Polish AI harness, part 2: AI in the team without losing skills
Series · episode 2 of 3
An AI assistant speeds up work, but the success of a deployment depends on whether people still understand their work, learn the profession and have influence over the changes. The second part of the series is about people
We return to the salesperson from part 1. The assistant finds the price list and creates a draft offer, and the salesperson finishes the task faster. After a few weeks, however, the expected number of offers grows, there is less and less time for checking, and a junior member of the team starts approving proposals they cannot explain on their own. A technical measurement still shows success, although dependence on the system is growing
The second part of the series is about how to avoid this. A good harness gives employees influence over the changes: the ability to check answers, question recommendations and develop their own skills
Six principles of a good harness for people
How SAIE helps
In SAIE, people have the final word
- TWIN:DESK quotes a specific fragment of the source, so the answer can be checked straight away
- the agent presents a plan for approval before it does anything
- the human competence gate in the CDF methodology helps to check whether the approving person understands the premises of the decision
- new users are guided through TWIN:DESK by a walkthrough tailored to their role
Success will mean that an employee who understands the terms of a reliable offer and has time to check them prepares it more efficiently
The full post is on our blog