Last week, Google revealed that the viral AI image editor known as Nano Banana was in fact its new Gemini 2.5 Flash Image model.

The tool’s speed and accessibility are already putting Adobe on the defensive, forcing the company to rethink how it positions Photoshop.

The Bigger Story

This wasn’t a typical product launch. For weeks, an anonymous model was quietly dominating LMArena’s image battles.

Users quickly noticed it wasn’t like the others – edits came out sharper, characters stayed consistent, and results appeared in 1–2 seconds instead of 10–15.

Clues soon followed. Banana icons slipped into prompts. Google engineers posted cryptic emojis on X. The performance felt like DeepMind’s. By the time Google confirmed the truth on Tuesday, the community had already connected the dots.

And that reveal landed at a tough moment for Adobe. Photoshop still depends on manual skill and layered workflows.

Google’s model flips the dynamic: anyone can type “add a leather jacket,” “change the background to a beach,” or “remove that person from the photo” and get professional-grade results in seconds. At $0.039 per image generation, it opens up advanced editing to everyone – not just trained designers.

Investors have already been bracing for this shift. Adobe’s stock is down 35% year-to-date, with analysts pointing to mounting fears that AI will undercut its subscription model. 

Against that backdrop, Adobe acted fast. Within days of Google’s reveal, it folded the model into Firefly and Express – a clear signal that hosting disruption is safer than resisting it.

Why You Should Care

Photoshop isn’t dying tomorrow. Professional designers aren’t abandoning Creative Suite for a chatbot. But Nano Banana represents the moment AI image editing became genuinely useful for everyday editing tasks.

Adobe’s 35% stock slide reflects investor recognition that the $300/year Creative Cloud monopoly faces pressure when comparable AI capabilities cost pennies.

Adobe’s swift decision to integrate Google’s model also shows they see it too. Their strategy is shifting toward aggregating the best AI tools rather than relying solely on their own technology.

For compete pros, watch how Adobe frames this move in its next earnings call. The language they use will be a live case study in defending market position when your differentiator starts to look like a commodity 🍌

You weren’t hired to build battlecards.

You were hired to help your company win.

But tactical busywork is stealing your time and your influence.

That’s why we built Compete Agent – your new competitive intel operative.

Compete Agent automatically:

• Monitors every sales call for competitive threats
• Instantly alerts sellers with deal-specific responses
• Generates fresh competitor profiles in seconds
• Keeps all intel updated daily (not quarterly)

The result? Your team stops getting blindsided by competitors and starts winning more deals.

📌 Teams like Gainsight are already saving 250+ hours on research while hitting 72% seller adoption! Don’t let your competitors get there first.

See Compete Agent in action 👇 

ChatGPT can feel like a cheat code for competitive analysis – until it isn’t.

I’ve seen product marketers type “Tell me everything about Competitor X” and get something that looks solid: well-organized, confident-sounding, and ready to drop into a slide deck.

The problem? “Sounds right” isn’t the same as “is right.”

ChatGPT has several inherent blind spots that make it seriously risky for competitive intelligence work – and it’s worth understanding these before putting any of its outputs into use.

Some of these limitations include:

• It can’t guarantee real or accurate citations – You can’t be sure if what it gives you is outdated, biased, or wrong until you verify it yourself.
• It lacks your context –  It doesn’t know your competitors, product, or messaging unless you give it that information, which requires manual upkeep.
• It struggles with recency – It often misses launches, pricing changes, or market shifts from the last few weeks or months, even when you ask it for this info directly.
• It’s useless for competitive deal support – It doesn’t have access to deal‑specific data, and it can’t automatically push insights into the tools and workflows your sellers already use.

But of course, that doesn’t mean it’s worthless.

Used well, ChatGPT can help you speed up research cycles, break blank-page syndrome, and turn verified insights into usable content faster.

But confuse “directional inputs” with “verified intel,” and you risk eroding trust with sellers and buyers alike.

📌 I put together a full guide on how to get the benefits of using ChatGPT for research without stepping on the landmines.

Read it here: How to Do Competitive Analysis with ChatGPT (And When To Avoid It)