By Phil Spalding

cambolc.co.uk,  LinkedIn, BlueSky, Facebook

In the last blog I talked about the different types of AI and how content creators were starting to group them in a taxonomy. Here we are going to start the process of grouping the types of AI. This particular blog is going to start the conversation about Productivity and AI.

First what do we mean by productivity. I am going to use the definition that it is the amount of work in a given time. For instance the number of customer enquiries handled per hour could be one measure. Would the use of AI for frequently asked questions be appropriate? How are we going to communicate with the AI Agent? Is it going to be text based or is it going to be speech recognition in a chatbot?

On a simplistic level the questions asked about how the example above could be approached are throwing up some quite large questions as to how the AI would work. If there is to be any answers that are generated then a Large Language Model (LLM) may need to sitting somewhere to interpret the question. There needs to be a mechanism by which the answers are checked and refined. Each company or organisation will have it’s own expert or system vocabulary. A general LLM might be needed to be tweeked so propriety knowledge is not released out into the wild of internet. If an answer is not immediately forthcoming the interface then needs direct the customer to a skilled human operator. This relies on the customer either accepting or rejecting the proposed solution.

Making the switch to an AI driven workplace might seem attractive from head count related productivity but how much can you trust that the system you employ is good enough for that task. What is the equivalent development costs compared to the person hours that you are no longer utilising? How should you adopt the new system? Trial it in one department or make it organisation wide from the beginning? These are all considerations that have been tried and tested with every new innovation that has come along. We need to make sure that the shiny new AI is the best thing since sliced bread for the particular problem that needs to be solved.

So what is out there?

10Web – A site that will design your website using AI tools for you following a prompt of what you require. Combines I would imagine at first glance a number of AI technologies including what would be a LLM and a code generating AI. Agentic AI would also probably be another layer to handle customer queries?

Balloonary – This is an Ad producing AI that will apparently allow your products to take off.

Bright Eye – Is on the Apple Appstore and is described ”  a multipurpose, generative, gaming and analytical AI app” on a mobile. Being an android phone user I will not be trying this one out.

Chatfuel – This is a sales assistant from a Meta partner. This is therefore presumably a component somewhere, of the Amazon linked part of the digiverse.

Chatsimple – A sales and support assistant you can add to your website.

Dialogflow – A Google Cloud service which uses deterministic and generative AI to allow the user to produce customisable agents for interactions involving customer conversations. Has a large development community.

Pebblely – A photoproducing AI.

Sortext – An AI that converts first time visitors to customers using E-commerce focused AI to make offers tailored to the site visitor.

Wix – A website building AI. Wix has been around a long time (since 2006) and has been through the stage of WYSWYG online website builders.

The above are a few productivity AIs. The majority have a pay model for any serious use so it is a little difficult to try before you buy in some cases. A few may be placed in the wrong category but as with all classifications they suitable for revision with more experience of their use.