How to choose an AI consultant

Adding AI or RPA for that matter to a business can add huge value, it can reduce labour costs, provide cutting edge decisions and add real efficiency to a wide range of industries and processes. But how do you ensure that you pick the right consultancy for the job. Making the right choice can ensure the best delivery for your organisation. 

Firstly though, what are you expecting from a AI consultant?

The most important part it the ability to understand needs and successfully translate those initial conversations into real plans and recommendations. From there, finding the right skills to implement a development plan can also often be supported by a good consultant. 

Top AI consultants need the following to be successful:

  1. Be able to understand the technology

Often consultants are focused with one platform or technology that they use because they are partnered with the tech supplier. This can often lead to shoehorning tech into solutions to fill a quota of delivering a certain platform. The best AI consultants will be have a holistic view of the field of AI and know what works and how. 

  1. Have strong links to development.

Half the job is providing a proposal for work. Having links to development is beneficial in many ways. It allows them to offer a way to deliver any plan, putting there commitment in what they have proposed. It often means that they are working closely with the development teams, which can mean that they understand the detailed considerations of a data scientist. 

  1. Know the right processes 

Process management can be the make or break for a digital project. Even more so in AI. The technology needs to be delivered in such a way that it integrates with the real world to add value, testing this and developing this needs the correct processes to ensure success once the development is over. This includes how to test and how to evaluate a AI project during and even before development

  1. Willing to partner to drive success 

Consultancy is often considered as a mercenary of the business world, picking a consultant that can take the understanding is one things, but a consultant should want to drive the ideals and project forward to ensure success. This is especially true if development and delivery is required. Make sure the consultant can also wear the project management cap when needed to make things work and operate as needed.

  1. Understand risk

Risk and risk mitigation is essential for any project. As for an AI or RPA focused project being able to generate the risks ahead of time and provide real mitigation plans up front is very important for an AI consultant. These mitigations can either be pushed though better planning and information control or by working closely with the teams and persons that are requesting the solution. Building a good working relationship with a consultant is another skill that a consultant should be focused at to better the project. 

  1. Be focused on security 

Security systems and adhering to GDPR almost goes without saying. Picking a focused dedicated consultant is all well and good, but they always need to build correct security procedures into everything they do. Very often, especially with new AI systems, they will include a GUI (graphical user interface) that needs a login to access the data rich system; having the knowledge, experience and ability to deliver such essential requirements is mandatory for any AI consultant. 

7. Understand the value of MVP or proof of concept.

For a deep dive on the topic of the value of MVP’s see Pro AI’s article on the topic. To summarise, a consultant needs to have the knowledge of process that includes focusing on a small number of use cases to maximise success in many different ways. 

8. Be confident in deployment 

Deployment is the area in AI development that contains the most risk. All the work and new code needs to be integrated into an operational environment (assuming there is one already). Any bug can cause catastrophic damage to customer facing use and operation. If the system is one that pertains urgent information or is relied on my users, this can be drastic. Not only does a consultant need to be focused at delivering a system free of bugs, they need to ensure that integration to the existing system is secured and has backups. This could range from having a sophisticated automated staging environment that pushes docker containers of new code through rigorous automatic code evaluation checks before merging with production, to having a high speed backup of the operating code to revert back to in the case of failure. Provisioning of these plans is essential for a consultant. 

Big companies are assumed to have more resources to attribute to projects, but the real demands should be the ones above, as these 8 points are the real focus to ensure the consultancy role is what is really needed. This in turn gives a better chance of success. 

Leave a Reply