{"id":3700,"date":"2025-10-02T11:00:00","date_gmt":"2025-10-02T11:00:00","guid":{"rendered":"http:\/\/www.zoomlavilin.com\/?p=3700"},"modified":"2025-10-09T13:16:17","modified_gmt":"2025-10-09T13:16:17","slug":"heres-exactly-how-the-hubspot-blog-team-uses-ai","status":"publish","type":"post","link":"http:\/\/www.zoomlavilin.com\/index.php\/2025\/10\/02\/heres-exactly-how-the-hubspot-blog-team-uses-ai\/","title":{"rendered":"Here\u2019s exactly how the HubSpot blog team uses AI"},"content":{"rendered":"

I\u2019m going to be real with you: I\u2019m exhausted. We blinked, and it felt like everything<\/em> was run by artificial intelligence (AI). I\u2019m being dramatic, of course, but AI<\/a> is<\/em><\/a> transforming how we work<\/a>, play, and live faster than any tech I\u2019ve seen in my decade-long marketing career.<\/p>\n

\"Download<\/a><\/p>\n

With so many changes and unknowns, it\u2019s been overwhelming even for the seasoned marketers and writers here at HubSpot. But the more we learn, the more we realize AI doesn’t have to be the enemy of content writers.<\/p>\n

When used as a thoughtful addition to a human-led content creation process<\/a>, AI can be a powerful tool that helps us all work faster and smarter \u2013 without sacrificing our human touch.<\/p>\n

Here on the HubSpot blog team, we’ve been experimenting with AI for a while now, and, like everyone, we\u2019re still figuring it out.<\/p>\n

But we’ve also uncovered some exciting use cases that have streamlined some of our more time-intensive workflows, improved our output, and given us time back for deeper topic research.<\/p>\n

We\u2019re not about gatekeeping, so let\u2019s talk through how we\u2019ve brought AI into our writing process.<\/p>\n

<\/a> <\/p>\n

How HubSpot Bloggers Use AI in Their Writing Process<\/h2>\n

Step 1: We use AI to break through creative blocks and generate ideas.<\/h3>\n

Every great article starts with a spark of inspiration. Sometimes the ideas flow freely; other times, we need help getting off the ground. This is where our relationship with AI begins.<\/em><\/p>\n

I\u2019ve penned around 800 articles as a content marketer, and even I still get writer\u2019s block. When I had trouble in the past, I would chat with teammates for thoughts and ideas. But working remotely, there are fewer opportunities to talk casually like this.<\/p>\n

While AI can’t replace chats with teammates, it has been a game-changer when I want help fleshing out an idea or thinking about how to approach a topic. <\/p>\n

“When I\u2019m planning topics to cover, AI is an incredibly helpful brainstorming buddy,\u201d shares Fellow HubSpotter Caroline Forsey.<\/p>\n

\u201cI just throw an idea or even a general topic in a chat, and it kicks back angles, subtopics, and even potential titles. I don\u2019t use its suggestions verbatim. But starting off with that creative spark saves me hours on ideation, letting the human team layer in voice and nuance from the get-go.”<\/p>\n

I couldn\u2019t agree more.<\/p>\n

Recently, I was coming up with topics for HubSpot\u2019s Breaking the Blueprint<\/a> column, and had some trouble getting started. I tagged ChatGPT in for some help, prompting:<\/p>\n

\u201cI\u2018m brainstorming article topics for a blog focused on giving marketers and business professionals from underrepresented groups advice and information they need to know. I want to target specific topics and concerns that are unique to these groups. I\u2019d also like to touch upon topics related to the evolving role of AI on society and business. Can you give me some ideas?\u201d<\/p>\n

Here’s a peek at its response:<\/p>\n

\"screenshot<\/p>\n

From there, I workshopped the suggestions and then met with a colleague to finalize the topics that best suited our audience.<\/p>\n

Like I said, AI doesn\u2018t fully replace bouncing ideas off a colleague, but it can provide a promising starting point when you\u2019re stuck. This got my <\/em>creative juices flowing, and that\u2019s all I needed. <\/em>Once we have our direction, it’s time to dive into research.<\/em><\/p>\n

