The AI Endgame Is Decision Automation
In the past year, almost every organisation has initiated an artificial intelligence (AI) proof of concept (POC), but only 13 percent of these have translated into real projects. With organisations increasingly fascinated by AI, it’s crucial to focus on the tangible return on transformation investment (RTI). The key to achieving this lies in decision automation.
Decision automation involves applying business rules, data analysis, workflows, and AI to automate decision-making processes in both operational and strategic contexts. For Customer Experience (CX) leaders, this could mean determining optimal times for ad buys in a campaign, adjusting pricing strategies for dynamic discounting, sending follow-up messages for upselling or cross-selling, or gauging customer satisfaction following a purchase. Essentially, it reimagines every end-to-end CX process through a five-step cognitive framework:
- Learn – AI systems replicate the five human senses to collect contextually relevant information such as time, location, and business process context.
- Understand – These systems employ reasoning by comparing past interactions with future predicted interactions to derive an understanding of the business environment.
- Recommend – By studying past behaviours and current conditions, AI systems make recommendations that will influence future decision-making.
- Act – The decision automation lifecycle is actualised by executing these decisions, thereby revealing the consequences of actions taken.
- Refine – Systems analyse the outcomes of their actions and strive to reduce inaccuracies, ultimately refining the decision-making process.
Building the foundation for decision automation necessitates a holistic approach that begins with the creation of an abstraction layer on transactional systems. These include data, customer journeys, and user experiences. Organisations have been working diligently to maintain and augment these systems by incorporating context, security, identity, and intelligence. This forms the building blocks for intelligent orchestration.
As a consequence, there's been a surge in efforts to develop Customer Data Platforms (CDPs) and interlink them with intelligent processes and experience hubs. These form the intelligent orchestration services necessary for creating business graphs and multimodal models that drive decision automation.
Achieving decision automation enables organisations to deliver on several fronts, including personalisation, AI-driven decision engines, and situational awareness. Through this approach, organisations could realise the long-envisioned 360-degree customer view, with significant impacts on their bottom line.
According to Constellation’s “State of AI in the Enterprise Survey,” 76 percent of respondents cited operational efficiency as their highest RTI. Basic CX automation, seen as the lowest-hanging fruit for AI applications, promises both cost-savings and increased revenue. Additionally, 73 percent of respondents reported RTI in terms of revenue and growth, while 52 percent cited cost reduction as a top benefit.
For early adopters, Constellation’s AI 150 executive network suggests asking five key questions to test decision automation readiness:
- Where and when do you insert a human? Identifying the correct points for human intervention enhances the efficiency and effectiveness of the AI system.
- Can you operate at machine scale with humans? While machines make thousands of decisions per second, integrating human-scale operations with machine-scale efficiency is vital.
- Do you have enough data to achieve precision decisions? Precision requires extensive data sources. For instance, 85 percent accuracy might be acceptable in CX but unacceptable in finance.
- Who do you partner with for the last mile of data? Collaborating with partners across value chains is essential for achieving high levels of trust and data accuracy.
- Who is accountable when something goes wrong? Establishing blame is complex; it could lie with the system, operator, partner, or another third party.
These considerations highlight the intricacies involved in moving towards decision automation. The transition from AI concepts to actionable projects that deliver measurable outcomes is complex but achievable with a strategic and informed approach.
Source: Noah Wire Services