Agents for growth: Turning AI promise into impact

Key takeaways

End-to-end value is generated by change. Expansive productivity victories are table stakes. Impact is achieved through focusing on the largest growth issues and resolving them. Reprofit work processes and not tools. Expansion is achieved by mapping decisions and handoffs and placing agents at the point of change, rather than attaching them to existing steps. New operating model scale. End-to-end transformation necessitates cross-functional human-AI groups, overlapping data products and governance in which agents are viewed as assets that are managed as talent. A multinational retailer will experience a boost in demand of a high-quality product in one area and excess stock in another. A team of AI agents can reallocate the ad spend, optimize pricing, stock reroutes, and refresh creative assets within seconds to offers that align with the shopper intent. The next step, in this case, is the coordinated work activated by the signals of the customers and the organization of the growth of business on the spot. This is not an imaginary situation; it is the future of AI in increasing functions. The agentic AI places automated reasoning within the marketing, sales, and customer service processes. We put the number of deployments in marketing and sales that will generate more value using agentic AI at over 60 percent. Early proofs indicate that gen AI has the potential to release up to $2.6 to 4.4 trillion of annual value, and as much as 20 percent of the projected productivity augment is in the regions of marketing and sales. Marketing and sales are, quite literally, the end of the spear in terms of quality delivery of the potential of agentic AI into any real value. The foundation models of gen AI are the basis of agentic AI which is a system that is capable of acting in the real world and is applicable in performing multistep processes. Automation and complex tasks can be automated and done by AI agents with the aid of natural language processing, which would have otherwise been done by human beings. The impact already can be observed among early leaders. As an illustration, based on the analysis it has been estimated that some of the fortune 250 companies are experiencing a 15-fold acceleration in the speed of campaign creation and execution due to an increase in the speed of innovation and process optimization.

They can, say, streamline prices, push ahead leads, customize offers and can handle customer dealings up to the end. Gains can go up as organizations further embrace agentic AI. It has been analyzed that properly and scaled agent deployments can provide productivity gains of between three and five percent per year and may also get growth to rise by 10 percent or even higher. The meaningful value of AI, however, has not yet been achieved by most organizations. Almost 8 out of 10 report no material bottom-line benefits of AI in general, primarily because of the limitation associated with fragmented pilot programs, poor data, and lack of sufficient bases of governance. The leaders that have broken through and extracted value out of AI are re-architecturing the process of growth by adopting AI agents into their process. Through experience in industries, there are four lessons that are being learned by organizations that are making breakthroughs and transforming agentic AI promise to performance in marketing and sales:

Impact starts with the location of where the agent can move the needle, be it conversion, pricing accuracy, or engaging the customer, and mobilizing the agents to hasten those results. Think of the subject of personalization where the chance is not only tested but also deep. It has been research-proven that 71 percent of the consumers anticipate personalized interactions, and 76 percent are frustrated when they do not occur. The results of AI-personalization in increasing customer satisfaction by 15 to 20 percent, growing revenue by 5 to 8 percent, and decreasing the cost to serve by up to 30 percent are impressive. This becomes possible at scale with agentic AI, which applies contextual reasoning and real-time decisioning to perfect offers, content and experiences with each interaction.

European insurer success story

As per the analysis, a European insurer, e.g., redefined its sales process using AI agents who tailored campaigns in hundreds of microsegments, customized scripts according to buyer cues, and trained sales teams with real-time feedback. The outcome:

• two to three times more conversion rates
• a quarter reduction in the length of the customer service call
• ongoing learning loops that the manual reviews were unable to equal in effectiveness

US airline case study

Similar practices are being applied by other organizations with the help of AI to raise the customer experience by understanding what each customer wants next and offering it at the right time. US airline applied predictive insights by applying the same compensation to flight disruption by distinguishing between frequent travelers and occasional travelers. The impact:

• 210 percent increase in targeting at-risk customers
• 800 percent increase in customer satisfaction
• 59 percent decrease in the churn amongst high-value travelers

Dynamic pricing applications

The same sort of intelligence is also finding application in the area of improving pricing. It is possible to have agentic AI perceive market changes, calculate the results and respond immediately by either changing the prices or redistributing the inventory in real-time depending on what competitors are doing, consumer changes, or demand predictions. An example is airlines which already leverage agentic AI to generate custom packages that are based on fares, seats, and optional offers that are dynamically updated to react to live information, including search trends, weather, and booking behavior.

