More than a robot copywriter: how AI unlocks hidden productivity gains in B2B SaaS
Last updated: April 20, 2023
Most SaaS marketers’ opinions on the use of AI in marketing range from maximalists, who believe it will change everything and replace all marketing jobs, to AI skeptics, who believe it lacks soul and won’t disrupt any jobs.
We see AI slightly differently and instead believe that AI can assist humans and help them become more effective marketers.
So instead of replacing humans altogether, we believe marketers can apply ChatGPT and other generative AI to existing workflows to attain more volume and velocity.
While we don’t believe AI will take a marketing strategy from zero to one, it will help marketers get there faster by automating steps in the process. Additionally, it can make the final product yield a higher ROI if you use it properly.
As a result, marketers that know how to use AI effectively can produce better results in less time and with less resources.
In this post, we’ll discuss:
- How we used ChatGPT to improve our understanding and analysis of our total addressable market
- Three underappreciated ways that AI will improve marketing team productivity
- Setting expectations and guardrails for how AI can help marketers is an important next step for most marketing teams
- Actionable Ways Marketers Can Leverage AI to Produce Better Results
By the end of this post, you’ll have a different perspective on how to use AI within your marketing organization: it’s not just a robot copywriter.
If you’d like help navigating a tougher economy and finding out how AI can improve your marketing team’s productivity in a time when everyone’s being asked to do more with less, we should talk. Start by getting your free marketing plan here.
How We Used ChatGPT To Improve Our Understanding and Analysis of Our Total Addressable Market
We were inspired to share our viewpoint on AI after seeing promising results applying it to our own internal workflows.
The first problem we used AI to solve was manually cleaning over 6,500 records pulled from ZoomInfo and Crunchbase as part of our own ABM strategy.
While ZoomInfo and Crunchbase offer basic filters (like employee headcount, B2B, and revenue), the data isn’t always specific enough to identify whether or not a company fits our ICP, and it can be inaccurate.
For example, Crunchbase shows that the memory card company, Lexar, is private equity backed and generates between $10M and $15M ARR. At first glance, it looks like it fits our ICP, which is B2B SaaS companies doing between $10M to $100M ARR. However, further research shows that Lexar is a manufacturing company.
This is a critical piece of information that Crunchbase didn’t provide, and if we didn’t clean the data, we would have wasted thousands of dollars running an ABM campaign for a company that doesn’t even fit our ICP.
Data cleaning is a mundane task – but it’s also a high stakes task as just one incorrect record could cost us thousands of dollars in wasted ad spend when it comes to targeting the right prospects.
To clean the data, we previously had only two options.
First, we could use a skilled human from our team to manually clean about 50 records per day. While this option produces a low error rate, the opportunity cost of the team member’s time is very high, and it would require about 18 weeks to clean all 6,500 records.
Alternatively, we could hire a team of VAs to complete the task faster, though the error rate might be slightly higher, and we would have to pay for the manual hours required to complete the task.
Now, AI is a third option to solve this problem.
The error rate of AI falls somewhere between a skilled human and a team of VAs, the execution timeline is reduced from weeks to minutes, and the total cost to execute the task is just a few dollars.
These benefits made AI the obvious choice to clean the data, so we created a prompt and asked ChatGPT two questions:
- Can you clarify this URL is a B2B SaaS company?
- If you say yes or no, tell me why.
In under 5 minutes, ChatGPT enriched all 6,500 records for just $1.23.
This made it the cheapest option and helped us launch the campaigns months in advance of when we had originally planned.
The team was also able to reallocate the time that would have been spent either cleaning records or managing the VAs to instead test variations of the campaigns and make them more effective.
This scenario demonstrates how AI automation can make marketers more effective by increasing volume and velocity.
Three underappreciated ways that AI will improve marketing team productivity
In our initial experimentation with AI, you saw how it helped us shorten execution time from months to minutes.
But speed and removing unskilled work is only one part of the puzzle of how AI can help marketers.
Simply by increasing speed and removing that work, marketers are able to focus on higher leverage tasks and improve ROI with less resource.
Below we’ll discuss some specific ways AI can make marketers more effective by enabling them to deliver a higher ROI with the same amount of time and resources.
Reduce Campaign Execution Timelines
AI allows you to save time by automating tasks. But the impact of the time saved is far more important because it means you can launch campaigns faster.
In our prospecting list example, we reduced the campaign preparation time from 18 weeks to just five minutes. This also means that we’ll receive the ROI of that campaign months earlier, which translates to faster growth.
