Over the years, the quality of tools available to the Capacity Manager have improved immensely. Little's Law may have been written before the advent of mass computing, but its application is crucial to predicting computing capacity. One of the earlier applications of the Law was to predict queuing on telephone exchanges. Whilst I am too 'young' to remember those days, I am reliably informed that it was common practice to employ teams of people to watching a bank of lights and manually count the number of flashes occurring within a defined period. The Capacity Manager would then use these observations in their analysis and predictions.
Things have moved on immeasurably since then. Not only has the use of computers increased, but the automatic collation of performance and capacity data has meant that the Capacity Manager can effectively define the metrics that they want recorded, go home for a day or two, and then return to analyse the output.
Capacity Toolsets (even simple spreadsheet programs) make the effort of analysing, modelling, and reporting so simple that many Capacity Managers in the industry today do not need to know what Little's Law is or how to apply it using pencil and paper. But can the tools do even more? Are we able to do away with the trouble of having a specialist Capacity Manager completely? Surely with the intelligence and capabilities of the Toolsets today, anybody that can point a mouse and click can be a Capacity Manager?
A major step in the work of a Capacity Manager is to identify the cause and effect. What business activity is causing the profile of resource utilisation, and what will be the effect of a change in that activity? To do this effectively, the Capacity Manager needs to understand the business processes that the IT service is supporting. They needed to ensure that they are measuring application transactions that relate to the business processes, and then they need to find a robust link between the application transaction and a measure of resource utilisation. Throw in a business forecast (which will need manipulation to turn it into an effective forecast of application transactions) and you have a recipe for quite a powerful artificial intelligence neural network.
Certainly, a Capacity Toolset can try to correlate EVERY resource metric against EVERY application transaction, and report the closest fit. But what good is a correlation between the count of inodes in use and the number of new customer records, if the system can only store a limited number of customers due to licence restrictions.
So, to answer the original question... No. A tool will never be able to replace an analyst. I would suggest the reader considers the analogy of Word Processing. These days, anyone can produce a newsletter or simple flyer for themselves... however Word Processing tools haven't turned us all into publishers able to produce a newspaper or magazine!