Organizations that see meaningful contribution to agentic AI are not merely putting into action new agents to enhance the current tasks, but are redesigning workflows. The agents add value when applied to the end-to-end processes and journeys by adding automation and coordination-limited strength, but when applied to individual steps, their strength is restricted.

Traditional processes can be described as flowing one way, typically between departments: The Marketing department passes the work to sales, escalation to the service department, and pricing. The advancements in each of these functions have been enormous over the past few years with the integration of digital and analytics functionality and agentic AI builds on these advances by automating and coordinating activities at the team and functional levels. By overcoming the endemic issues of coordination between the complicated silos and workflows of operations, organizations may be able to attain shorter cycle times, in addition to more consistency and responsiveness at a scale that no amount of human coordination could match. More importantly, success requires the process design based on agents- not attaching agents to old processes. As an illustration, instead of assigning agents to assist customer service departments to answer their grievances more quickly, leading organizations deploy agents to forecast possible problems, make pre-emptive calls to a customer before a call, and address situations in advance with their personal offers.

European insurer transformation

To obtain a clear picture of what this would be in practice, the European insurer provides a clear perspective of the same. As per the analysis, within 16 weeks, the firm re-architected a commercial model of a network of connected agents operating throughout the entire customer experience. The resulting improvements:

• Knowledge agents had centralized more than 1,000 policy and product documents allowing frontline employees to find out the relevant answer immediately
• The AI-powered call transcription and grading by the coaching agents ensued that 95 percent of the sales calls were reviewed automatically compared to 3 percent before
• These capabilities were integrated by the integration agents into the available CRM and agent portal--following single-sign-on security policies and offering real-time performance dashboards These agentic systems, collectively, resulted in a 25 percent reduction of the average call time, less cross-functional handoffs that were manually performed, and a feedback loop. This experience made the agents keep improving next-best actions, sequence of messages, and product matching to keep abreast with changing customer demand.

Agent types and roles

End to end change will require creating value with AI agents, however, requires the right agent to the right task:

• domain specific agents that perform complex, contextually sensitive actions
• generalist agents that perform actions like data synthesis or content creation
• error checking agents
• orchestration agents to coordinate and align the entire system The most important role in this endeavor is played by humans. They are able to closely collaborate with the agents to oversee and check and handle issues that the AI agents present to them. The most evolved firms integrate these human-agent interactions into adaptive workflows, which change in every iteration and customer signal.

Organizations must abandon viewing an agentic AI as an add-on tool and begin to view them as digital partners. That is to give the roles of the agents, onboard them appropriately and manage them with clear performance expectations- just like human team members. The appropriate metrics of the performance of AI agents, however, vary with the conventional productivity KPIs. Instead of counting calls, or the number of campaigns, top-performing organizations consider a combination of measures:

• quality of conversation
• accuracy in completing tasks
• accuracy in responding to escalation
• learning rate, which demonstrate the ability of agents to absorb feedback and adjust to changing buyer behaviors These metrics are continuously monitored because all the actions of the agents are recorded and can be traced. Real-time dashboards indicate performance changes, compare benchmark results with human threshold, and indicate when retraining or recalibration is required.

