AntiCon LX: Key takeaways around generative AI for Sales and MarketingKatherine Hockley
Last week, hundreds of marketers flocked to “AntiCon LX: The MarTech, AdTech, NextTech, and SalesTech ‘anti’ conference.” The main topic that echoed throughout the conference was AI, so in this blog, I delve into the questions raised by the theme, such as “Will AI steal my job?” and “How can AI improve content marketing?”
How AI can drive Sales and Marketing efficiency at scale?
Frans Riemersma, Founder, MarTech Tribe
Joyann Boyce – Founder, Inclued AI
Tim Bond – Associate Director – Ipsos
Sumeet Vermani – SKV Consulting – Founder and CMO
First, the panel drew the distinction between analytical AI and generative AI. The former is not new: we are using it when we use facial recognition to log on to on our phones, for example. However, the boom has recently been in generative AI, which is AI capable of generating text, images, or other media in response to prompts.
Challenges of generative AI
The panel pointed out the dangers of AI, as it is being trained on historical data, and human beings don’t have a great history ‘overall’. This could affect the outputs in terms of inclusivity. One member of the panel compared it to a puppy – it’s learning for rewards, but you can’t just ‘trust’ it. Similarly, it’s important to tell AI ‘what not to do’ so that it doesn’t spit out negative outputs (for example, when using it as an online customer service support tool).
Positive use cases of generative AI
It can be a great shortcut and ‘bouncing off’ point, which can sometimes be a blocker for marketers. For example, when creating customer profiles, you can ask ‘what else would you include?’ which may help you fill gaps you have missed or not even thought of.
Similarly, you can use it to automate prospective outreaching, which is escalated to a salesperson in house when necessary.
It could also help marketers scale and interpret data faster, which would allow us to improve how we communicate with customers; human relationships can’t be replaced, however.
How will it affect creative careers?
Vermani pushed home the point around getting ahead of the curve and starting to be proactive with it – or it could be that you are on the receiving end of a decision from someone else about how it will be used.
One of the panel members talked about an image that was generated that was of National Geographic quality. The reason this was output was such high quality was due to the specialist knowledge of the photographer inputting the prompt. He inputted which lens, aperture, and so on for the AI to ‘use’ when formulating the image.
This was the overarching message – your specialist knowledge will be key in managing generative AI. The panel predicted that job descriptions in the future will include ‘prompt engineer’ as standard, because “AI and humans together will always beat one of these on their own.”
There was also a comment that we often see AI represented in popular culture as broadly negative (Ex Machina, for example), when the reality is that it can be a positive, so there is perhaps a warped view of AI in the public consciousness.
The panel then talked about AI not necessarily putting jobs at risk, but democratising access to certain careers. For example, AI will enable more people to do your job, which fed back to the point around ensuring you are proactive and not reactive to using it in your current role. It will also open up new careers and opportunities, so it is likely new roles that don’t exist today will exist in the future.
Regulation / legislation
Will regulators hit ‘pause’ on AI as so many tech leaders have called for? The panel felt this was not possible, as ‘bad actors’ who are using it for scamming activities wouldn’t stop, regardless of legislation. The feeling was that even if there were regulations, it would be too late.
Organisations, they felt, are likely to end up with their own AI tools, which would lead to governance developing internally, as it would have to be overlayed on top of the existing infrastructure layer of the business.
Ending the big content gamble: How big data and AI are fuelling a revolution in content ROI
Nick Mason, CEO, Turtl
Turtl is a content platform and they have been developing their AI to improve content predictions for users of their platform. This means it can predict the ROI for your content before it goes out, and if you feed into it your structure, audience, and so on before writing, it will predict how it will perform in terms of conversion and provide recommendations for improvements.
If the use of this type of AI becomes widespread, then this will change the landscape of content marketing. It takes the ‘gamble’ out of big content decisions and may be helpful when trying to explain the value and ROI of content in the future.
I looked at this in the same way I look at the predictions you see for paid ads - is not a crystal ball, but will certainly be helpful in getting buy in from stakeholders and certainly in content optimisation.
