Just over three weeks ago, the curtain came down on Adobe Summit in Las Vegas, Adobe’s annual customer event.

Last year‘s event was dominated by news of how Adobe was leveraging generative AI and infusing it across its platform and product set to help brands achieve personalization at scale. Of all the announcements made, what stood out to me, in particular, was how Adobe was using it to help marketers overcome traditional time and cost challenges in the content production process through its Content Supply Chain approach.

This year, however, much of the talk focused not on generative AI but on agentic AI and the new possibilities it offered marketers.

There were too many announcements to describe here, so here are my four big takeaways from this year’s event:

1. Adobe released 10 purpose-built AI agents, two of which should have some data analysts worried.

Adobe announced Agent Orchestrator, a new capability in the Adobe Experience Platform. It features 10 purpose-built AI agents that cover everything from audience optimization, content production, and customer journey optimization to data insights, data engineering, and site and workflow optimization.

Adobe Experience Platform Agent Orchestrator. Source: Adobe

Now, while all of the purpose-built agents are powerful applications that will undoubtedly increase the capacity and productivity of marketers and creatives, it was the data insights and data engineering agents that caught my eye.

Why?

The Data Insights Agent simplifies and enhances the process of deriving insights from signals across an organization, enabling marketers to visualize, forecast, and improve customer experiences in real-time. Meanwhile, the Data Engineering Agent supports high-volume data management tasks such as data integration, cleansing, and security- a complex operation as businesses continue to connect disparate data within their organization.

These agents are taking on what is often arduous and time-consuming work, and while the demos I saw were powered by synthetic, clean data, the power and value were clear to see.

Now, assuming that an enterprise has clean data, and that’s no small assumption to make or challenge to overcome, these particular agents are going to have a significant impact on productivity and efficiency and could potentially call into question the number of data analysts that a marketing team needs.

That’s not just me talking.

I discussed this with an executive from a large telecom firm at the event, and they mentioned that, due to the rapid evolution and increasing sophistication of AI-powered analytical tools in this space, they are considering resizing their data analyst team to concentrate on the top 25% of their analysts while redeploying the other 75% to different areas of the business.

Hence, my quick take …. traditional data and business analysts look out!

2. Brand Concierge….The future of guided experience just arrived.

There is not a lot to say about Adobe’s new Brand Concierge other than…..

If you want to see what a multi-modal, brand-centric concierge-type experience looks like – one that can take a known customer from discovery to recommendation to purchase – then look no further.

Check it out.

It’s going to be the future.

3. Adobe leans into B2B GTM orchestration

Many Martech-focused events are dominated by B2C case studies, use cases and applications.

This is despite the fact that, according to data cited at the event, B2B transactions typically account for around 70% of the GDP of most advanced and industrialised economies.

So, it was refreshing to see Adobe lean into B2B marketing and Go-To-Market (GTM) orchestration at their event and illustrate how their platform can be used to deliver personalization at scale for B2B brands.

Moreover, of the ten purpose-built AI agents Adobe announced at the event, one of them is a dedicated B2B application that focuses on account qualification, where the AI agent supports the evaluation and advancement of opportunities in a sales pipeline by automating the provision of the right content to the right people, within an enterprise buying group (i.e. the committee of individuals responsible for major purchasing decisions at an enterprise), at the right time based on where they are in their journey.

Now, this sounds great and is in line with research that shows that B2B buyers spend 70% of their buying journey doing their own research before talking to vendors.

However, what wasn’t clear was what happens in those cases where a member of a buying group has a pre-sale question, for clarification, say, that comes about as a result of consuming some of the content that has been shared with them.

It may seem like a small thing, but how that hand-off is managed and how context is passed through to the pre-sales rep or customer service agent can play an outsized role in how the buyer feels about their whole experience. A bad experience could lead to the souring of the individual’s and, possibly, their whole buying group’s view of the vendor.

4. Finally, do we know the rules of the emerging AI search game?

While at the event, I got a chance to talk to Vivek Pandya, Director at Adobe Digital Insights, about some new research coming out of Adobe that highlights the fast-growing proportion of web traffic that is now originating from generative AI assistants.

Specifically, the research found that traffic to UK retail, travel, and financial services websites from AI Assistants has surged in the last 6-7 months. Banks are experiencing the largest increase in this AI-driven growth, with a 700% increase in visit share since August, compared to around a 500% increase in the Retail and Travel sectors.

Moreover, these volumes are growing at a rate of more than 30% each month, suggesting that web traffic originating from generative AI assistants is quickly becoming a significant channel that brands need to navigate.

Over the same period, brands in the US have experienced a 1,200% to 1,700% increase in traffic growth.

But here’s where it gets interesting. Digging into the data with Pandya shows that ChatGPT is by far the most commonly used assistant among those who have used Al assistants in the past month, followed by Google Gemini and then Microsoft CoPilot.

Now, while to feature in the search results coming from Gemini and CoPilot, it would seem pretty safe to assume that the same rules of good search and Search Engine Optimization (SEO) would apply. The same is not true of ChatGPT, as speaking to two senior executives at the event, one from an international financial services company and another from a large US telecoms firm, they and their teams were not 100% clear on what is driving their search results, where all of their data is coming from and how up-to-date it is, despite the fact that OpenAI announced in May 2023 that ChatGPT would have access to real-time data from Microsoft’s Bing search engine. As a result, they find themselves in near-constant contact with their search agencies about this issue to better understand how it affects their business and, particularly, how it affects their real-time pricing offers.

What’s clear from this is that there are lots of assumptions being made, and there is some confusion about what it takes to win in this new and fast-growing AI-assistant web traffic space.

That needs to change if brands are to prosper.

This post was originally published on Forbes.com.

Credit: Photo by Mareks Mangūzis on Unsplash