Step 2: We power our research with AI as an information assistant.<\/h3>\n

With our topic and angle defined, we need to understand our subject inside and out. This is where AI becomes our research powerhouse.<\/em><\/p>\n

As a content marketer and editor, my top concern is always quality and accuracy. I want to make sure the information I put in front of any audience is valuable and trustworthy. That means a lot of research, and good research takes time<\/a>.<\/p>\n

Generative AI can help exponentially here with its ability to gather and summarize information from multiple sources, often in seconds. In fact, 47% of marketers report<\/a> using AI for research already.<\/p>\n

When it comes to my own research process, I no longer have to filter through pages and pages of Google results to find the information I need. Instead, AI does most of the heavy hunting for me and wraps it up into a nice, easy-to-understand package. It doesn\u2019t just give me a resource that might<\/em> answer my question like a search engine; it gives me an answer directly.<\/p>\n

Forsey also sees the benefits of using AI<\/a> for content research.<\/p>\n

She recalls, \u201cOnce, I needed to write a post \u2014 ironically \u2014 about the differences between ChatGPT, Google\u2018s Bard [now Gemini], and Bing\u2019s<\/a> AI-powered search engine. So, I asked ChatGPT: \u2019Can you tell me the differences between ChatGPT, Google\u2018s Bard, and Bing\u2019s AI-powered search engine? Please put the pros and cons in a table format.\u2019”<\/p>\n

Shortly after, Caroline got back a detailed chart illustrating exactly what she wanted in a few short paragraphs:<\/p>\n

\"screenshot<\/p>\n

Kaitlin Milliken<\/a>, Senior Program Manager for our Freelance Writer Network, considers AI a big part of her research process as well. She shares how it\u2019s helped her distill information about HubSpot products when writing.<\/p>\n

\u201cWhen we write about products, I’ll ask AI to explain in more detail what makes a certain product unique, and it gives me a list of specifics that can help inform my writing\/editing.”<\/p>\n

The result? Using AI for content research and analysis has helped us dig deeper and uncover angles and understandings we may have missed or taken more time to come across without it. Now comes the actual writing.<\/em><\/p>\n

Step 3: We collaborate with AI on rough outlines and messy first drafts.<\/h3>\n

Here\u2018s where things get interesting. We don\u2019t just research and brainstorm with AI \u2014 we actually let it help us write. But there are parameters to our approach.<\/em><\/p>\n

The HubSpot blog team doesn\u2019t just copy, paste, and publish AI-generated content. Rather, we use AI as a jumping-off point \u2014 that messy first draft to fuel ideas.<\/p>\n

One of the big reasons people read HubSpot articles is to learn from real-life marketers, sales, and service people. They want expert, verified examples and accounts of what works and what doesn\u2019t.<\/p>\n

That kind of human expertise doesn\u2019t come from AI. Human judgment and storytelling can\u2019t be replicated. AI can\u2019t produce lived knowledge, experience, and insights, but it can quickly gather initial, general information for our writers and editors to build into a final piece.<\/p>\n

Our writers use AI to save time, particularly for more general, educational articles. Many, including myself, use AI to draft the more straightforward sections of blog posts so they can focus on more impactful ones.<\/p>\n

For instance, I\u2019ve often turned to AI to draft conclusions in recent months. With the bulk of my article content finalized, I pop into ChatGPT and ask it to draft a conclusion that matches the tone and ideas shared. From there, I refine and edit to my liking.<\/p>\n

\"graphic<\/p>\n

HubSpot writer Tristen Taylor<\/a> also uses AI for blog intros and conclusions. She explains:<\/p>\n

\u201cBefore using AI, I often found myself rewriting introductions and conclusion paragraphs multiple times. Now that I have access to these tools, my time for completing blog articles has been cut down tremendously. AI can capture the topic, intent, and targeted keywords I need through refined AI prompts.\u201d<\/p>\n

