Stop the AI "Money Pit" with this Quick Wins Practical AI Approach from DBMT

Get AI both right and right now by delivering in the following four areas: Quality, Value, Adoption and Satisfaction. Here's how DBMT delivers quick wins projects.

  If you are scratching your head trying to figure out when your AI investments will justify the huge investments you made, you are not alone.

The Problem. For many companies, AI initiatives were developed as a “build it and they will come model” assuming that the value the corporate AI solutions deliver will be so compelling that employees will rush to use them. But just as early initiatives in Business Intelligence, CRM, and ERP, the overly optimistic assumptions that big spend in AI would result in unquestionably compelling gains ended in a loud thud.

​

So what do you do now? It’s time to right the ship and score some big points with senior management. The key is Practical AI, delivering high quality, value, adoption and satisfaction in the short term. Getting AI both right and right now depends on delivering in the following four areas:

• Quality.  The results must meet or exceed existing corporate standards.

• Value. The results must provide tangible value to senior management that is measurable in terms of timeliness, quality and cost.

• Adoption. The capability’s use must maximize the impact of Quality and Value to the company through augmented workflows.

• Satisfaction. Stakeholders, including users and corporate internal customers (Brands) must perceive value and have positive experiences with the company’s implementation of these capabilities.

Okay, so how do we do that?  Here’s DBMT’s approach. It delivers through three key actions:

• Overcoming barriers to adoption through proactive assessment, development and communication of the vision, equipping the team with playbooks and essential tools, and identifying, celebrating and consolidating wins.

• Ensuring successful use of AI capabilities through proactive risk and capability assessment, definition of key usage scenarios and guided augmentation of key workflows.

• Quantitatively assessing the cost and value impact of AI use, both to measure overall net impact, as well as to prescribe ongoing adjustments of usage scenarios and workflows to achieve cost, value and efficiency goals.

So what are the Steps? The following is a breakdown of how DBMT gets businesses/teams to quick wins and net positive return AI initiatives.

• Strategy and Vision.   DBMT develops a Strategy & Vision presentation for the use of AI Capabilities as a proactive step to assure client stakeholders that AI will profoundly help them rather than hurt their careers.  An essential part of this step is forming a stakeholder team that contains diverse representation.  The vision will express how AI/LLM tools will transform processes and outcomes in a compelling and easily communicable way to express the 'why' behind the change.

  • Essential to this initial step is an assessment of barriers to adoption based on anecdotal and empirical evidence.

• AI Readiness and Measurement Assessment.  We begin with a review of AI/LLM tool usage, outcomes and satisfaction.  DBMT then develops an AI Adoption Framework making use of anecdotal usage data and knowledge as well as available usage tracking data from regions and areas with rich use, resulting in a comprehensive assessment of current use.

• Usage Assessment

  • Usage Data.  A review of usage data to determine how the tools are being used.  DBMT’s AI Adoption Framework will use data from rich use areas,. Usage Scenarios will be extrapolated based on usage patterns. For the ongoing process, structured data from the Impact Monitoring & Learning node will be the primary source.

  • Value Data. A review of the value provided by AI in terms of benefits like time-to-market improvements, cost savings, and improvements to workflow.

  • Satisfaction Data.  A review of satisfaction from users.  DBMT’s AI Adoption Framework will use stakeholder interviews and anecdotal communications for review and assessment.

  • Stakeholder Interviews.  Based on the Satisfaction Data, key users will be interviewed to determine specific pain points and opportunities for improvement.

• Assessment of Tools and Capabilities

  • Update Tool Details.  Research and Review of updates that might alter preferred usage scenarios.

  • Update Foundational Details.  Research and Review of updates to foundation elements (such as new/updated underlying models) that may alter preferred usage scenarios.

• Quality & Risk Assessment

  • Quality.  Determine impact of findings on the quality of output/results

  • Risk. Determine impact of findings on legal and regulatory areas

• Determine Cost/Value Metrics & Benchmarks

  • Review Current Measurement.  Examine the current performance measurement approach

  • Augment Metrics.  Define augmented Metrics that will elucidate AI Impact

  • Define Baselines.  Determine the baseline to which AI impact is measured against

• Usage Guidance Refinement.  Ensuring quality, adoption and satisfaction requires a proactive approach to guiding AI solution use.  The DBMT approach includes the following:

• Define/Revise Usage Scenarios.  Develop detailed Usage Scenarios designed to ensure teams leverage the strengths of each tool, avoid pitfalls and manage to success.  Each Usage Scenario will include limits, guidelines and output expectations. Usage Scenarios will be constructed to deliver on the typical team workflows that will support production uses of tools.

