal-driven Campaigns that deliver on the promise of AI as part of robust AI Engineering capabilities. The CognitiveScale Cortex Enterprise AI platform allows users to:

  • Clearly define business objectives using observable data-driven KPIs from connected data, models and actions

  • Generate AI applications that drive value by continuously measuring and optimizing to stated business goals

  • Build reusable Campaign components like Mission Plans that help clients achieve a goal state using interventions or recommendations

  • Leverage an intelligent automation program that connects groups of cognitive interventions with bots (e.g., RPA, cognitive and chat)

  • Run intervention simulations using simulated or synthetic datasets prior to deployment to production

  • Update customer profiles in real-time and "learn" about customer preferences so that the next deployment of interventions is incrementally more effective

  • Select the best implementation strategies using feedback from intervention simulations that link AI-powered applications to business goals

    Learn More

    Contact us to explore how we can help your organization deploy and scale Business Goal Optimized solutions. 

    Email: sales@cognitivescale.com

    Phone: 1-855-505-5001



Tags for this resource
Share this page