She adds, \u201cIt also allows me to refocus on the content or information of value that readers are looking for, instead of ruminating over small details at the beginning or end.\u201d<\/p>\n

Our team has used AI in our creative brief and outlining process for over a year, and we\u2018ve started experimenting with AI-generated first drafts. So far, we\u2019ve found AI drafting to be a helpful time-saving measure for some topics, though not all.<\/p>\n

For more general topics, such as \u201cwhat is\u201d and \u201chow to\u2019s,\u201d it frees up more of our writers\u2018 and editors\u2019 time to speak with other experts, gather unique examples and case studies, and \u201chumanize\u201d more generic topics overall, but AI can\u2019t offer the original ideas needed for op-eds and thought leadership pieces.<\/p>\n

Read: <\/strong>How to Humanize AI Content So It Will Rank, Engage, and Get Shared in 2025<\/a><\/strong><\/p>\n

\"graphic<\/p>\n

This is where our philosophy comes into play: \u201cAI assists, humans lead.\u201d AI supports ideation, drafting, and iteration \u2014 but people make the final call. AI gathers the pieces; we put together the puzzle. We never publish AI-generated content without thoughtful human editorial oversight.<\/em><\/p>\n

Learn more about our AI principles here.<\/a><\/strong><\/p>\n

Step 4: We use AI to analyze data and uncover insights.<\/h3>\n

Numbers don\u2018t lie. They add credibility to our articles, but they don\u2019t always tell their story clearly either. Here’s how AI helps us make sense of our data.<\/em><\/p>\n

Numbers are not my strong suit. When faced with survey data or performance analytics, my head spins a bit, so when the prospect of using AI for data analysis came about, I couldn\u2019t have been more excited.<\/p>\n

Earlier this year, I tested this use case for our State of Inclusive Marketing 2025 report<\/a>. ChatGPT allows you to upload spreadsheets, so I attached the file and asked it to give me the most significant trends coming through the data.<\/p>\n

\"screenshot<\/p>\n

Initially, the agent just returned a simplified spreadsheet (womp), but when I requested that it explain the qualitative trends in paragraph form, here\u2019s what I got:<\/p>\n

\"screenshot<\/p>\n

Could I have distilled these trends without the help of AI? Yes \u2014 but not in the minute or two it took ChatGPT.<\/p>\n

Senior Staff Writer and Website Editor Amy Rigby<\/a> echoes this, reflecting on analyzing data for our website marketing report<\/a>.<\/p>\n

“LLMs are incredibly powerful at identifying patterns in data, especially ChatGPT\u2019s o3 model. I used ChatGPT to help me highlight trends and further segment our website marketing survey data, resulting in a much deeper, richer report. I used to get frustrated with ChatGPT because its math was frequently wrong, but the o3 model<\/a> has greatly improved.\u201d<\/p>\n

While it\u2019s tempting just to take the trends AI finds and run, double-check them. As we know, AI can make mistakes and make things up. This is <\/em>the biggest challenge marketers report<\/em><\/a> facing with AI, so take the time to spot-check your data to ensure the agent didn\u2019t take any liberties.<\/em><\/p>\n

(Note: OpenAI just <\/em>launched GPT 5<\/em><\/a>, so hopefully, it\u2019s improved in this area.)<\/em><\/p>\n

Step 5: We experiment with AI-generated visuals to enhance our content.<\/h3>\n

Great words deserve great visuals. Here’s how we use AI to create compelling imagery and interactive elements that would otherwise be out of reach.<\/em><\/p>\n

So, I fancy myself a keen self-taught graphic designer, but writing remains my first and most refined skill. While written content still matters, consumer content consumption habits<\/a> and buying behavior increasingly demand more visual and interactive mediums like videos, infographics, and quizzes.<\/p>\n

This kind of advanced content is out of reach for me and marketers with limited design and development skills.<\/p>\n

Thankfully, many AI tools are available now to help us. In fact, our AI Trends for Marketers report<\/a> found that image and design generators are tied with general chatbots for the most popular AI tool marketers use.<\/p>\n

