Idea management software needs AI to deliver value

In large groups, tens of thousands of employees are likely to take part in a process of collective intelligence and corporate innovation. Even if only 10% of the employees play the game, this can represent several hundred ideas, suggestions and comments per day.

It is, of course, impossible to sort all these contributions by hand. This is why integrating an artificial intelligence (AI) component into an idea management software is highly strategic. If it is well structured, it can provide comfort and support in use that serves both the back office and the front office. A good way to save time, generate engagement, develop more relevant ideas and measure and structure the ideas submitted. When it comes to corporate innovation, AI mainly performs decision analysis work. Intelligence then becomes collective and artificial, resulting in a massive impact for your users.

AI is a term that has been talked about a lot for several years. Sometimes falsely associated with programming or automation, it is increasingly presented as a sales argument. As in the comics, we would like to imagine it as relevant as the one used by Iron Man with his famous Jarvis – the artificial intelligence that serves as an omniscient assistant. Sadly, we are still a long way off. In most professional solutions, for an AI to be functional, it must be based on a proven self-learning structure.

What is deep learning?

It is a type of AI that gives a machine the ability to learn on its own. Where programming is based on rules developed by engineers, deep learning is based on a network of artificial neurons inspired by the brain. Each piece of information is analyzed, decoded and interpreted according to the data previously received. Thanks to deep learning, a computer is able to learn by example. To achieve this, the more the system accumulates different experiences, the better it will perform. It is a technology that requires significant computing power to train neural networks.

A decision support method

When it comes to corporate innovation, it is simply indispensable. It is an assistance tool that will boost human capabilities to work faster, process large volumes of data and provide human decision-makers with keys and objective KPIs to make important decisions.

AI can, therefore, make a real difference between classic or redundant topics, and truly relevant, strategic and new topics. Based on a fine-grained semantic analysis, it will be able to bring up the right topics at the right time, and above all, to the right person according to the users’ interests and their interactions. A fairly classic approach that can be found on many platforms such as Netflix or Amazon.

Why use AI in an idea management software?

Without this technological building block, the idea management platform is ultimately just a box of ideas that must be manually sifted through. This represents a considerable amount of work and is difficult to sustain in the long term. But this is not the only advantage.

To promote engagement

To work, corporate innovation needs an idea management software. It is via this tool that the contributions will be submitted, enriched and commented on. Its principle is quite similar to that of the newsfeed on social networks. The topics most likely to generate interactions are prioritized according to the user profile. This way, everyone can see the things they are really interested in. Fostering engagement is essential to keep participants coming back regularly. And today, there is no better system than AI.

To not disappoint users

AI is an important qualitative filter. Without it, we make investments to bring in users at a loss. If they are disappointed, they will not come back. Someone who wastes personal time on an unsatisfactory experience may even infect novice users, discouraging them and turning them into resistant or disengaged users. The level of expectation generated by the launch and communication must, therefore, crystallize in the use of the platform to align resources with the carefully engineered impatience.

To filter relevant topics

Even with the best will in the world, one cannot be interested in everything. Choices must therefore be made, and AI brings significant added value. The topics are thus classified according to the daily challenges of the employees’ work and include peripheral topics likely to interest them. However, this profiling system is not solely focused on the business. It can also be based on employees’ interests in order to maximize their engagement potential. In order for it to be truly useful, the idea management platform must therefore avoid subjects for which the employee has no particular knowledge or interest.

To create a market for innovative co-creation

A market represents a supply and a demand. If the two parties meet and get along, then everything goes smoothly. Collaborative innovation coupled with AI-optimized decision analysis is therefore a market: the supply is based on proposals and ideas for improvement or creation, and the demand on the willingness to get involved and participate. AI aims to create and balance supply and demand to satisfy all parties. Avoid non-committal ideas and value rare nuggets, while offering new choices or engaged discussions for contributors.

How can we use artificial intelligence in a corporate innovation process?

AI is a term that can be scary. In reality, it is mainly there to support the proper functioning of the idea management software. But it is still necessary that its implementation be well implemented, tested and integrated at the right time.

Do not wait to deploy it

Adding an AI component to a idea management platform – whatever it is – necessarily increases the cost and deployment of an ideation tool compared to a traditional system. However, this remains an important strategic decision for maximum scalability. I speak from experience, as I have experienced collective ideation situations where AI was either absent or inoperative. As a result, when you access an idea management platform with a volume of ideas that can reach 10,000 contributions, you find yourself overwhelmed. You can’t find what you want, the search is dysfunctional and time-consuming, and nothing on the platform is ultimately relevant to employees. If you look for everything, you will find nothing.

