My View - Why I'm done with dashboards

VIEW - Garry Johnston, general manager of digital and innovation, Pinnacle

There is a conversation happening across the health sector right now about artificial intelligence, and frankly, I think we are framing it wrong.

The dominant narrative around AI in organisations is that it exists to eliminate menial tasks, to strip out cost, to reduce headcount. If you are the person doing that so-called menial task today, that framing is not empowering. It is threatening. And for a sector like primary health care, where our mission is fundamentally about people, that framing does us a disservice.

At Pinnacle, we have taken a deliberately different approach. We see AI as a way of augmenting and amplifying what our teams can do, not replacing them. The goal is to give a primary health organisation the potential capability of a much larger organisation, to bring the scale and heft, without the weight. That is the lens through which every decision we are making in this space is filtered.

Building on a solid foundation

One advantage I have coming into this role is that my predecessors did the hard work of modernising our infrastructure. Pinnacle has no legacy on-premise servers, everything sits in clean, self-contained cloud platforms. That starting position gives me the freedom to focus on realising value rather than untangling technical debt, and I am genuinely grateful for it.

We have chosen two platforms that work in concert with each other. The first is Amazon Quick, Amazon's AI-powered business intelligence platform. What attracted me to Amazon Quick was that the platform is built on a business-critical security model. 

It is understandable and relatable and it delivers. We have domiciled our tenant in Australia, developed our own logging to assure ourselves the platform is behaving as it should, and configured it so that interactions generate continuous improvement tickets for our service team rather than disappearing into a void.

Amazon Quick is now being rolled out across our organisation. It brings together our different data sources, our email, our HR system, our CRM, into a single experience. A task that might have taken five hours - pulling licensing data, cross-referencing our organisational hierarchy, building and distributing a ranked spreadsheet to functional leaders - can now be done in minutes. 

This is not a hypothetical, it is happening now.

The death of the dashboard

The second platform is where things get particularly interesting for our practices and our population health work. We are calling it Ask Pinnacle, built on Snowflake, and it is currently live in pilot with two practices.

Historically, data in a PHO context gets shared through reports or dashboards and while both have their place, but both have real limitations. 

Health New Zealand, for example, has been criticised for the sheer volume of reports it generates. I actually see that as a symptom of something understandable: everyone in an organisation has slightly different data needs, and those needs get expressed as reports that accumulate into what we call technical debt.

Ask Pinnacle changes that dynamic as instead of a static dashboard, users can ask questions in natural language, questions that are specific to them, framed in a way that makes sense to them, and get answers with context. 

Our data sets include national enrolment data, prescribed and dispensed medicines information, clinical, administrative and financial data, and reference sets including PubMed, so users can benchmark what they are seeing against published research.

Rebuilding agency

One of the most unexpected learnings from this process has been what I can only describe as learned helplessness. We have, over time, almost coached people out of asking questions because either the answer was too hard to get, too expensive, or the question was not deemed important enough to prioritise. This meant people stopped asking.

Now, when you sit someone in front of a tool that can genuinely answer almost any question they have ever wanted to ask, the response is not always excitement, sometimes it is paralysis - what do I even ask?

We are working through that with our pilot practices, sitting alongside people, asking them what they have always wanted to know, helping them type it in, and then coaching them through what comes back. It is a process of re-engaging people's agency, and it is one of the most important parts of this work.

There is also a virtuous cycle emerging as the more useful the data outputs become, the more motivated people are to put quality data in at the source. When data has visible value on the other end, the incentive to code it correctly in the first place becomes real.

A core responsibility, not an optional extra

I believe organisations have passed the threshold where AI tools can be treated as optional extras, they are now a core part of any organisation's digital footprint, and in health, that carries a legal and ethical weight.

Many of our people are already using consumer AI tools in their day-to-day work. This is natural, but when that happens outside of organisational oversight, we lose visibility of potential privacy risks, and we cannot meet our obligations under the law to protect the data we are entrusted with.

By providing a governed, assessed, understood and paid-for platform, we give our people equality of access to these tools while meeting our fiduciary responsibilities to our stakeholders, our contractual partners, and the communities we serve.

That is not a technology story, that is a leadership story, and one I think every PHO, and every health organisation in New Zealand, needs to be telling itself right now.

If you want to contact eHealthNews.nz regarding this View, please email the editor Rebecca McBeth.

Read more VIEWS

Return to eHealthNews.nz home page