So instead of just saving our team a few hours per week, the real impact of AI automation is that it allows us to close deals worth tens of thousands of dollars months in advance.
Enable More Iterative Testing
If you use AI to automate basic campaign execution, you’ll have more time to test the campaigns, make them more effective, and therefore see a higher ROI for every dollar spent.
For example, let’s say half of your time is currently dedicated to creating and launching campaigns while the other half is spent testing them. If you use AI to assist with the campaign creation process, you can reallocate that time to testing more variations of the campaigns, which helps you produce a better ROI from each campaign.
Automate Best Practice Tasks
There are probably plenty of “best practice” tasks that you’d like to take care of, but the time and headspace investment required to perform them is higher than the potential ROI. As a result, these tasks get pushed to the bottom of your list indefinitely.
An example of one of these tasks is updating hero images on your company’s archive of older blog posts.
In a perfect world, you would update all of your old hero images, but the ROI of increasing clicks by a small percentage is probably far less than the investment required to pay a creative agency to execute that task.
Now, AI can execute this task for you.
For example, you can enter a prompt into ChatGPT asking it to summarize each blog post. Then, use a Zapier GPT plugin to use that prompt with an AI image creation tool like Midjourney.
From there, Midjourney creates the image.You can add it to a Google Drive folder and have a separate automation that updates the post in WordPress.
Thanks to AI, all your hero images are up-to-date, and the entire process only costs about $25.
Setting expectations and guardrails for how AI can help marketers is an important next step for most marketing teams
As we saw in the previous section, AI is clearly effective as a tool in the broader automation strategy of any company. On that basis, we have to ask: is it only a matter of time before it’s good enough to replace marketers entirely?
There are many high-level tasks that are unlikely to be replaced by AI. For example, positioning of a product in the marketplace and taking a more nuanced view of the strategic opportunities for growing a company.
But that doesn’t preclude marketers from using AI as a ‘copilot’ in this kind of work
For example, a generic prompt fed into any AI tool such as “create a landing page for a document and workflow management tool selling to architecture firms,” will produce a generic landing page.
Sure, it will probably drive some conversions for people that are looking for commodity products, but the overall conversion rate will likely be relatively low and those who do convert are probably low-value customers anyway.
As we have mentioned many times on our blog already, buyers in B2B SaaS buy in buying committees and are often deeply evaluating the differences and nuances of all their available options in any given market.
On the other hand, a marketer that has accurately identified how their software solves a meaningful pain point better than any other solution available and accurately communicates that in their marketing messaging will likely outperform generic messaging.
Marketing teams can still use AI as a partner in producing that end result.
For example, they could simply ask the AI tool a more specific question that includes information about the ideal buyer, the pain points they experience and then a specific prompt that doesn’t attempt to get a polished, public facing result right away.
Here’s an example of the kind of detail that will get good responses and start a meaningful ‘conversation’ with an AI tool:
As you can see in this example, you still need a live marketer to establish the company’s unique positioning and accurately identify your target audience’s key pain points to provide the right prompts for AI to execute.
This highlights two particularly important point for marketers who are embracing AI in their workflows:
- In the same way that you’d review an employee’s work when they first started at your company, you need to review the AI’s work too
- You must ensure that you don’t put proprietary information into an AI tool
Two critical controls we’ve implemented at Powered By Search include human review requirements and privacy rules.
1. Review all AI work manually
Our first rule is that a human must manually review and edit any AI generated content before publicly publishing it. The ultimate responsibility for quality control and quality assurance belongs to the human.
While AI is becoming increasingly effective at creating content, it still isn’t perfect and might produce content that misrepresents your brand and the core message you want to deliver to customers.
2. Respect privacy and proprietary information
Twitter, Facebook, LinkedIn, and Google, are all free for consumers to use because user data is the product they leverage to sell to advertisers.
OpenAI is no different.
It will take any of the data you feed it and incorporate it into the large language model’s neural net. In short, this means any data you enter into the system is no longer private.
As a rule of thumb, we only share data with AI that we would share openly with anyone who we haven’t signed an NDA with.
For example, we would never run ChatGPT on a cell that contains annual customer revenue data from Salesforce.
Actionable Ways Marketers Can Leverage AI to Produce Better Results
Now that you’ve seen the possibilities of AI, how do you actually implement it into your workflow?
Below, we’ll give you some actionable prompts you can use to automate your workflow and improve your marketing efficiency.
You can use these basic prompt frameworks to help you get started with your AI or modify them to fit your specific use case, though as a general tip, we recommend that you make your prompts as specific as possible. If your prompts are too general, you’ll generate lower quality responses.