\"bar<\/p>\n

Source<\/em><\/a><\/p>\n

Senior marketing manager and AI blogger Martina Bretous<\/a> showcases generative AI’s prowess in this area in her series \u201cIs it Real or AI?\u201d<\/a><\/p>\n

In it, she crafts a variety of multimedia content (i.e., video, audio, photos) using AI and challenges readers to see if they can identify them. For instance, this video of a beautiful yellow bird she featured was AI-generated on PikaLabs<\/a>.<\/p>\n

Impressive, right? I also used ChatGPT to create this highly specific featured image of a piggy bank representing a Facebook budget.<\/p>\n

\"screenshot<\/p>\n

Source<\/em><\/a><\/p>\n

Overall, AI is helping us create visual elements that can truly enhance the experience of our content and make the concepts easier to grasp. There are tons of robust <\/em>AI video<\/em><\/a> and <\/em>AI image generators<\/em><\/a>, including <\/em>HubSpot\u2019s Breeze<\/em><\/a>, that you can use to do the same.<\/em><\/p>\n

Step 6: We use AI as a quality control check.<\/h3>\n

Before any piece goes live, it goes through our editing process. And yes, AI plays a role here too \u2014 but not in the way you might expect.<\/em><\/p>\n

Whether we realize it or not, we\u2019ve been using AI for proofreading and editing for decades.<\/p>\n

Think about it: spell and grammar check on Microsoft Word and Google Docs, \u201cautocorrect\u201d on iPhone, Grammarly \u2014 even non-writers rely on these tools to minimize misspelled words and grammar issues that could come off as negligent or unprofessional.<\/p>\n

Today, this has just gotten more advanced.<\/p>\n

“People are rightfully concerned about AI\u2019s potential to hallucinate in editing, but one thing that gets lost in the debate is that humans make mistakes, too,\u201d reflects Rigby.<\/p>\n

In recent months, she has turned to AI agents as proofreaders and even editors for her articles.<\/p>\n

She continues, \u201cEvery writer knows the experience of being too close to their own work to see it clearly or being in such a rush to meet a deadline that they miss something. So, in addition to human editorial oversight, I see LLMs as another set of eyes, an extra layer of verification.\u201d<\/p>\n

\"graphic<\/p>\n

\u201cI fact-check all of my articles, but for additional assurance, I sometimes run them through ChatGPT or Claude and ask the AI to flag any errors, which I can then look at more closely. AI often does catch typos and errors that I missed.\u201d<\/p>\n

She doesn\u2019t just stop at grammar and spelling. Rigby recommends asking agents for feedback on even how to improve your writing.<\/p>\n

She says, \u201cI sometimes ask AI to critique my work, poke holes in my logic, and point out weaknesses I may not have noticed.\u201d<\/p>\n

I like to prompt it with something like, \u2018Think like a digital marketer who will read this piece to decide which software to buy. What concerns would they have after reading this? What questions would come up that were not answered in the piece?\u201d This helps me go back and re-examine and strengthen my arguments.\u201d<\/p>\n

Rigby has also built a custom GPT called StyleScout for our team’s internal use, which helps writers quickly scan their drafts for style inconsistencies. This was as easy as uploading our internal style guides to the GPT and letting AI handle the rest to ensure our content is consistently on brand<\/a>.<\/p>\n

\u201cBefore I submit an article, I can run it through the GPT and get a line-by-line breakdown of anything that might violate our style guidelines. Yes, AI sometimes returns outputs that are overly pedantic or heavy-handed. That\u2019s why human judgment always gets the final say. It’s my responsibility to check that what it tells me is sensible and accurate. If it isn\u2019t, I disregard it.”<\/p>\n

Check out our free Masterclass \u201c<\/em><\/strong>How to Create Custom GPTs That 10x Your Marketing Results<\/em><\/a><\/strong>\u201d to learn how to build a custom GPT for your team.<\/em><\/strong><\/p>\n

Step 7: We use AI to whip up \u201cmicrocopy\u201d (i.e., meta descriptions, social posts).<\/h3>\n

