Configurable collective intelligence & no-code

ALC moves from informal exchanges to the possibility to configure collaboration to get smart results more efficiently

The insufficiency of informality

Our current digital collaboration tools privilege informal exchange which can severely handicap group intelligence. Since ever we favour informal exchange – talking and chatting. According to global historian Yuval Noah Harari, language was invented to gossip. We have essentially kept these habits for our face-to-face meetings and have extended these habits to our digital collaboration.

Often “natural” communication is however deeply biased and ineffective. It is inadequate to collect contributions in parallel. Also, it is uneasy to progress stepwise and in a focused way towards common purposes. The typical real-time applications for this purpose create a lot of noise and render cumbersome even simple tasks such as finding a common date.

This does not mean that today we do not already practice structured interactions: we fill forms; we vote online; we use business process management applications. These applications tend to focus on form-based data collection and / or application specific rigid interactions. There are simple tools such a list fillers (e.g., a la Google Keep) or finding a common date (Doodle, Calendly etc.), template-based electronic whiteboards, e.g., for brainstorming (such as Mural, Miro or Klaxoon). And collaboration platforms such as Teams or Slack have integrated / connected to simple polling and form filling apps. Information collection is often rigidly set by specific “solutions” which collect biased inputs. As soon as we ask for more than a bit of basic data collection and form filling, these current solutions turn out to be functionally poor; they are inadequate for many more complex collective purposes.

Configurable interactions to get smart results and save time & effort

To cope with complex collaborative undertakings, we need to propose effective means to organise “collective intelligence” i.e., to smartly coordinate individual intelligence, in view of common purposes. To mobilise intelligence, collaboration dynamics need to be managed effectively, and as a whole. Sometimes contributions need to be collected anonymously – to ensure that participants do not feel socially pressured – sometimes nominatively, to catalyse mutual stimulation. Sometimes results should not be shared before everybody has contributed, to ensure independence of opinions; sometimes it is better to show what is “on the table” in the real-time, so that everybody feels to compelled to contribute more. Sometimes a simple vote is better; sometimes ranked votes or auctions with tokens from a limited budget are to be preferred, e.g., to allow minority preferences to emerge in repetitive situations. Science tells us that “collective wisdom” often requires that we compensate social and cognitive biases. How we collect inputs from groups affects deeply the symFlo outcome: asking anonymously or nominatively; sharing results as they are collected or once everybody has contributed; giving 2 minutes, 2 hours, or 2 days; single versus multiple votes; etc. The “best” scheme depends on the aim, on the group, on the context…. There are thousands such issues. The right structured interactions can free a lot of energy and reduce the duration and overhead of collaborative processes by orders of magnitude (the wrong ones can just be painful).

At SymPlace we are working on a no-code collaboration modelling language relying on configurable ALC building blocks, and the means to connect them, to create sophisticated symFlos combining them for larger purposes.

Use case: Asking a group for advice

Many of us know this situation: “client” AA asks a group for advice on a somewhat difficult perhaps somewhat controversial subject – by email, in a meeting or a social media group – and receives nothing or on the contrary triggers a rather confusing discussion, which leads nowhere. Of course, AA can ask each of her consultants individually, but that is time consuming and then the results still need to be grouped and consolidated. Perhaps she wants to ask several questions, some of which call for a fast response; inversely, others will require a deeper exchange, depending on whom is asked. AA could organise a survey which is sent to each person individually, but surveys do not allow for interaction.

An ALC-enabled process involving symFlos could obtain a more inclusive and smarter result, as follows:

  • Participants receive an invitation to the symFlo, with an explanation of how they will be invited to contribute advice, and what they receive in exchange. Accepting implies agreement with the terms of the consultation.
  • They receive a short summary which they acknowledge upon request from a symBot.
  • They can ask clarification questions which they get answers to.
  • They submit recommendations anonymously and without seeing each other’s inputs before everybody’s contribution is submitted.
  • (Only) Now participants provide feedback on each others' inputs.
  • In parallel, an informal discussion explores the bright and dark sides of each recommendation.
  • “Client” AA closes the consultation thanking the “consultants”.