One of our favorite AI use cases for demand generation is offer optimization. Here are just a few prompts you can use to improve your offer:
- “Our CTA is ‘Schedule a demo.’ Provide three variations that are loss aversion focused.”
- “Our product is X, it solves Y pain points for Z customers. Create an offer that resonates with those pain points.”
- “Create a headline for (URL X) that highlights (unique value proposition) and includes the keyword (your main keyword).”
- “Write a detailed feature expansion for these three features that explains the pain points they solve and key benefits.”
AI is excellent for automating text ad copy creation, landing page generation, and A/B testing. Here are a few prompts you can use or modify for each of these tasks.
Text Ad Creation
- “Our new feature does X and its benefit is Y. Then, write three text ads. One that is gain focused, one that is loss focused, and one that is logic focused messaging.”
Landing Page Generation
- “Design me a landing page with a demo opt in form on the right hand side and some testimonials on the left hand side for an invoicing software company.”
- “Create three different variations of this landing page with loss aversion messaging.”
Some of the most common methods to incorporate AI automation into your organic SEO strategy include keyword research, content brief creation, topic clustering, and link building.
- “Identify keywords that a VP of Marketing at a Saas company struggling with poorly performing Google ads would search.”
Content Brief Creation
- “Create a content brief for the keyword (X) that includes target word count, secondary keywords, the post format, and questions the content should answer.”
- “What are keywords related to (X) that a person about to purchase a (Y) tool might search for?”
- “Identify blogs that are an authority in X topic that have previously accepted guest posts.”
- “Write a humorous outreach email to publishers of those blogs pitching them on guest post idea X. Include the benefits of accepting our offer and make it something they wouldn’t hit the spam button on.”
Some content marketing tasks that AI can help automate include topic research, content refreshing, and content amplification.
- “Tell me what pain points someone searching the keyword (X) is struggling with and how they’re currently trying to solve it. Then, tell me why their current solutions are ineffective.”
- “What are the most recent studies on topic X and what point do they prove?”
- “Tell me what subtopics blog posts ranking for (keyword X) cover.”
- “Tell me what topics the top ranking posts for (keyword X) do not cover that a searcher might benefit from knowing.”
- “Take this blog post, summarize the main points it’s communicating, and then use it to create five Tweets, each with actionable takeaways.”
Some marketing automation tasks that AI can help you automate include lead magnet creation and onboarding optimization. Here are some prompts you can use to get started.
Lead magnet creation
- “Can you basically create a checklist that I can transform into a PDF that covers seven or ten problems that somebody has about Topic X?”
- “Can you take those seven to ten problems and turn them into a five email onboarding sequence with actionable takeaways in each one?
- “Can you create a flowchart of a 21 day new user onboarding sequence from the point that someone signs up to a trial to the point where they pay for a trial?”
We’ve started using AI to automate aspects of our web design processes, and here are a few of our favorite prompts:
- “Here’s a website targeting X customers with Y problems. How would you optimize it to drive more conversions for a product that has Z benefits?”
- “We want to create a website in the style of webflow for a company that sells X. Can you create three different mockups?”
- “What kind of color scheme would fit with the main color X for a luxury brand selling high end vacation rentals?”
Our View of AI And How Marketers Can Embrace It To Produce Better Results
While some marketers are afraid of AI taking over their jobs, we believe that marketers who know how to leverage it effectively will become linchpins to their organizations as they’ll be able to produce better results with the same time and resources.
We don’t see AI taking over marketers’ jobs entirely, as a human still needs to direct the general marketing strategy and positioning, though it will help them get from zero to one faster and cheaper.
This is a key reason why the entire Powered By Search team is actively investing in learning how to leverage AI to produce better results for our clients in less time.
To learn more about how we can help you incorporate AI into your marketing strategy to produce better results in less time, claim your free marketing assessment today.
What you should do now
Whenever you’re ready…here are 4 ways we can help you grow your B2B software or technology business:
- Claim your Free SaaS Scale Session. If you’d like to work with us to turn your website into your best demo and trial acquisition platform, claim your FREE SaaS Scale Session. One of our growth experts will understand your current demand generation situation, and then suggest practical digital marketing strategies to double your demo and trial traffic and conversion fast.
- If you’d like to learn the exact demand strategies we use for free, go to our blog or visit our resources section, where you can download guides, calculators, and templates we use for our most successful clients.
- If you’d like to work with other experts on our team or learn why we have off the charts team member satisfaction score, then see our Careers page.
- If you know another marketer who’d enjoy reading this page, share it with them via email, Linkedin, Twitter, or Facebook.