H4MoD Problem Sponsor Interview: Flt Lft Peter Kennedy

We caught up with Flight Lieutenant Peter Kennedy to talk through his experience of being a problem sponsor on H4MoD. He recently sponsored a Hacking for Ministry of Defence (H4MoD) problem at the War Studies Department, King’s College London looking at multi-domain operational planning. His team came up with a solution that will enable planners to collaborate more effectively across multiple locations.

H4MoD is a postgraduate programme in which teams of four to five students learn and apply the lean start-up methodology to understand and solve national security and defence problems in 10 weeks. Each team is given a different problem, each with its own government Problem Sponsor. The Sponsor is someone who understands and has experience of the problem.

 Could you talk us through multi-domain operations and why they are important?

Multi-domain operations come down to how we plan military operations that span two or more “domains” which can include: air, land, maritime, space and cyber/electromagnetic. We’re getting better at this, but we still have a long way to go.

The UK’s Doctrine and Concepts Centre talks about our principal adversaries operating in all five domains in order to undermine our ‘cohesion, erode economic, political and social resilience, and challenge our strategic position in key regions of the world’.  The UK military has recognised the threat, however, we have a tendency to remain fixed in our own ‘domain’ siloes when it comes to planning and our ability to transition between domains as we plan operations needs to be more efficient and flexible.   

Could you tell me about the problem that you submitted?

The problem that I originally submitted was: RAF front-line operators need a more efficient and flexible means of planning front-line operations, in order to manage the increased complexity that will result from the incorporation of multi-domain operations. 

When I submitted my problem, my initial thinking was that the key pain point was an inefficient planning process that wasn’t agile enough to respond to emerging threats and opportunities.  For example, looking at the air domain alone, the air tasking cycle can take up to 72 hours. From my perspective, this was a problem of rigidity in the top-down planning process. However, as the students interviewed more and more stakeholders, several common frustrations emerged. Time and time again, they heard about the lack of common data types and flawed assumptions leading to flawed plans. This included the lack of a common language and a baseline for planning operations, including time zones, data types, units and so forth. Further, the students discovered another key pain point in the planning process, namely, the (in)ability to collaborate remotely with individuals dispersed across different locations.

Could you talk through the solution that your student delivered on the course?

Addressing the pain point around collaboration, the team initially focussed on ways to adequately translate the “whiteboard” idea into an online environment so that dispersed groups/individuals could work in tandem on one unfolding plan. 

Usually when everyone is physically in the same room, operational planners can use a whiteboard to plan and aid the visualisation of the unfolding operational plan. If people are not physically together, then words alone are often not enough to be able to effectively communicate complex plans. Further, if you are copying whiteboards into different “rooms” or across different bases - this complicates the ability to collaborate effectively.

As the team started to look at how to extend this whiteboard idea across multiple locations, they discovered a piece of software, called Widow, that the US Air Force is currently using and developing. This software allows core assumptions, equipment constraints and the like to be added. For example, you could enter in key constraints related to particular air frames such as their maximum speed, landing mass and so forth. Such would help to mitigate potential flaws in plans related to erroneous equipment assumptions. 

So the team’s final Minimum Viable Product (MVP) was an online dashboard which operational level planners would populate with assets, capabilities and a broad overview of how they would like a plan to play out. Like Widow, this would help mitigate flawed planning assumptions whilst helping to forge a common language and set of baseline assumptions for planners at all levels. Unlike Widow, the team’s dashboard would be applied to all domains rather than just Air. Testing this MVP with end users at Permanent Joint Headquarters, they found that it resonated strongly. 

For their final brief, the team presented a depiction of what the planning dashboard might look like, with users being able to manually enter details (assets, objectives, constraints, etc.) which would then be visible to those granted the appropriate permissions. Many of the stakeholders who were interviewed throughout the project suggested that this dashboard could draw from existing databases to automatically populate asset capabilities, or to automate parts of the planning (scheduling, deconfliction, etc.).  However, the team recognised that this functionality would rely on building a much simpler platform in the first instance, before any more sophisticated functionality could be added. It's for this reason the team decided to focus on getting the basic requirement of the dashboard right first.

Great to hear that the team’s MVP resonated strongly with the end users. So what happens next? Are you intending to implement what the students delivered?

Next week I’m talking with the Defence Science and Technology Laboratory (Dstl) who are working on a similar problem. I’ll be talking with them about how the work that the team has done can plug in to what they are doing. I see the dashboard that the students came up with as being the interface between humans and systems which can support decision-making. I think up until now, we have focussed more on the back end “system” without thinking about that interface.

I am also continuing to work with the student team and we are meeting up at the end of this week. In fact, I am currently trying to figure out if I can get the students security clearance whilst also investigating the possibility of bringing them on board as consultants to the MoD to help with implementation.

Would you encourage colleagues to submit a problem to H4MoD?

Yes, absolutely. My student team which included: Alice Merriman-Jones, Brayden Beckius, Matilda Grow, Alyn Hoefer and Nolan Morrison exceeded my expectations. They were professional, keen to solve my problem, and just excellent.

One of the best aspects of the course was getting a completely different perspective on my problem. The military can feel like a bubble sometimes with people recycling the same ideas and arguments. However, on H4MoD I had access to a group of students from a completely different range of backgrounds. Consequently, they took a very different approach to looking at my problem and took it in directions I had not previously considered.

It was also just good fun getting to interact with the students. I had the opportunity to chat through a problem that I find intrinsically interesting and to bounce ideas off a diverse group of people.

Overall, I will definitely want to sponsor another problem in the future. 

Are there any particular lessons that you would highlight for future problem sponsors?

Your student team will be putting in a lot of time and effort to look at your problem. In return, you need to be willing to commit to the process and give them your time each week. With my team, we set up two calls a week - these were short and sharp updates to ensure that I followed the team’s progress. I also chose to attend my team’s weekly presentations. That was not something I had to do, but I found it really useful to see who they’d been talking to and to understand how their thinking was developing.

My other key lesson would be to avoid solutioneering. It is something that we are pretty bad at in the military and I know that I was initially guilty of this with my team. I became aware early on that I was trying to shoehorn in my own objectives as to how I wanted it to go. At that point I decided to take a step back and let the team freely conduct their discovery and see where it took them. To be honest, that was the best thing I could have done as the students ended up going down paths that I would not have considered.  

That’s a great conclusion, thank you Peter for sharing your experience of being a Problem Sponsor with us.

Flt Lt Pete Kennedy has a background flying the C130J for the Royal Air Force. He is currently studying a MSc in Artificial Intelligence with Cardiff University sponsored by one of the Chief of the Air Staff's Fellowships.