Whether you’re a startup or an enterprise, developing AI products is challenging.
Not only do you have to wrestle with the challenges of finding a use-case that where AI can actually deliver value into an enterprise workflow, but you also have UI concerns, and – often – much higher demands to monitor algorithmic drift and other technical issues.
This article will be the first in a three-part series with a focus on AI product development. While this article will cover the ideation process for AI products, the second article will cover the AI product development roadmap.
We advise our clients to think about developing their product in the short term in order to work towards long-term goals. Asking the following questions during the product development stage will help generate the best decisions for AI product development.
Each short-term idea should be seen as a step towards attaining a long-term goal – mirroring our advice for anchoring the strategic priorities in any near-term AI initiative in an enterprise.
We’ll begin by sharing questions for prompting AI product ideas with long-term strategic value for your desired users:
Long-Term AI Product Ideas
Ask “Where is My Industry Ultimately Going and Who Will Ultimately Win?”
When our clients are brainstorming long-term AI product ideas, we recommend they consider the future environment of their business sector. The most important factors to consider are what the most successful initiatives are, and which types of data will be most important to making that possible. This could be considered a proposed explanation of who you think will “win” in the future of your business sector.
- A CRM vendor for high-touch sales might speculate that allowing sales teams to see a visual map of an organizational chart would be most mission-critical to improving sales results in the years ahead. This could enable them to more quickly prioritize the right targets within that organization chart in order to save time. Additionally, these visual maps could help them take the right steps to ensure they close the sale.
- A hospital or healthcare network may suppose that the leading healthcare companies of the future will have access to automated electronic medical records (EMRs) systems. These systems offer medical transcription via AI voice recognition and automated coding of injuries and illnesses according to their classification. In the future they could automatically update and save new records to the network’s database, and separate individual fields within the records to save as individual data points.
Ask: “How Will These Future Winners Move Forward Into the Future?”
This question is designed to answer where you think your sector is moving in the long term. The “winners” of the future will be those who can take the most data points into their software that are relevant to the most important success factors. The more relevant data added to a machine learning module the better a module can become, thus it is important for a company to identify which data (and features) they want to be collecting as early as possible.
- For the CRM vendor, the company that knows when the sales agent made a phone call and what they talked about will be the winner. The company that can find the right ways to take in and utilize data most effectively will be the one that can best prompt salespeople about the next best step.
- For the healthcare network, the company with the fastest transcriptions and deliverance of EMR data across providers will win. Those companies will have an even stronger advantage if they can automate requests for specific details about patients without the need of providing their entire record.
Ask: “How Will AI Enable These Future Market Winners?”
Once you have established where you think your business sector is heading in the future, you can begin to consider how AI can help future leaders move in that direction. This involves collecting the most important types of data for the business areas you have already supposed to be the source of the most future success.
- For the B2B CRM vendor, this could be an AI application for predicting the lead score of various people within an organizational chart they are working with. It might use data from the organizational chart to suggest next actions for the sales agent. It could also utilize historical data involving selling to similar accounts and then suggest actions based on that. The vendor could see this as AI helping clients discern when and how to reach out to potential buyers.
- For the healthcare company, this would likely be an automated EMR application that is compatible with multiple interfaces so that the information is accessible no matter where patients go for care. It would be equipped with voice recognition technology and would require live recording data of healthcare providers speaking medical notes into a microphone.
Ask: “What Aspects of My Product Would Make it so Customers Would Never Buy From Anyone Else? What Data Can Get Us There?”
Now that you have brainstormed the data necessary to enable customers who will go on to be leaders in their own sectors, it is time to consider other selling factors of your product. These include price, features, quality, experience, and the value your product offers. Before brainstorming on new additions, consider which of these aspects would make it so that customers would not want to buy from anyone else. It is important to find a balance between these aspects when considering ideas for the long term.
Consider the data that will allow you to deliver that product to your customer base. This will usually be the same types of data that you have already brainstormed future leaders will use to drive customer satisfaction. The capability to incorporate the most important types of data customers will need to use in the future may play an important role in the success of your product.
For example, AI vendors could help businesses in the HVAC industry deliver the best possible product in terms of price, features, experience, and value by using energy and sustainability data from past buildings their clients have worked on. Identifying that energy and sustainability data before anyone else is the race toward what we call “data dominance.”
Suppose Amazon were brainstorming how to continue to monopolize the business of American eCommerce. They might come to the conclusion that they want more quality products to offer. They want to be able to recommend the best products to their customers, have the best search function, and to deliver items to customers faster. AI plays a role in all three of those areas. It may not help them directly recruit more customers, but it helps potential customers find what they are looking for faster and with logistics and delivery.