• Define/Revise Supervision and QC Requirements.  Define Oversight and QC Layers needed specifically for AI/LLM, and setup/training required.

• Define/Update Capability Decision Trees.  Define overarching guidance for expected uses mapped to specific tools.

• Develop Scenario Specifications.  Develop full scenario documentation, including illustrations, examples and outcomes.

• Develop Scenario Universe.  Create an inventory of Scenarios and Outcomes/Deliverables.

• Workflow Augmentation.  A critical pathway to adoption of an AI Capability is augmenting workflows to incorporate AI tools in existing day-to-day work.  DBMT will propose updates to Entry Points, Supervision Needs and Decision Points to ensure successful execution. This includes the following:

  • Reconcile Usage Scenarios and Workflows.  Recommend additions to workflow to incorporate usage scenarios into standard procedures.

  • Reconcile Entry points and Supervision Requirements.  Incorporate supervision and entry points into workflow.

  • Augmentation Process.

    • Assessment.  Review current workflows and workflow systems for efficiency, quality, and implementation considerations.

    • Adjustment.  Incorporate AI/LLM setup, use, QC and delivery into current ways of working.

    • Review.  Hands-on review of recommended adjusted workflows by the Stakeholder Team to inform final set of new workflows.

    • Integration.  Incorporate new workflows into workflow system, ensuring an effective testing process.

    • Testing.  Test implementation of workflows to validate success.

• Training & Rollout. To ensure success, DBMT clients get support needed to smoothly adopt the new capability into their day-to-day work. This includes:

  • AI Playbook.  A gateway document that guides users to the knowledge and resources they need to most effectively use AI.  The focus of this deliverable is both on general AI resources and corporate specific solutions.

  • Training Requirements & Guidelines.  A structure and guidelines that ensure that the prerequisite knowledge is documented, accessible and referenced in training materials.  This information is both new material and links to existing material.

  • Process & Training Rollout Support.  Assist in the presentation, refinement and adjustments made to the playbook and training resources.  This includes a rollout to a pilot audience as well as assisting in the creation of materials to roll out to a larger audience.

• Impact Monitoring & Learning.  DBMT works with our clients to define and present usage data and metrics that drive actions focused on improving process and usage that will serve as the basis for continual improvement. Here’s how we break it down:

• Metrics Design and Development.  We define and present usage data and metrics that drive actions focused on improving process and usage that will serve as the basis for continual improvement.

  • Usage Scenario Definitions

    • Conforming

    • Non-Conforming

  • Setup Level of Effort (hours, costs)

    • Operator Layer

    • Supervision Layer

  • Use Level of Effort (hours, costs)

    • Operator Layer

    • Supervision Layer

    • Comparative effort to similar outputs created without the tool.

  • Analyze Comparative time-to-market

    • Measuring time-to-market for results against similar outputs created without the tool

• Reporting and Visualization. DBMT will design and implement reporting templates and scheduled output/delivery to stakeholders. This includes the following:

  • Define Recommended Reporting Templates

  • Define Recommended Delivery Approach and Schedule

• Usage Assessment and Action.  DBMT will recommend a process through which the team leverages and acts upon findings from usage and outcomes. This gives the client:

  • Attribution of Efficiency Gains to Usage Scenarios.  Determine impact of usage on efficiency, showing champions and challengers.

  • Usage Scenario Optimization. Assess non-conforming usage scenarios and identify candidates for Usage Scenario Change Controls.

  • Training Optimization.  Determine the usage patterns that can be improved through training and prioritize training topics

• Short-Term Wins Identification & Consolidated Gains.  Here, DBMT Identifies short-term wins so that clients can call them out and celebrate with stakeholders and senior management. This is crucial for maintaining AI program momentum and motivation. These wins provide tangible evidence that the change is beneficial. As successes accumulate, it's important to consolidate these gains and use them as a foundation for further change.

So there you have it, DBMT’s step-by-step approach to quick measurable wins with AI. Use it as a guide or shoot us an email (info@dbmt.com) and we’ll set up a meeting to walk you through our approach - addressing your company specific needs and goals to make the conversation more relevant and valuable to you.