The article is written, edited, and ready to publish. But our work isn’t done\u2014now we need to promote it. This is another area where AI shines.<\/em><\/p>\n

For some reason, writer\u2019s block strikes hardest when I write meta descriptions or social posts for my blog articles. Perhaps it\u2019s topic burnout, or I\u2019m just too close to the pieces to distill the real golden nuggets of information.<\/p>\n

Whatever the reason, quickly generating the right messaging to promote your content is one of my favorite uses of AI. For instance, here\u2019s what ChatGPT gave me when I asked it for some social copy based on this article:<\/p>\n

\"screenshot<\/p>\n

It\u2019s not perfect, but it gives me something to take and make my own.<\/p>\n

Jamie Juviler<\/a>, HubSpot Manager, Staff Writing & AEO, has also leveraged AI for meta descriptions. He suggests feeding your introduction paragraph into an AI chatbot and then asking it to produce a 100-150-character meta description.<\/p>\n

The beauty of this approach is that AI handles the initial creative lift, freeing us up to focus on refining the message and ensuring it aligns with our brand voice.<\/em><\/p>\n

Bonus: AI helps us stay current on the industry.<\/h3>\n

It may sound clich\u00e9, but the rumors are true: Marketing changes quickly. Thankfully, AI helps our blog team keep up with it.<\/p>\n

Juviler explains his approach. “In this new realm of AEO and, more broadly, AI-powered marketing, I’ve been leveraging Gen AI as an information consolidation agent.\u201d<\/p>\n

\u201cWe\u2018re constantly flooded with AI-related news, tips, and technologies, so whenever I come across a new piece of information \u2014 whether it\u2019s a video, blog post, podcast, newsletter, etc.\u2014 I run it through a custom GPT<\/a> I’ve set up that helps me determine whether it\u2019s important or new. If it is, the AI modifies a personal tracking document I maintain for the industry as needed.\u201d<\/p>\n

This system allows Juviler to stay on top of the stream of AI news without getting overwhelmed. This concept could also be used for other trends we cover, like email marketing, creator monetization, and social platform algorithm changes, or any topic of interest in your industry.<\/p>\n

We\u2019re in this together.<\/h2>\n

Like everyone today, our team is learning as we go when it comes to AI. The technology has its benefits but also drawbacks, and with constantly evolving technology and unclear regulations, it\u2019s a bit of the Wild West. But we\u2019re figuring it out.<\/p>\n

We want to lead with transparency with our audience, so we\u2019re committed to sharing how we use AI in our work, as it evolves, and every step of the way.<\/p>\n

We will include a disclaimer on any post we write that has been AI-supported, letting you know that we used AI to complete the post. And you\u2019re welcome to dig into the HubSpot Media AI guidelines<\/a> to learn more.<\/p>\n

We don\u2019t know what tomorrow holds, but we know AI will help us all navigate this changing landscape. We know AI can’t replace the humanity behind the HubSpot blog \u2014 and we will never take that human connection for granted.<\/p>\n

\n

Editor’s note: This post was originally published in May 2023 and has been updated for comprehensiveness.<\/em><\/p>\n<\/div>\n

\"\"<\/p>\n","protected":false},"excerpt":{"rendered":"

I\u2019m going to be real with you: I\u2019m exhausted. We blinked, and it felt like everything was run by artificial intelligence (AI). I\u2019m being dramatic, of course, but AI is transforming how we work, play, and live faster than any… <\/p>\n","protected":false},"author":1,"featured_media":3703,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[11],"tags":[],"_links":{"self":[{"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/posts\/3700"}],"collection":[{"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/comments?post=3700"}],"version-history":[{"count":2,"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/posts\/3700\/revisions"}],"predecessor-version":[{"id":3726,"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/posts\/3700\/revisions\/3726"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/media\/3703"}],"wp:attachment":[{"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/media?parent=3700"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/categories?post=3700"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.zoomlavilin.com\/index.php\/wp-json\/wp\/v2\/tags?post=3700"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}