I recently had the opportunity to interview the chief strategy officer at OpenAI. Preparing for that conversation — and thanks to those of you who contributed questions via Slack — got me thinking about what we, as news organisations, actually need from consumer answer engines.

Answer engines are quickly becoming an interface for knowledge. But if we want them to work for news publishers, there are a few core areas where we need clarity, collaboration, and innovation. We need to be forthcoming about that, which is something INMA is working on. 

This conversation isn’t unique to OpenAI. It’s for all LLMs and consumer answer engines, but he gets kudos for showing up and talking to us about it.

Here are the four areas I bought up in the discussion with Jason. 

1. Data: understanding how people use answer engines

Right now, we have almost no visibility into how consumers engage with answer engines when it comes to news. Ideally, we’d get hard, individualised, interactive reports showing what users ask, what they click, and where they engage. At a minimum, we need aggregate insights on how audiences are consuming news inside these experiences.

Without better data, it’s impossible to make strategic choices around monetisation, product, and discovery

I’ve spoken to the team at OpenAI about this, and we may have some information we can share soon. I’ll keep you posted. To set expectations, it definitely won’t be a full data console but some broad strokes on how ChatGPT is used for news. 

2. Traffic: Will answer engines drive audiences back?

Many of us would love answer engines to act as a marketing channel for our owned-and-operated products. But today, I see two major challenges:

Firstly, user intent. Answer engines give people exactly what they need, instantly. If they want more depth, they simply ask another question, they don’t necessarily click through.

Secondly, a UX mismatch. Moving from a clean, simple conversational interface into a busy, ad-heavy Web site often creates friction and drop-off.

The latter is something we’re actively working on with OpenAI over the next six months: workshops with media executives exploring better, smoother pathways between answer experiences and publisher environments. I’ll be back with a report on the findings in due course. 

3. Brand: How do our titles show up?

As answer engines become a gateway to information, brand attribution becomes critical:

• Do consumers see our reporting surfaced transparently with brand attribution?

• Are trusted sources knowingly elevated above low-quality content?

• How does this influence audience perception of credibility?

If our brands disappear inside generic responses, the long-term relationship with readers is at risk. One big advantage we have in an AI tsunami is trust. We need to ensure this is reflected and leveraged. 

4.  Monetisation: beyond licensing

Licensing deals will likely play a role, but we need to think beyond them.

Some possibilities:

• Integrated subscriptions. Imagine if an answer engine understood who was a paying subscriber and could serve deeper, tailored answers accordingly.

• Commerce opportunities. Could answers link seamlessly to relevant products, events, or affiliate experiences?

• Premium bundles. high-value add-ons for highly engaged users.

A model where subscriptions are part of the answer engine experience could actually make the product better —surfacing richer, more personalised information, while extending the value of the subscription itself. That opens up stickier engagement and deeper opportunities for both publishers and answer engines.

The vision on all of this sounds obvious to media professionals. But the reality is that it’s a lot more complex.

• Product development must happen on both sides.

• Commercial agreements need to be structured and tested.

• Publishers require onboarding, integration, and ongoing support.

• Customer support systems must evolve to manage a hybrid ecosystem.

And quite honestly, this has to fit in with a number of other priorities answer engines are grapling with. This isn’t a “switch it on” scenario; it’s an iterative collaboration between answer engines and publishers (which is why I am such a big advocate of dialogue).

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .