Chester Tech Meetup: AI Marketing Plan (from Poll Responses)
Using ChatGPT to create a marketing plan using the data collected from the meetup Polls.
Thank you to everyone who attended the Chester Tech Meetup last week. During my talk, we looked at the potential dos and don’ts of B2B digital marketing and how AI could support marketers during tactical-based tasks. Part of the talk was focused on gaining the opinions of the audience via a hypothetical task to find a suitable vendor to support with an aspect of their digital marketing.
At the end of the presentation, I revealed that the anonymous first-party data collected would be used to attempt to create a marketing plan solely using AI. With the ultimate goal to determine whether AI was capable of providing some suggestions that align with the needs of its audience (you, the audience for the evening), the “buyer” in this scenario.
The poll was conducted in two parts. Part one is “strategic” and part two is “tactical.” The strategic poll was designed to identify preferences and information about the user. The second tactical poll was to generate some ideas on how to reach the audience in the most effective way.
Below is a flavour of the poll questions and answers, followed by the end product of this exercise - a proposed marketing plan. I took five steps to get to this point, and this post will walk through those steps.
Before…
Example poll with audience responses.
Example word cloud with audience responses.
After…
Marketing plan overview provided by ChatGPT following the step by step process (below).
Using AI to analyse data and suggest a marketing plan
For this exercise, we’ll be using ChatGPT (GPT-5) and some simple prompts without scrutinising or building on the responses too much.
Once you have reviewed the outputs, please provide your feedback in the comments section of this page.
Step #1 - Analysing the data
The raw data (available here), which includes unfiltered individual rows of answers, was exported from the polling app, Poll Everywhere, and then uploaded to ChatGPT. Poll Everywhere does allow for reporting, however, it would be beneficial for us to keep all the data in the same place and have ChatGPT learn about the dataset as we progress.
I asked ChatGPT to create similar charts to the ones which Poll Everywhere provided on screen during the talk. (Once data as such has been uploaded, it will ask a series of questions and suggestions for what to do next. As it’s used to handle data as such, there are times where prompts are suggested and just a simple response - yes/no - is required.)
Next, I wanted a summary of the dataset to identify some common themes. This will give me an overview of what I have to work with, as well as an opportunity to review anything that may seem out of the ordinary. Here are the insights generated (unedited from the ChatGPT response), which attempt to summarise the audience’s preferences and needs, as well as how we may be able to effectively reach them with our campaigns:
Buying factors: Clear dominance of product quality.
Ethical stance: Mostly very important, with small minorities saying not at all.
Marketing etiquette: Again, strongly very important.
Partner preference: Beer buddy is the runaway favorite.
Feelings during search: Very scattered; no single dominant feeling, but pressure, curiosity, confusion, and excitement appear often.
Browsing history use: Majority want no creepy activity using data. Some like clever personalization, very few want generic recommendations.
Tailoring the pitch: Strongest preference is for pitches focused on challenges relevant to job and goals.
Content types: Most want a balanced mix of human + AI-assisted content with fresh ideas.
AI in marketing: Most are okay with AI; a significant group are unsure, and a minority are unhappy.
One thing to tell the marketing company: Most common themes: listen, understand me, relevance, honesty, keep it simple.
Example chart created by ChatGPT.
Step #2 - Create a buyer persona
Based on the insights from the dataset, I wanted to develop the audience pool into a workable buyer persona. Whilst the dataset didn’t include explicit demographic details (age, gender, income, etc.), ChatGPT can infer likely traits from the patterns in the responses and create a buyer persona with demographics, interests and psychographics.
Some of the inferred demographic data is as follows (unedited from the ChatGPT response):
Age Range: Likely 30–50 → responses suggest mid-career professionals making rational, responsibility-driven purchase decisions.
Role/Job Function: Marketing, operations, or procurement decision-makers (given focus on pitches, content, ethics, partner styles).
Education: University educated, possibly postgraduate (analytical responses, emphasis on ethics and quality).
Income: Mid-to-high bracket (decision-making authority).
Company Size: Mid-sized to large organizations (formal marketing/partnership procurement processes).
How accurate was this? Please provide your feedback in the comments section at the bottom of this page!
I then asked it to create a buyer persona - for which it has labelled “The Pragmatic Evaluator.” Specifically, a persona poster enriched with demographics, psychographics, values, traits, barriers, and interests. We need a persona as such for any marketing campaign, but we also need a defined audience profile to give ChatGPT and prevent any confusion on the audience we are looking for it to reach.
As you can see below, there are some formatting issues which is often indicative of AI when generating images without having to go through multiple iterations. But it’s detailed and perfectly functional for the marketing campaign we are going to ask it to build.
“The Pragmatic Evaluator” persona via ChatGPT.
Step #3 - Determine the strategy
With the audience set and an idea of communication preferences established, we then need to formulate a strategic plan. So, as the “seller,” I now need to ask ChatGPT to provide some strategic ideas of how to promote the marketing agency and its services, specifically its PPC services, to this persona. The initial prompt here is as simple as “determine a marketing strategy when promoting PPC services to this persona”, which based on the outputs, could be tweaked were you to be creating a real marketing plan.
Here are the strategic considerations for a PPC-focused marketing strategy for a marketing agency, as provided by ChatGPT (unedited from the ChatGPT response):
Be clear, be human, be transparent.
Avoid gimmicky retargeting → stick to professional relevance.
Build trust through proof (case studies, certifications, testimonials).
Frame AI as an efficiency enhancer, not a replacement for expertise.
