AI assistants: does a local gain benefit the whole team?
Building an assistant without coding makes task automation accessible, but does not guarantee an organisational benefit. The management question is specific: how can we check that time saved by the user does not become additional work for colleagues?
What the literature establishes
Thomas W. Malone and Kevin Crowston (1994, ACM Computing Surveys) propose understanding coordination as the management of dependencies between activities. Their review shows why a task cannot be considered independently of the resources, prerequisites and outputs it shares with other tasks. Applied to an AI assistant, this framework directs attention not only to what it produces, but also to what that output demands of its recipient.
The mechanism: organisational complements
Erik Brynjolfsson and Lorin M. Hitt (2000, Journal of Economic Perspectives) review research linking the benefits of information technology to complementary investments, particularly in working practices and organisation. Installing a tool therefore does not amount to transforming a process effectively. For a no-code assistant, the hypothesis to test is that benefits also depend on handover rules: who receives the output, in what form and for what purpose.
The tempting conclusion
The tempting conclusion is that an assistant useful to its user must also be useful to the team. Yet a rapidly produced summary may require the receiving department to reclassify information, seek clarification or re-enter data. The point is not to demand wholesale reorganisation before every trial, but to avoid confusing a local saving with a collective benefit. (our executive and employee training programmes)
What the evidence does not establish
These articles address coordination and information technology, not today’s generative assistants; they provide an analytical framework rather than an estimate of their effects. Nor does the economic research discussed justify attributing every performance difference to the tool alone. An evaluation restricted to generation time or user satisfaction leaves the recipient’s work outside the measurement boundary.
A practical check in Sion
Within SHR — Swiss Human Resources’ programme « Créer ses assistants et agents IA sans coder », an exercise in Sion could track an administrative handover: a hydropower maintenance request, a non-clinical healthcare file, a wine order or an alpine tourism booking. Over a pilot period defined in advance, compare cases of similar complexity with and without an assistant, adding the sender’s preparation time to the recipient’s processing time through to acceptance. Also record requests for clarification and rework, using case identifiers rather than their contents. The criterion to check would be a reduction in total time per accepted case without an increase in rework: an improvement for the sender alone would not be sufficient. To go further: explore the Building AI Assistants and Agents Without Coding training in Sion in the canton of Valais, or browse our executive and employee training programmes in Switzerland.
In pictures: Building AI Assistants and Agents Without Coding in Sion
- building ai assistants and agents without coding
- Sion
- research
- training sion