Reading time: 4 min
There is a gap between the willingness to listen to collaborators or customers and the reluctance to ask simple unbiased open questions. Semantic AI is a powerful tool to bridge this gap and fully leverage the power of open questions.
You are not listening
I was getting impatient. I had asked my co-founding partner a simple question (about some commercial options). I had laid out the situation pretty clearly, named the major options and ask him about his advice: "in your opinion, how should we approach this?"
And he was not answering. He was babbling on life, death, human nature… and I was wasting time listening to his digressions that didn’t really help me.
"You are not answering my question" I said, annoyed.
Are you open to consider an answer that does not fit in the preset list of what you expect?
"Well", he replied, bluntly and calmly, "it is because you are not really ASKING my opinion. You want me to pick one of the options you have defined. Doing so, I compel to the the way you analyze and break down the issue. Except that I have a different angle. There is no point for me to choose between options that I don’t feel relevant. What I am trying to tell you is that there are other ways to grasp the issue… but you are not ready to hear it. So you feel I am not answering your question, whereas in fact you are simply not open to consider an answer that is not under a preset format."
I was speechless. He was right.
Every conversation is not a problem to solve
In my previous business life - I did spend more than 20 years in management consulting - I had acquired the most efficient way to solve complex problems:
Break it down into smaller manageable parts. Make a list of MECE hypotheses. Go searching for the facts to assess whether each hypothesis is valid or not.
Pyramidal thinking is not just a gimmick. It is incredibly efficient. Because you split workload and delegate, leave no stone unturned and always argue on facts rather than opinions. So powerful you use it for everything.
But every conversation is not a problem to solve. Coming with hypotheses and using people as "fact checkers", asking them to help (un)validate your hypotheses, comes with a major drawback: you are not really listening, rather just using them. You are bound to miss relevant points: maybe they don't agree to the way you breakdown the problem. Maybe things outside your box - matter to them. Maybe it makes a difference that they say things the way they want to say it.
The downside with asking with hypotheses and using people as "fact checkers" is that you are not really listening
And it’s fine as long as your sole leading metric is efficiency.
But don’t expect this to help you actually learn something you don’t already know, nor make people feel better about expressing themselves.
Asking without prejudice
We have been helping for the last 10 years managers to leverage practical and meaningful collective intelligence for remote teams. By enabling collaborative, efficient asking (and use semantic A.I. to capture insight).
And the importance of listening carefully to all the stakeholders of importance: collaborators, customers, citizens… is now well established as a management best practices.
As a consequence, digital tools to ask questions and run conversations online - on collaborative platforms or on a simple online questionnaire - are spreading to collect authentic feedback, establish dialogue, build buy in.
And part of our daily work is to coach our clients into asking simple unbiased open-ended questions. Because they often tend naturally to influence respondents in imposing a predetermined analytical grid, openly, with a multiple choice or more subtly like in the following examples:
One client was once performing a full visioning exercise with a broad group of top managers, scattered across the globe. He wanted to hear them about growth opportunities.
I have prepared 3 open questions about growth:
- What opportunities of external growth?
- What opportunities of geographical growth?
- What opportunities of growth in integrating new businesses?
People are all the same. You can’t nudge them to fill out your spreadsheet
The rationale was clear. Leave no stone unturned. Cover all the options. Collect the maximum information.
But people are all the same. You can’t nudge them to fill out your spreadsheet. If you hand them a microphone, they will tell you what they have and want to say about it. In this example, they will put all their answers into the first question … and wonder why you ask the same questions three times. Because, in fact, there is only one open question. As far as getting insights was concerned, breaking down the questions had only downsides: People would get bored, feeling they were repeating themselves. Some 'good pupils' would force themselves to provide ideas they were not fully endorsing... and in the end, you would miss on evaluable insight: some options arenot considered spontaneously by the management!
The power of open questions lies in the ability to ask without prejudice. In a questionnaire, after Yes/No or a multiple choice question, it boils down to a simple "Could you elaborate?"
The flip side and the remedy
Naturally, people tend to ask close or multiple choice questions for a very practical reason: efficiency. Because it generates more analysis work, because it might introduce unexpected things that will create again additional work. So asking true open questions is much more challenging. Because it leaves you with the burden of analyzing hundreds, thousands of text answers. As a result, there is today a gap between the willingness of the organizations to nurture open dialogue... and the ability to process and extract insights out of these conversations.
We strongly believe that semantic AI is now able to close this gap. Not as a black box, but rather as a super-powered assistant that will perform semantic tasks automatically under the guidance and the control of the analyst.
Semantic A.I. can be the super-assistant that performs super-efficient analysis of these open questions
As an example, our latest import console is able to import any qualitative data and perform the following tasks:
- Auto-detect tags such as keywords, locations, names, organizations
- Suggest content for user-defined topics, based on the meaning of answers
- Identify spontaneously consistent groups of answers topics, of similar meaning
encapsulated in a simple interface, enabling filtering, searching and counting, in 50 languages.
Full analysis of a few hundred of comments can now be done in less than two hours, as opposed to a few days.
There is no longer a downside to asking open questions.