Product Ops Is the Afterthought that Everyone Now Needs
The hard work organisations really need to do to maximise the impact of AI
Recently I took part in an AI Hackathon where a bunch of product managers got together and spent a day thrashing out some fairly sophisticated POCs using powerful AI tools. What’s evident is that with the advent of AI, it’s no longer the build time or the ‘how’ that acts as a constraint to ambition, it’s the underlying setup.
Case in point: my team hit an obstacle early. We were trying to assess progress towards a product goal, but immediately we required data to help us answer this question. We needed to determine where that data lived for us to easily integrate with it, and find a way to ensure that data was high quality.
This experience almost certainly reflects a shift in the entire industry. For large and small organisations, the pivotal question has moved from ‘what problem are you trying to solve and how will you solve it?’ to ‘is your organisation set up to solve problems?’ This gap has a name – it’s called product operations, and AI just made the major concerns of this discipline much, much more pressing. Concerns like:
How is your data set up?
How do your teams collaborate in prioritisation?
Where are you sharing insights and decisions and how can people access them?
Who is empowered to make decisions and who makes the final call?
Why this is happening now?
Broadly, we’re seeing three trends intensifying this problem:
AI tooling is increasing speed. There used to be some delay where teams could refine plans, but now the gap between vision and execution has shrunk. It’s a bit like swapping a screwdriver for an automated drill. It might be quicker, but discernment becomes ever more important. Organisations want the speed and efficiency of AI without ensuring a robust decision making apparatus. Without that, you’re just multiplying chaotic outputs.
Data lives in more places than ever. One of the biggest challenges organisations face is storing and structuring their data properly, which is the essential ingredient for leveraging AI. Off the shelf AI is exceptionally good at coding, but still patchy in being a personal assistant. Why? The data source for one is regimented. The data source for the other is scattered, incomplete, and requires judgement to parse. Organisations should be shifting their data assets as close to the former as possible and that’s not easy to either do or prioritise above more ‘visible’ outcomes.
Good product strategy depends on good business strategy. Who ‘owns’ strategy still can feel unclear. If the business strategy is woolly, that impacts the effectiveness of the product strategy. Leaders who do not feel digitally confident still have a vital role to play in clarity and leadership that requires very little digital understanding – yet still, organisations are being vague in their business aims and expecting product to close the gap unassisted. Decision-makers need calling out and given the right forums.
None of these problems are new. What’s new is how fast they’re compounding together — and how everybody is looking to product to solve the mess.
So what does product ops actually do?
The phrase simply covers three core jobs:
Data governance — making sure there is one source of truth (or an interconnected system of truth) so PMs aren’t pooling from sources that quietly disagree.
Tool standardisation — a central stack for planning, AI, delivery, research, experimentation and more, so teams aren’t independently adopting overlapping (or conflicting) tools and progress can be consistently measured
Decision making forums and workflows – determining roles, responsibilities and escalation pathways for decision-making. This includes selecting the processes of choice (e.g. continuous discovery, strategy frameworks, measurement mechanisms) and consistently communicating them.
It’s not glamorous. It’s also the difference between a product org that can move forward and one that falls into the same disagreements every planning cycle.
Fix the right problem
If you can fix your data issues and get your teams working on the same tools in a clearly articulated decision-making structure, you have the groundwork essential for capitalising on AI.
Keeping ahead is not about jumping on AI as fast as possible – that will only ever magnify the problems you already have. Keeping ahead is about recognising the underlying structural problems and doing the hard work to fix them.
So if your organisation has more than a handful of PMs and more than one AI tool in active use, you already need this function — whether or not anyone’s given it that title yet.
 | A guest post by
|