Last week I was lucky enough to spend a few days in Dublin with my colleagues. I spent Monday at Microsoft’s (very nice) offices, meeting members of the Microsoft sales team. A team we interact with a lot but rarely get to meet face to face, so it was nice to spend some quality time with them. This was followed by a thoroughly nice meal in Dublin city centre and a few drinks in Temple Bar. We then spent Tuesday working together in Dublin, something we don’t often get to do as a marketing, sales, and presales team here at Infinity Group due to our distributed geography these days.
On Wednesday and Thursday, it was time to attend the European Power Platform Conference 2023, an event dedicated to Microsoft’s Power Platform technology stack, with 3 keynotes, 6 full day tutorials and over 90 specialist sessions. With such a great lineup the first challenge was deciding which sessions to attend!
In this blog, I’ll pick my top 5 highlights from the conference.
Power Virtual Agents and AI_
The Power Platform 2023 Conference opened in style with some traditional Riverdance, followed by an opening keynote from Charles Lamanna, Corporate Vice President, Business Applications & Platform at Microsoft. Charles’ keynote covered a lot of content around the direction of the Power Platform and some impressive stats. The session highlighted the success of the platform’s low-code development journey and some of Microsoft’s ambitious plans to grow user numbers in the not-too-distant future.
One highlight for me personally from the keynote, was around AI Integration with Power Virtual Agents. We were shown a demo with two impressive features, the first was the ability to boost your bot by sharing information with it such as a SharePoint site or public website. This facilitates the ability for the bot to answer questions from a user using the context provided by these additional data sources without the need to author specific topics (this is game changing!).
I had to give the preview feature a go, so I set up a bot in a development Power Platform environment. I went on to browse through our company documents via SharePoint and started asking questions required to fill in a tender response for a prospect — such as whether we were ISO27001 certified and when our certification expired, or whether we had a business continuity plan. The bot cleverly answered everything accurately and speedily without the need to author a single topic.
What followed was even more impressive: A demo bot where a couple of custom actions had been set up (one to retrieve booking availability on cruise ships and the other to check the weather in a given destination), using preview capabilities was able to extract information from the questions to dynamically populate the required input parameters for the custom actions. With AI functionalities, the bot was able to call and present results back to the user; offering available slots for cruises in their destinations of interest and weather readings.
Where any inputs were missing, the bot was able to understand this and asked for extra information before executing the required action. Again, this was all achieved without authoring a single topic manually.
Automated Machine learning (AutoML) in Power BI_
A very interesting session on utilising machine learning within Power BI was run by Andrea Martorana Tusa, Microsoft MVP – Product Manager in Pandora. It included some background on the features along with interactive demos. Available as a premium feature for dataflows, AutoML within Power BI covers the following types of models-
Binary classification: This gives a ‘yes/no’ output based on input data, for example whether a customer is a credit risk or not.
Classification: This allows the model to distinguish between three or more different outcomes, such as determining the category of a product.
Regression: This is used to estimate a numeric value, such as the % chance of a customer leaving.
Implementation is as simple as following a wizard within the Power BI service, with the need to have data to hand, to train and test the model. Once setup, the model can be evaluated, customised and retrained until its optimised; before being applied to your live data set(s). Once the model has run, the finished data is returned along with ‘explanation’ columns and tables to describe the reason for the ML model’s decision.
Power Platform adoption success stories_
The Power Platform 2023 Conference was aimed at Microsoft partners and customers alike, so there were multiple sessions run by end users of Power Platform to explain how they had achieved success.
One such session, run jointly by Microsoft and Defra, explained Defra’s journey to its Power Platform adoption. The organisation of 30,000 staff which started with no adoption, governance or centralised management, now has all of those things in place, along with a well-defined Power Platform strategy. Listening to the session it was clear that it was not an easy journey and the organisation had invested a significant amount of time and energy to get to where it is today.
Defra now have an ecosystem of app makers spread throughout the organisation, working on a centralised platform with clear guidelines in place for app innovation and development, governance, security and compliance. End users throughout the organisation have the tools they need to solve complex business problems in an agile way, creating apps that add real value to the organisation, with a process in place for the app to be approved and distributed to the wider community.
The important of a Power Platform environment strategy_
There was also a very interesting session on the importance of Power Platform environment strategies. It highlighted the positive impact a comprehensive approach could have on your organisation.
Quite often, usage of the Power Platform starts organically within organisations as curious users play with the tools available within the platform. Typically, they will have access via their Office 365 licencing, so there are no barriers to building apps and flows. Whilst we would always encourage this innovative mindset and problem-solving attitude, it’s important that these users are given a secure environment to work in.
For example, without the necessary governance in place, a Power Automate flow written by a user, could read confidential customer information from an internal SQL database, and may accidentally post the results to LinkedIn or Twitter! Fortunately, the Power Platform has a set of controls and data loss prevention policies, which allow administrators to ring-fence certain connectors so they cannot be used in the same flow, thereby preventing the scenario of reading internal data and sharing it externally.
With the correct environment strategy, you can also support your users by giving them safe environments to build in, coupled with defined approval processes to deploy changes to production environments for apps that are going to be used by wider teams.
The session also covered other key topics such as scalability, administration, lifecycle management, standardisation and consistency and licence management.
Impacting the community: Power Platform 2023 Conference_
Lastly, and perhaps most importantly, it was amazing to see the number of passionate Power Platform enthusiasts from Microsoft, their partners, and customers in attendance at the conference. With over 1,200 attendees, compared to 350 the year before, the growth of the Power Platform community is impressive.
It’s reassuring to know that as the platform is rolled out across more and more organisations, there’s a huge group of passionate enthusiasts ready to help. There were over 90 sessions run by individuals who had given up their time to create content to share with their peers to help spread knowledge. All the speakers stayed after their sessions to answer questions from members of the audience, to assist them with practical challenges in their businesses.
This blog was provided by our Chief Innovation Officer, Tristan Shortland.