Flow strategy versus Next Best Action strategy for visit recommendations
Key Takeaway
AI model-based visit recommendations from Einstein Discovery are available only when the request uses a Next Best Action (NBA) strategy. A flow strategy can produce rules-based recommendations but not model-based ones.
Einstein Visit Recommendations in Agentforce Consumer Goods suggest which stores field reps should visit and when. Each recommendation request runs a strategy. The strategy is either a flow strategy or a Next Best Action (NBA) strategy. Both types can produce rules-based recommendations. Only an NBA strategy can carry AI model-based recommendations from Einstein Discovery, so the strategy type decides which kinds of recommendations a request can return.
Rules-based and AI model-based recommendations
Visit recommendations come in two kinds. Rules-based recommendations follow business priorities that the company defines. AI model-based recommendations come from an Einstein Discovery model trained on historical store visit data. Each model has a specific objective, such as maximizing revenue.
Rules-based recommendations can combine several rules. For example, a company might give priority to:
- A store that hasn't been visited in the last 21 days
- A Tier A store over a Tier D store
- A store one week before a promotion start date
- A store that has run short of stock for the last five visits