"Stuck in Half Manual mode"
Most (all other?) algorithmic asset management products are very similar: An R&D department continuously develops what they refer to as algorithmic systems, usually a combination of proprietary as well as common TA-methods that are tweaked to correlate with the current market situation.
Typically these algorithms can show performance during 0-36 months, but used over longer time they don't add value. The developers therefore need to continue to develop algorithmic systems and the asset managers constantly need to decide which ones are to be used for the actual asset management.
With the traditional algorithmic methods one can say that asset management has gone from selecting financial instruments to selecting algorithms. Unfortunately there is little evidence to support that it is easier than other forms of asset management, it requires quite a few extremely skilled and experienced human resources to be success over time. In any case, note the limitation that because of the manual R&D and selection process it is actually not possible to backtest the long term performance of the total portfolio.