US homebuilder case study

One of the largest homebuilders in the US shows the way this discipline can lead to an effect. In order to enhance online communication and appointment booking, the company trained AI sales representatives and trained them to act like its highest performing salespeople. When over 500,000 sales transcripts were analyzed, dozens of conversation states - greeting, objection handling, follow-up, close and which patterns are most commonly linked with success were identified. Based on these insights, the team created agent personas with distinctive styles, tempos and conversational styles. All AI dialogues were then compared with human baselines through a scoring agent that assessed the accuracy, personalization, and flow. The Dashboards showed drop-offs and tone mismatches, which allowed quick setting of tunes. Results:

• The rate of conversion-to-appointment has increased by three times
• weekly appointments increased by two times
• the most effective agents were showing human-like equality in empathy and flow

As agents assume the workflows that cross marketing, sales, and customer service, businesses should reconsider the process of organization of growth. The historic approach of each of the functions working in its silo is being replaced with an integrated approach where agents coordinate their activities, share data and join the whole customer experience to create a customer experience between the moment of awareness and loyalty. The design of campaigns, conversion of leads, and customer interaction are not in a sequence anymore, but in one, learning loop. This cannot be done without a new, hybrid human-AI operating model. This system involves agents to do the coordination and implementation with humans to supply strategy, creativity and control. The cross-functional nature of growth teams is constructed in such a way that marketers, sellers, customer care representatives and data scientists work together based on common workflows and goals.

Agent factories

Scaling requires that the leading companies are establishing agent factories: specific hubs that industrialize the process of agent creation, deployment, and regulation. These hubs formalize reusable architecture, common data products and security and compliance guardrails. And the standardized agents that they construct are reserved to stakeholder, role-based obligations, in such that:

• lead agents plan the work
• practitioner agents perform the actions
• QA and compliance agents oversee the performance Examples of this strategy are several global banks, who have gone ahead to set up agent factories to remodel their due-diligence efforts. Every factory has its own agent squads to process individual steps, including data extraction and validation and quality assurance, which minimize the number of manual steps and enhance the accuracy and control.

Outdoor lifestyle producer example

One of the major North American producers of outdoor lifestyle products used the same principles in their approach to customer service. As the analysis indicated, following the analysis of over 30,000 service tickets and call transcripts, the company redesigned the function to have agents deal with diagnosis, data retrieval, and summarization and leave the human-centered treatment to humans. The implementation has been successful due to the customized change-management plan that involved the training of leaders on KPI dashboards, frontline employees in job aids in AI assisted workflow and technical staff in model maintenance and tuning. Lasting loops of feedback and joint dashboard will ensure that both human and digital agents are oriented towards achieving quicker resolution time, increased satisfaction and quantifiable revenue growth.

Human capability transformation

As such systems evolve, the differentiator is transformed into human capability. People no longer perform; they now manage, perfect and enhance the manner in which the job is done. The managers and specialists should be taught to delegate to the agents, review the output, identify exceptions and direct the learning loops. New competencies, including timely design, result monitoring, and escalation control, are rapidly becoming part of the current growth functions. A significant number of organizations already aim to employ agentic AI with 25 to 50 percent of employees on a regular basis a good indicator that the ability to collaborate with AI is becoming a business advantage.

A year later in the agentic AI era, the lesson is now obvious, development will not be a growth of tools but rather the way leaders will establish and implement them. The competitive advantage will not be determined by the number of agents that a company launches but their design, management, and scale effectiveness. New mindsets are already being implemented in the companies that are taking the lead. It is just the start of the transformation agentic AI will cause, soon larger questions will arise, such as:

• When your sales agent is negotiating with the buying agent of your customer, how will your company stand out?
• In case execution is commoditized, what do you have to offer as your brand?
• With workflow cross silos how will you maintain accountability and control? These are the agentic era of leadership tests. Very soon, the main question to leaders will no longer be what this agent can do to us but what results do I intend to achieve with it, and how can I make the most of the space it provides to enable humans to do whatever only they can do even better? The faster the organizations incorporate the agentic AI in the marketing, sales, and customer support processes, the faster they will be able to find the responses to those bigger questions.