Map the processes first, then bring in the AI agentsMap the processes first, then bring in the AI agents

CEG is a Córdoba-based company with more than 14 years of experience providing import services to B2B clients. The team set out to adapt its operations to the macroeconomic cycle, which required an organizational restructuring: some areas were overstaffed and others understaffed. As a result, the company couldn't serve its entire customer base efficiently or maintain the service quality it was known for.

The most concrete challenge was commercial scale: the company had a database of nearly 10,000 customers, but the sales team could actively work only about 2,000. The remaining 8,000 (active and inactive customers and new prospects from across Argentina) fell outside any follow-up structure. On top of that came a high volume of after-sales inquiries that demanded staff time the team could no longer sustain.

The central question was: how do you scale by evolving structures and internal processes and making better use of resources, without neglecting your people? The answer started not with technology, but with an in-depth redesign of its processes.

Before deploying any tool, the team spent six months redrawing its entire digital ecosystem: the end-to-end customer journey, from first contact with a new prospect through the full after-sales process. They defined behavioral metrics, calibrated the sales funnels and set a measurable baseline. Only once that map was clear did they start looking for tools that would fit the redesigned processes, not the other way around.

They chose to deploy conversational AI agents through the Atendium platform, which lets teams design and adjust agents in natural language, without specialized technical training. The team worked in complementary roles: one person defined the tone and brand identity, another designed the processes and response criteria, and another handled weekly moderation. Two main agents emerged: one focused on early lead nurturing and prospecting, which captures information, qualifies it and leaves a summary in the CRM; and a hybrid customer service and after-sales agent that answers frequent questions and automatically escalates critical cases to the human team.

The integration was deliberately non-invasive: the agents didn't change existing tools or processes. They acted as a support layer that freed up valuable time to focus on strategic accounts, without losing commercial reach or growth.

Just two months after the agents went live, building on eight months of redesigning processes, digital ecosystems and experiences, the results are concrete and measurable. The two main agents have started more than 5,200 conversations, with effectiveness rates of 65% to 75% depending on the agent. First-contact resolution in customer service holds at 95%, while the team has cut its working time by 35%.

On the operations side, the customer experience coordinator reports a 25% efficiency gain in customer service and collections. The tools made it easier to process high volumes of conversations, tagging and segmentation. The team's talent was redeployed to higher-value work: analysis, agent moderation, prompt design and sales strategy.

The result is a more efficient sales force that sustains sales levels with a leaner structure. The story isn't over: the team has already identified next steps to extend the agents to messaging channels and e-commerce platforms.