The agent's AI recipe: Discover, Decide, Deliver
The Discover, Decide, Deliver framework is a useful model to frame how business leaders should think about their investment and approach to adopting AI and agent's work flows.
Let's explore the process of bringing a revolutionary medicine to market with agents using Discover, Decide, Deliver. Traditional drug development is a long and costly journey, often taking several years. agent's AI can automate repetitive manual steps and create more seamless transitions between stages, accelerating time to market.
The process begins with Discovery. This is where initial research takes place to understand the disease, decide which aspect of disease biology to target, and then develop a compound that will alter that aspect of the disease. It also includes data analysis to test molecules and compounds. Next is the
Decision phase, where researchers and clinicians land on how and where the drug will be used, from speeding up new clinical trial protocols to analyzing massive amounts of complex data from these trials. Then there's the Delivery phase, which involves producing the drug, educating patients and doctors, and bringing it to market.
While discovery is happening faster than ever, manual processes between and within stages can cause delays, errors and miscommunications, slowing down overall progress. In this case, speed to value isn't just a metric – its the difference between getting a drug to market faster or losing market share. For instance, one of the world's largest pharmaceutical manufacturers, Novartis, applied Generative AI in its drug development process even before agents came along. Imagine what they could do now. Ideally, agents can cut time lines from decades to years and years to months, enabling faster decisions, delivery and learning. Here, agents don't just help you do more with less – they accelerate value by getting you to the right outcomes faster.
This concept applies across industries. For example, Levi Strauss & Co built a powerful but complex traditional AI system to detect fashion trends. Agents could optimize this process by analyzing online images daily, tracking the shift from fitted to baggy jeans across demographics and delivering insights faster. The Coca-Cola Company is taking it a step further by using AI to enhance its relationships with retailers. Imagine an agent that gives recommendations on how to arrange products in coolers based on current consumer preferences. In these cases, speed to value is achieved by 1) accelerating the transition from analysis to action and 2) reducing the delay between insights and action across companies.
Leadership considerations for the agent's age :
In a recent article, we explored the evolving role of leadership in the age of AI. The rise of agents requires a similar mindset shift. Below, we apply the Discover, Decide, Deliver framework to help leaders ask the right questions in an agent-driven world.
1. Discover: Getting to better answers, faster
The key to leveraging agents in discovery is understanding your objective and using the right data to fuel faster insights and analysis from more places. Agents can quickly pull data from multi modal sources – text, audio, visual – and spot patterns that reveal opportunities. Using reasoning capabilities and accelerated computing, they help connect the dots faster, turning data into answers.
Leaders need to ask:
What sources and types of data could uncover untapped opportunities?
How do we redesign work flows to let agents drive insights?
Where should human judgement come in to ensure the business objective is met or course corrected?
2. Decide: Getting to actions faster
Across industries, leading companies like Netflix have built their success on high-velocity decision making. They thrive because they minimize the iteration cycle and execute quickly to deliver innovation. Agents that can reason allow any organization to streamline decision making, reduce the cost of experimentation and take action faster. With agents handling manual, repetitive, error prone tasks, leaders can focus on more strategic decisions sooner, accelerating the innovation cycle.
3. Deliver: Getting to outcomes faster
Today, the speed of delivery depends on how well you collaborate with customers, vendors and partners. Agents enable seamless collaboration across organizations, speeding up delivery by ensuring that processes are in sync between teams, suppliers and customers (just like APIs did for digital transformation).
Leaders need to ask:
How can we collaborate with customers and partners to create agent-driven work flows that speed up delivery and unlock new possibilities?
What do we need to understand about agent-to-agent collaboration across organizations versus within our own?
How can looking at interactions from human-human, human-agent, and agent-agent perspectives help us optimize the customer experience?
Equip yourself for an agent's future
Agentic AI is changing how businesses drive outcomes by accelerating speed to value. The companies that will dominate the next decade won’t just use AI – they’ll fully integrate agent-driven actions across their business and partnerships. To capture the benefits of speed to value, this will need to be both a process change and a mindset change across leadership and throughout the organization.
How to get started:
Align agent's work flows with operational capabilities.your abilitiesYour digital work flows have to be in sync with your physical delivery model. If there is a disconnect between how work is processed digitally and how it is executed in the real world, you risk losing the benefit of speed to value. Regular alignment checks can keep both elements functioning together efficiently.
Define ideal outputs and outcomes.Since the full potential of agentic workflows is still unfolding, set ambitious goals. Start by defining the ideal outputs and outcomes, then break them into small, manageable steps. This approach will allow quick iteration while making tangible progress toward your long term vision.
Identify human-in-the-loop insertion points.agent's work flows are not a “set it and forget it” solution. There will be inflection points that require human oversight to ensure quality and correctness. These human-in-the-loop points are crucial for intervention and decision making. If these checkpoints are missed (or misaligned) early on, small missteps can become bigger issues.
Set up your agent's feedback loop.agent's AI improves through a continuous feedback loop, where data from each interaction is used to refine models and enhance agent performance. This ability to adapt and apply fresh knowledge helps companies become more effective over time, getting to value sooner with each iteration. Set guidelines to distinguish between feedback that drives value and feedback that simply provides knowledge and ensure that agents effectively integrate feedback to drive ongoing innovation.
Success comes down to implementing a scalable and repeatable framework that works within your existing processes. Discover, Decide, Deliver is one example that will help you break through the complexity and unlock the promise of agents.
You must be logged in to post a comment.