A MarTech supercollision: Smashing together AI, data, composability, and no-code
Scott Brinker – VP Platform Ecosystem, HubSpot
Scott Brinker, the ‘godfather of MarTech’, closed the conference with a look at how we were witnessing a MarTech supercollision across Data, AI, and Composability.
Data used to be ‘locked’ in apps, or that data (often too much of it) would be pushed one way towards a cloud data warehouse and remain there. Sometimes, it would be pushed out towards a BI which had limited analysis and slow feedback loops.
This has evolved with the introduction of AI and CDPs. These flows are now multi-directional, transforming the use of data and insights across an organisation as data becomes more portable and is no longer ‘locked away.’
Software / Composability
It used to be that software was built and a company had to adapt their ways of working to fit the views of that software vendor (as this would be infused into the products by those vendors, even unintentionally).
However, with the introduction of the idea of composability, we see the use of software in a way that fits the buyer without having to take on the deep technical build themselves. Brinker used an analogy of Buy, Build, and Adapt with toys.
You could buy a fully-formed toy (low/no effort to create, but not flexible/adaptable to your needs) or create one from scratch with raw materials (hard work but fully flexible). The alternative would be to build from packaged apps with composable blocks. This makes software more malleable and adaptable, enabling the customer to take control and use it how they need, rather than how it is prescribed to them.
This change has been driven by the democratisation of technology, which leads to technical developments that empower less technical users. This is where generative AI comes in.
Brinker shared an example where somebody challenged themselves to see how much paid generative AI products could do for them to launch a new product in 30 minutes. You can read the article here, but it created a market positioning document, an email campaign, a website, a logo, a hero image, a script and animated video, and social campaigns for five platforms. Of course, the quality of this output can be disputed, but ultimately it demonstrates the possibilities, especially for small companies or freelancers.
As 'VP of Platform Ecosystem' for Hubspot, he mentioned the development of the natural language interface in HubSpot, ChatSpot, where you can type prompts such as “Remind me to follow up on X contact in 3 days” or “Create X report and compare it against last month. Can you email me this report every month?”
This empowers users, who may be waiting a long time on IT for their small automation use cases. Brinker noted that these are often not prioritised over high-end use cases, leaving many projects abandoned.
So what does the future look like with generative AI in sales and marketing? Brinker shared some of his predictions:
- Content and relationships: Buyers will ‘shut out’ pushed marketing and sales content, even if hyper-personalised. This will make trusted sources even more valuable. It may also mean buyers ‘pull’ content using AI, with AI service agents talking to one another until progressed to a human interaction.
- Most software becomes disposable: It is easier to recreate from scratch than maintain. This means software becomes ambient – we pick and choose from multiple software. There will be an increase in ‘zombie’ software apps, agents, and automations.
- API services become a first-class marketing channel: Imagine asking ChatGPT to recommend a two-week holiday in New York for your family, and then asking it to book it for you. If you are a travel company, having an API that allows people to then book this trip on your website becomes incredibly valuable. Alongside this, analytics is massively democratised due to generative AI.
Key takeaways from LXA
Generative AI was the topic that dominated the entire conference. It was suggested that marketers should get ahead of the curve by taking advantage of software (composable or otherwise) that could give them a competitive edge. However, Gartner’s ‘hype cycle’ was mentioned often, so it is worth considering just how urgently and to what extent one needs to heed that advice.
There was also the underlying elephant in the room that not many people understand the ‘black box’ that goes into building these models. It is therefore likely that organisations will be looking to technical folk more than ever to help create and manage the internal AI infrastructure that will help future proof their operating model.
Savvy CMO Survey (Part 2): The Sprint for Marketing Innovation
Technically Minded | Data vs. Marketing: Who should control the view of the customer?
How psychological quirks can give your customer experience (CX) the edge
The personalisation playbook: A guide to data-driven personalisation at scale
Customer Data Platform (CDP): A game changer or just another acronym?
Over the waterfall: Is Agile the next big push in Marketing Ops?