Artificial Intelligence to Improve Project Communications

< Go Back
How can we demonstrate and use AI in our domain of project management to improve the quality of our communications outputs?
20 October 2021 - ERNACT  project
Dr Margaret Quinn | ERNACT Programme Manager

Artificial Intelligence, AI, is a high-profile subject so well-known though perhaps not so well understood. As a disruptive technology it has the potential to transform speed, safety, precision, effectiveness, greenness of our efforts. It is already having an incredible impact in many areas ranging from Medicine/Healthcare, Industry/Manufacturing, Finance, Environment, Transport ..across the whole of society. Yet its uptake appears to be slow especially for small and medium sized organisations as noted by the Advanced Technology for Industry report 2020. So, what is the issue? 

There are multiple factors and perception plays a role. The idea that AI is very difficult, it is only suited to big companies, that it is very expensive, that you need a lot of data to do anything with it. This may not be helped by the images you see in articles connected to AI which tend to be futuristic, holographic brains and sci-fi robots. Major barriers include having an understanding of how to use it, availability of skills to implement, high costs, trust issues and governance, data availability.

In ERNACT we typically lead projects which aim to increase the uptake of disruptive technology in our region. So rather then just preaching, we thought we should try and practise using disruptive technologies ourselves. The challenge was set: how can we demonstrate and use AI in our domain of project management to improve the quality of our communications outputs – in time for a demo at our EU Regions Week workshop.

We may not be aware but most of us are already using AI in our daily lives, via spam email filters, recommender systems in online systems. Also in daily communications for projects, there are a range of AI based tools such as grammar/spell checkers, predictive text, plagiarism checkers to name but a few. We can use AI to help us produce well written text but the text may not be very relevant to a project’s communications objectives. This is where we decided to focus: to try to demonstrate the AI can be useful to assess the relevance of project communication outputs. The challenge though was that this needed to be accomplished and demonstrated in a short time frame of a few weeks, in time for Eu Regions Week, without AI experts or programmers and without internal big data. 

Having abandoned a few unsuccessful routes, we simplified the goal to checking that a set of key project topics are contained in project communications outputs such as project news items. We opted to use the Microsoft Power Apps platform which provides an AI builder and pre-built modules. This allowed us to make a topic model for a project without any coding, just configuration, following the easy step by step guide. There is no need to have in-depth understanding of the AI or Machine Learning supporting the module. The approach was hackathon style, iterative and collaborative. We trained an AI model on a set of project topics with the DIGITAL REGIONS Interreg Europe project as a test case (“SME”, ”Industry 4.0”, “Manufacturing”, ”Regional”...). We then created an app to be used for demonstration purposes, again using the Power Apps platform and lo-coding only, to run the AI model on a set of news items. The output is a confidence score that each news item is about the target topics. So happy days! A demo was produced in a few weeks. On the short journey we learnt about using the topic model by using enough training data, being consistent and balanced in labelling of training data and starting simple (fewer topics).

The conclusion for me was that there are pre-built AI modules that are easy to use without coding and that these can be used in the project management domain of improving project communications. We foresee using this in daily activities using the topic checker each time we produce a communication output such as a news item for a project. We will need to add new topics that arise with new projects and to re-train the project topic model. To this end we plan to train our programme /project managers and communications team.

The next steps in ERNACT are more experimenting to increase the performance of our project AI model and testing on more projects, then training our team and embedding the AI checker tool in our project management platform, Teamworker, which we use to produce project news items.

My recommendation is to have a go, experiment! For re-assurance, we all have local centres of expertise who can provide great guidance on AI. In my case I thank Agustin Garcia Periera of Insight, NUIG Galway, as encourager and reviewer. To gain more understanding of what AI is about try taking an introductory course such as 'Elements of AI'.

Click here to email  Ianire Renobales at ERNACT for further information

Project Logo Project Graphic

Other Stories

€15 million, 17 projects, 15 digital tech pilots, 25 action plans and 173 good practices
25 November 2021
ERNACT regions and cities meet virtually to capitalise on the increased digital transformation opportunities in the new EU programmes.
more »
"Industry 4.0 is doing a lot towards green growth"
24 November 2021
ERNACT participates in Interreg Europe's 'Europe, let's cooperate!' big annual event.
more »
ERNACT gets ready to capitalise on the upcoming digital transformation opportunities
19 November 2021
Core network members will virtually meet to agree a firm project proposal development for the next year & discuss membership with interested regions.
more »
EU Intelligent Cities Challenge – A Year of Initiatives, Collaborations, and Initial Success for Derry
18 November 2021
Saverio Romeo, the Lead Expert appointed by the ICC programme, about Derry's progress towards achieving its objectives.
more »
Core ERNACT regions
Newfoundland & Labrador
South Ostrobothnia
West Romania
San Sebastián
Derry & Strabane
United Kingdom