To solve this problem, the principle must be reversed: the management of priority should not rest on the shoulders of the participants, but should be the responsibility of the machine. For many, after one unsuccessful attempt – or two at best – the retention rate plummets. Why spend time on a useless tool? This is why AI should not be an add-on feature that is developed and added in a second or third phase of development. It must be available and operational as soon as the tool is launched. Without it, there can be no idea management software.

Refine the profiling process

For AI to be effective, it must combine user data with the content presented on the platform. Upon first logging in, the user defines their skills, their profession, their service and their areas of expertise. Then, the AI refines its recommendations according to the interactions. In daily use, we are quick to forget that ideas are proposed by algorithms with a matching system. And that is precisely what makes a good AI. The important thing is to immediately find an idea that will provoke the employee to respond, without the need to use the search engine.

Choose the right technology

The choice of technology behind the AI is important, since it is an assistance system that must be appropriate. Some companies may want to go on their own and develop an internal system. I’d rather be honest with you. If that’s your intention, it’s a complete waste of time and money. Developing idea management software is much more effective when you rely on existing tools. At Yumana, we use Microsoft AI technology at the core of the ideation engine. Using this solution is an advantage because all companies know and use the relevant technology. In addition, I have noticed that large companies really trust Microsoft’s tools.

Manage multilingualism

In large groups established in several countries, it is natural to speak several languages. This must also be the case in idea management software solutions. Rather than using English throughout, which can inhibit ownership of the content and limit nuance and precision when expressing ideas, it is recommended to use the users’ native language. So all ideas, tools and interactions are handled in the default language.

How can you measure the performance of a idea management software integrating AI?

Like any system, we want to make sure it works well. To do this, there are several options. The first is the formal (questionnaires, panels, semi-structured interviews, etc.) or informal (rumours, impromptu discussions, feedback from the department, etc.) collection of internal feedback. These are practices that are easy to implement and which remain under the control of the company and the project team. Here we emphasize human interaction, we collect testimonials, and this is an approach that is generally used to understand the use made of the tool. It can also be used in terms of training or change management to support the transition.

There are other KPIs that the platform can centralize in a dashboard. This includes the most popular ideas, the best contributions, the popularity rate, the attitude of the community, social interactions, etc. These qualitative and quantitative indicators should serve the project. Ultimately, the goal is to determine whether an idea has convinced enough people to take it forward outside the platform. It can also be used to identify users who are likely to join a project team. Posting a comment is fine, but is the employee ready to get involved operationally? This is also a measure of the degree of relevance.

Can we create an idea management software without artificial intelligence?

On a technical level, yes. But would I recommend it? Not at all.

Doing without AI without falling into the suggestion box system that ultimately provides almost no value requires putting people behind the system. There are generally two main levels of action: firstly, there are referrers who can sort out the ideas, and secondly, there are experts who can become involved in the project. The connection is made manually, and it necessarily takes time.

I have already tested an idea management software without artificial intelligence and, to give you an idea, on the basis of 120,000 employees, you need at least 30 referrers and between 3,000 and 5,000 experts who can be called upon. On the other hand, it can be noted that human intermediation is more efficient because it is less automated, but the network of experts is difficult to maintain over time. It eventually loses its reliability and relevance due to the problem of renewal. Each time a referrer wants to ask an expert for input, they have to find the right one, make sure they are available, send them the information, follow up on the case, etc. The cost of doing this is very high, and the difficulty of getting the right ideas to the right people can be crippling.

AI-powered decision-making assistance saves time

Where it takes 6 months for an idea to be tested by humans, with AI, it can take 6 weeks. In fact, the limitations of the human model are well known. With human decision-making, the opinion may or may not be relevant, depending on the context, the mindset, the level of knowledge and the current mood. With AI, the approach is collective and objective. A bad idea will be quickly cataloged. Does this mean that AI can take the place of humans in sorting, selecting and linking ideas? No, because there is no perfect system. You have to accept that a great idea may fall through the cracks of the community, just like the fact that an individual is fallible and that their personal biases may lead to a bad decision on a subject that does not interest them, but that could prove to be highly strategic for the company.

AI and human intelligence combine

We are at a time that I would describe as a “crossroads”. It’s not either one or the other; it’s one at the service of the other. The machine is there to assist man. It does not make the decision, but provides insight and participates in the decision-making process.

Especially since it is important not to throw away or delete anything. An idea that doesn’t make it into a project today should be stored and kept for possible later use. It is common to see an idea emerge 12 or 18 months later, after it had been buried. It may regain favor in the eyes of the community due to a change in strategy, context or due to a change in the market. That’s why archiving and documentation are critical to reviving an old idea in seconds. On the other hand, if someone submits a new idea that joins an already archived contribution, the platform can reactivate them in order to test their relevance again. It is also useful to avoid duplication: it is better to collaborate on the same idea than to work in competition. And for that, AI is essential.