Advice is complex. Consider two financially identical individuals living in the same street with the same job, mortgage and commitments. Taking the bare facts, the robo adviser will reach the same conclusion for both. But throw in attitudes, opinions, gender and experiences and the advice outcomes can suddenly become very different.

Robots jump on attitude to risk which in relative terms is simple to capture with a psychometric questionnaire, but what of other attitudes? Attitudes to paying tax (or not), to charges and the perceived value of guru investment managers, types of savings vehicles, restrictions and commitments, these will all be important when making a recommendation. Opinions on social and ethical issues, and those arising from learned experience will all point to potentially different conclusions.

Overlay the fact that, for individuals, some attitudes and beliefs trump others and it becomes clear that the challenge for the robo programer is enormous. Marrying this with a requirement to work within a regulatory rulebook adds to the challenge. The response so far has been, not to build an all singing bionic adviser (it’s too big a job), but rather “restricted” parts. Arms and legs rather than a brain.

An arm might be a discretionary portfolio reflecting a particular attitude to risk and term. Simple to build and maintain, it is relatively easy to work out whether or not a specific objective is achievable and the probability of achieving it. Some may argue that this is simply replicating a traditional managed fund, albeit with a little more transparency and a different set of tax consequences. Now the customer does the selection taking the adviser out of the loop.

A leg might be a quotation engine comparing non-complex products such as term assurance. Here the moving parts are simple but understanding affordability and the levels and types of cover, while balancing these with other needs, adds a layer of complexity that is not, so far, seen in the robots available.

The danger of using separate mini robots rather than a joined up machine, is that users address one need in isolation to others. Invest rather than pay down debt. Perhaps buy a product that isn’t needed. The Financial Advice Market Review (“FAMR”) recognises the need to broaden access to advice. Robo advisers are a potential solution but while they remain fragmented parts of the advice process, they present a different set of risks. For customers who understand and know what they want, with cash sitting around, the robo investor is a potential option. But these are precisely the people already catered for and not the object of the FAMR.

For those who have worked hard and done the right thing, with modest savings, cashing in pensions or selling existing assets which may be fit for purpose, the robo adviser presents a potential mis-buying opportunity. Getting the disclosure right in a world where people click accept without reading the terms will be as important as programing the advice engine itself. It has the potential to become a turning point in future FOS cases.