When these questions are answered well, a company can collect the right data to make their product better than their competitors. When their product is better, they are able to keep collecting more data, which raises sales and thus generates more data and positive change in their product. This type of snowballing effect is called Data Dominance, and it’s exactly how Amazon has better recommendations than its competitors. Amazon has the data dominance to always get more purchases.
Near-Term AI Product Ideas
Ask: “Where Can AI Capabilities Deliver More of What Clients Want?”
An effective way of brainstorming near-term ideas for AI products is to consider what your business already offers to clients or customers and how AI can deliver more of it.
- For a fashion eCommerce store, this may mean a wider variety of products or better recommendations.
- Banks may use AI to implement facial recognition as an authentication option for their customers in lieu of a pin number.
- A trucking and fulfillment company might want to use AI to provide more accurate arrival dates.
Ask: “Where Can AI Fit Into the Core Initiatives of My Company?”
Another way to think about near-term ideas is to focus on the core initiatives of your company as a whole and identifying which of them are compatible with AI. This framing can help you get a wider view of where AI can be applied so that you can focus on areas where it would be most beneficial.
For this exercise, we use the example of a software company that sells paperwork and workflow management solutions to insurance companies. This involves creating an organized dashboard for completing paperwork and emailing it to all necessary people. This company would then ask themselves, “Where can AI fit in?”
One possible answer to this question is that there might be a way for AI to detect user activities and prompt them with a recommendation. This could be a simple recommendation of a next best step or it could link to learning materials to help the user better understand the software.
Additionally, the AI application could be made to source data from the client’s network about how the software is being used. This would allow it to scan for user error or any struggling employees. These types of automated prompts could form a robust feedback process that helps internal teams refine and improve the AI product. The prompts not only relate to the product, but also to existing business initiatives.
It is easier for larger companies to leverage AI in areas that their management has already approved of. If they understand that this type of innovation is happening within a business area that is already successful, it will be easier for them to invest in the required AI technology. Within larger companies this allows the company to permit AI investment without going directly against the grain of existing company priorities.
Existing business priorities are usually well thought out and revised over long periods of time, and there is no need to try and reinvent them when proposing an AI project. Instead, it is more effective to focus on how AI can enhance core initiatives and provide a better customer experience.
The Connection from Short-Term to Long-Term:
Drawing a Straight Line Between Your Short Term and Long-term Development Efforts
Connecting your short-term goals to your long term goals is perhaps the most important part of brainstorming new product ideas.
This will help you find a clear throughline showing how each short-term project will push your company closer to achieving its long-term goals in AI. The easiest way is to start with listing all of your long-term goals, and then listing short-term product ideas that might help the company move toward those long-term goals.
Ideally, the near-term products will put your company on track to succeed more in the future. Consider every possible way you could implement AI that furthers these long-term goals, and try to add as much detail as possible. Companies should be able to see a direct connection between what they are accomplishing in the short-term and where they want to be in the future in terms of market success.
Each short-term idea will likely require a fraction of the data required to achieve a long-term goal. It is important to consider exactly which types of data enable each idea. Gathering this data for both a short-term and long-term goal would drive efficiency in your AI initiatives simultaneously. At this stage, companies should be able to articulate how they can achieve their long-term goals and how their ideas for more immediate products helps them do this.
Building Core Capabilities
All of these brainstormed AI initiatives would undoubtedly help your company build core capabilities that facilitate more efficient work with AI applications. It is important to understand that gaining experience with AI is a benefit alone. Companies could gain experience in AI by starting small and only learning the skills needed to create AI even if the product does not generate a lot of revenue immediately.
The following are critical capabilities for any AI initiative:
- Data Infrastructure: Creating a clean and accessible data stack that can be used for Data science projects and training machine learning models.
- Data Science Talent: Hiring a data science team with the skills to facilitate short-term projects and ensure movement towards long-term success.
- Project Team Construction and Team Dynamics: How to get subject matter experts and data scientists to work together in functional AI teams.
- Building Towards Data Dominance: Thinking about the proprietary data necessary to pull in clients and thus more data.
Your short-term and long-term plans should be reviewed annually in order to ensure they still make sense together. Many companies will be limited if they only think of AI as a means to an immediate financial end. But AI is an entire paradigm shift that will help companies succeed in the future if they can establish these core AI capabilities.
Read the other two articles in this series: