Photo credit: venturebeat.com
This article is part of VentureBeat’s special issue, “AI at Scale: From Vision to Viability.” Read more from this special issue here.
Just a few months ago, when customers reached out to Baca Systems for assistance, human representatives would take several minutes—typically five to seven—to locate relevant information and begin addressing their queries. This process involved extensive searching through databases and analysis of technical documents before they could provide any substantial support.
However, thanks to the integration of AI agents powered by Salesforce, response times have dramatically improved to a matter of seconds, now requiring just five to ten seconds for an initial, meaningful answer. Andrew Russo, an enterprise architect at Baca Systems, emphasized that the company’s focus is not on reducing personnel, but on enhancing the speed and effectiveness of customer support.
Closing time gaps, delivering faster time to resolution
Since adopting Salesforce back in 2014, Baca Systems, a robotics manufacturer in Michigan, shifted from a basic service platform to the more robust Service Cloud. This transition laid the groundwork for a broader digital transformation that began in 2021 when the company integrated Salesforce’s enterprise resource planning (ERP) solution.
As part of this transformation, Baca Systems employed predictive AI for both sales and manufacturing analytics, ultimately advancing to the use of AI agents within the past year through Salesforce’s Agentforce. A major application for these AI assistants is in handling service calls, with approximately 57% of customer inquiries relating to hardware issues like machinery falls or the need for recalibration.
The AI now handles the task of retrieving relevant information, allowing human representatives to focus more on facilitating immediate resolutions instead of spending time on data retrieval. Moreover, it has the potential to proactively identify issues, such as wiring problems indicated by repeatedly tripping circuit breakers, thereby preventing recurring equipment failures.
“Our ultimate goal is to shorten the resolution time for our customers,” stated Russo.
AI agents generating sales leads, handling customer inquiries
Sales operations are another significant area for Baca Systems, which operates with a lean team of fewer than ten sales personnel. Russo noted that the company faces an abundance of potential leads that have gone unaddressed due to insufficient manpower. “The question we face is how to meaningfully engage those leads,” he remarked.
AI functions as a capable sales development representative, capable of sending out initial communications and engaging prospects in automated conversations before forwarding them to the appropriate sales team member. Hiring additional sales staff would entail considerable costs, making the cost-effectiveness of employing AI for lead generation clear and compelling.
In the near future, Baca plans to deploy customer-facing service agents that will interact with users via text messaging, managing inquiries autonomously unless the situation requires escalation to a human representative. Russo articulated that their ongoing mission is to deliver increased value to customers in both services and sales.
Additionally, Baca is leveraging AI for tasks such as email generation, account receivables, and crafting stringent collection letters when necessary. Russo has also harnessed AI technology for data cleaning tasks, employing retrieval-augmented generation (RAG) to eliminate duplicates and maintain data integrity within Salesforce. He reported that employee feedback has been overwhelmingly positive, with many team members eagerly seeking access to AI capabilities, underscoring the technology’s role in enhancing job performance.
The company is committed to a gradual and careful approach to AI integration to ensure flexibility and adaptability in operations. “Our objectives remain intact; we are simply discovering new pathways to achieve them,” Russo remarked. “It’s a more efficient and effective journey — akin to switching to the fast lane.”
AI serving up savings for ezCater
The corporate catering industry entails a host of complexities, including fluctuating headcounts, diverse dietary preferences, and various logistical hurdles. As ezCater continues to expand within this competitive market, the demands for high-quality service can become challenging to fulfill without increased technological support, according to Erin DeCesare, the company’s CTO.
With the implementation of Salesforce’s Agentforce, ezCater plans to streamline operations by enabling customers to interact with AI agents to modify their orders naturally. For more complicated requests, like changing venues or restructuring orders, the AI will automatically escalate the matter to a human representative, significantly reducing operational costs, as DeCesare pointed out.
Additionally, ezCater aims to utilize AI for “restaurant discovery,” allowing AI agents to recommend ideal venues based on user preferences, budget constraints, and location—backed by extensive data from millions of orders. “Utilizing natural language processing and AI is what we excel at,” DeCesare stated.
The initial focus is to integrate AI agents to support human agents internally, which has been well-received, as these tools enhance their capabilities and efficiency. DeCesare also noted that the comfort level among software engineers has shifted positively, as they can now approach AI development similarly to traditional software development practices. “It’s transforming their expectations and intuition around the development process,” she observed.
Excitement about the potential for broader applications is palpable among business partners, such as in areas like business analytics and process mapping. DeCesare remarked on the accessibility of AI technology that has surged in recent months, envisioning a substantial evolution in operational norms within the next year. “We will find ourselves in a vastly different landscape in just 12 months,” she concluded.
Source
venturebeat.com