Lucas Persona is Chief Digital Officer at CI&T, a digital transformation partner for global brands.
In my previous article, “Agentic, Generative and Predictive: How the AI Orchestra Works in Harmony,” I looked into the roles different AI capabilities can perform to harmonize customer experiences and generate business results. Now, we turn theory into practice. How do you contextualize this AI orchestra? How is agentic AI integrated into an existing landscape of AI solutions?
As I explained in my previous article, “Agentic AI makes use of the reasoning ability of generative AI to allow a solution that, instead of being programmed for a specific task, can evaluate, plan and execute tasks with the tools they have at their disposal to achieve their goal. The value of agentic AI is best utilized with larger, more complex tasks that require a sequence of different smaller tasks and the possibility to execute those tasks in various orders and scenarios.”
Success requires the strategic alignment of business objectives, user experiences, data architectures, ethics, team capabilities and governance models. Let us move beyond AI theory and possible solutions toward how to actually make it happen while delighting customers and engaging employees.
1. Define use cases as business outcomes, not features.
Begin by framing the use cases as business objectives that connect customer or employee experiences with operations. This shifts the conversation from “add generative AI to product search” to “increase the conversion rate of a shopping session by 10%,” and from “generate warehouse floor plan” to “enable same-day fulfillment on high-margin categories while maintaining the same staffing levels.”
Well-defined goals are critical when incorporating agentic AI because it uses the tools available in order to plan the best way to achieve those goals.
2. Create user journeys that incorporate humanized interactions.
The first impulse when trying to humanize the interaction is to turn chatbots into the main and only point of contact with customers. Look beyond that. Hybrid interfaces that provide users a mix of doing-it-themselves (DIY) and intent-based guidance—do-it-for-me (DIFM)—maximize the potential of outcomes.
This means incorporating additional ways of interacting beyond point-and-click, such as text, voice, images and documents, and designing intent/knowledge-harvesting journeys that can capture feedback, additional context and decision-making rationales that can be leveraged later.
3. Rethink your data and AI strategy with a new perspective.
Delivering those hybrid, new user journeys will require going beyond relying solely on historical transactions and unstructured and unmanaged data. An updated data and AI strategy leverages both data and the additional knowledge required to achieve business outcomes.
Seek out tacit knowledge that is not documented anywhere or is sparse but will be key in driving successful autonomous decisions.
4. Leverage the best tools and technologies for each task.
Work with technical teams to explore the distinctions between tools and agents; deterministic and probabilistic outcomes; and optimization, prediction and generative AI.
It’s natural for teams to try to apply new technologies in scenarios where they may not be well-suited, like using an LLM, which is non-deterministic, to perform a well-defined task that requires deterministic behavior. Selecting the right technology for each part of the solution can significantly reduce effort during the validation phase.
5. Embed ethical frameworks.
As agentic solutions are given greater autonomy, it’s necessary to incorporate ethical considerations from the start: in the selection of outcomes, initial solution design, guardrails implementation and accountability mechanisms.
The independence of agentic solutions does not imply that teams, leadership or the organization are not accountable for the decisions, actions and outcomes of those agents.
6. Focus on creating an AI-first culture instead of procuring the “best AI tool.”
The adoption of agentic AI requires a planned approach to education/literacy, experimentation, process redesign, role changes and skills development or upskilling.
Teams that incorporate an AI-first mindset will accelerate the adoption and success of agentic AI in their day-to-day. On the other hand, organizations that look at agentic AI as an external tool to adopt and hand over tasks to will quickly find limited effectiveness in their adoption.
7. Seek business buy-in on autonomy.
Organizations that had great success in implementing a data-driven culture are more likely to accelerate the adoption of agentic AI with increased autonomy.
As the reasoning and decision-making of agentic AI are strongly influenced by data, organizations that have resisted this cultural change will face stronger trust challenges in yielding autonomy to agentic AI. One option is to look for areas within the organization that have already embraced a data-driven culture to be the early adopters of agentic AI.
By approaching agentic AI implementation with these seven strategic considerations, organizations can move beyond experimental deployments to create sustainable, valuable AI-powered solutions that drive meaningful business outcomes.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
