In a not so distant future where we still would like to supply software, the primary user interaction with enterprise software will happen through natural language communication or someone else’s software. I strongly believe – together with others – that the days of specialized UI and applications offering UI, as we know it, are coming to an end. In essence it is about no-UI. From my perspective ERP is nothing more than a database containing structured data. It is – honestly – only marginally more complex than the underlying database engine itself.
Previously data capture happened by means of specialized UI. Each application had its own UI, which typically carried a remarkable resemblance with the underlying data model, represented as a database table in a relational database. Recent development in connectivity, data storing capabilities, compute and IoT – allows for capturing majority, if not all, data previously entered manually. The need for specialized UI will drop proportional with the increase of electronic data feeds. Making ERP and other enterprise solutions even more remote, from an end user perspective, it becomes distant, nearly non-existent – but still present and pervasive. Just like the database itself. No one worries about the database – maybe apart from very few specialized employees, same will happen to ERP and enterprise software as a whole. In the future no one will really care about ERP, like nobody cares about a database.
From an end user interaction perspective, then all interactions should be happening by using natural language or someone else’s UI. An UI that users are used to use as part of their private life. Something they are familiar with. The latter is also driven by the changes in how future employment will be. Recent research suggests a tendency to move toward an employment model based on contingent employment over permanent employees. To support such an employment agility, companies will be looking for enterprise software that requires minimum to no training to use, as it will otherwise, become too expensive and time consuming to adjust the work force. And what is better for supporting such agility and limit on-boarding time? A conversational user experience using natural language, and if absolutely required an UI embedded into platforms already used by the majority of people. A practical example of such could be using Twitter for workflow approvals and LinkedIn for HCM related functionality.
With the above being said, then we should not ask ourselves “what cannot be done by a bot?”, but rather “why can that not be done by a bot?”. Everything is either applicable for natural language or handled by underlying self-driving capabilities – in the case of UI, use someone else’s.