We use a multi-agent system to model how agents (representing firms) may collaborate and adapt in a business 'landscape' where some, more influential, firms are given the power to shape the landscape of other firms. The landscapes we study are based on the well-known NK model of Kauffman, with the addition of 'shapers', firms that can change the landscape's features for themselves and all other players. Our work investigates how firms that are additionally endowed with cognitive and experiential search, and the ability to form collaborations with other firms, can use these capabilities to adapt more quickly and adeptly. We find that, in a collaborative group, firms must still have a mind of their own and resist direct mimicry of stronger partners to attain better heights collectively. Larger groups and groups with more influential members generally do better, so targeted intelligent cooperation is beneficial. These conclusions are tentative, and our results show a sensitivity to landscape ruggedness and "malleability" (i.e. the capacity of the landscape to be changed by the shaper firms). Overall, our work demonstrates the potential of computer science, evolution, and machine learning to contribute to business strategy in these complex environments.