Deliver value-first offers (audits, calculators) that help them feel informed and empowered.
During this stage, I was more interested in the proposed messaging and positioning, which it clarified to me after asking for it to do so:
Lead with Quality & Outcomes
Ads should highlight results and reliability over gimmicks.
Example: “Proven PPC campaigns that drive consistent ROI.”
Show Ethics & Transparency
Make it clear how budgets are managed and measured.
Example: “Transparent PPC management. No hidden fees, no surprises.”
Tone: Professional, trustworthy, and simple — avoid flashy jargon.
I also asked it to provide this in a visual form and to cater to the traditional B2B buyer’s journey stages. This can be seen below. It has provided some tactical ideas (which we’d be looking at during the next stage), but this is the first time I began to feel that it needed more work. For example, I’m not so sure the “Beer Buddy”, who was the overwhelming winner of the “desired partner type poll,” features much in what we see above?
Whilst it’s certainly true that ChatGPT, and AI generally for that matter, can be trained and can learn, as well as iterate off the back of feedback given, the strategic recommendations felt generic on the whole. This is a reality at present for the B2B marketer - AI can’t replace him/her just yet! For a real-life plan, more time would need to be spent on this stage. But the aim of this exercise is to determine if a realistic and actionable plan could be generated from the limited data that was generated. And so far, it’s mostly succeeding.
Basic strategy map via ChatGPT.
Step #4 - Determine the tactical approach and campaign plan
As we now have the strategic outline, we need to create a mock marketing campaign plan. Here, I asked it to create one for a 3-month period. I confirmed with ChatGPT that I wanted it to create the following: I’ll build a 3-month campaign plan that ties directly to the PPC Strategy Map and the Pragmatic Evaluator persona. The goal is to balance trust-building, education, and conversion while avoiding “creepy” personalization.
The tactics/channels it has selected are as follows (unedited from the ChatGPT response):
Google Ads: Awareness → Consideration → Retargeting.
LinkedIn Ads: Sponsored content & Lead Gen forms.
Content: Guide, audit, case studies, webinar.
Email: Nurture sequence, follow-up offers.
Landing Pages: Clear, transparent, trust-building.
Whilst I would be inclined to make tweaks to the campaign plan, for the purpose of this exercise, I was largely satisfied for it to move forward and formulate this plan (as the selected channels are very much suited to the B2B audience which we work with on a daily basis). Specifically, creating a flowchart that includes a monthly rollout of the tactics.
I feel like we have gone far in just a few hours of working with ChatGPT on the approach and documenting the stages within this blog post as I proceed. This exercise was always a demonstration to be able to provide some ideas. But I would stress that more time is spent on each of the steps that we have gone through if ChatGPT were to be selected to support a task as such.
3-month campaign plan flowchart via ChatGPT.
Step #5 - Executing the plan
The final stage of this exercise is for ChatGPT to provide a structured way to approach actioning, monitoring and optimising its own 3-month campaign plan. Often, such a plan fails due to a lack of forward thinking within the execution phase, as ChatGPT rightly notes during our chat: “Planning is half the job; execution and monitoring is where success is secured!”
I was actually impressed by the detail provided here. Even though this is information readily available on the internet via multiple sources and formats, it wouldn’t be difficult for ChatGPT to actually generate this. But it has provided me with important considerations for actioning the plan, how to monitor performance (with suggested tools), as well as ensuring success and optimisation. I also asked it to create an infographic as a quick-reference guide which could be referenced mid-plan.
It has summarised the execution points (unedited from the ChatGPT response):
Plan carefully, test aggressively, monitor weekly, optimize monthly.
Keep everything transparent, clear, and value-driven — from ad copy to reporting.
Always tie activities back to ROI and clarity, since those are the strongest motivators for this persona.
Campaign success playbook via ChatGPT.
Final thoughts
ChatGPT has concluded that the “data tells us this persona values quality, transparency, professionalism, and reassurance.” Where every step of the strategy “should reduce anxiety and build trust, while avoiding gimmicks or over-automation.”
Finally, I asked ChatGPT to highlight anything else to be aware of during this campaign. I specifically referenced the original data within this prompt, so we didn’t lose focus of the information which we had gathered on the audience, questioning whether there would be any implications during this process. These considerations are below (unedited from the ChatGPT response):
Quality > Speed or Cost
The majority of respondents said “How good the product is” matters most, much more than cost or speed.
Implication: Your ads and messaging should lead with proven quality, reliability, and ROI evidence. Cost-efficiency is secondary.
Ethics & Transparency Matter
Many rated an organization’s ethical stance as “Very important.”
Implication: Position your agency as transparent and ethical. Avoid any dark patterns in ads or landing pages. Show certifications and clarify how budgets are managed.
Marketing Etiquette Counts
A significant majority said marketing etiquette is “Very important.”
Implication: Avoid pushy or intrusive messaging. Keep tone professional and human. Don’t overuse retargeting or “creepy” personalization — it will erode trust.
Preferred Partner Style = “The Beer Buddy”
Most respondents chose the pragmatic, trustworthy partner over flashy or “AI overlord.”
Implication: Brand yourself as approachable, collaborative, steady. Highlight long-term relationships, not just quick wins.
Feelings During Search
Many reported pressure, stress, confusion, uncertainty.
Implication: Your funnel should reduce anxiety with clarity, proof, and step-by-step explanations. Make the journey feel simple, structured, and reassuring.
Attitudes Toward AI
While some are comfortable with AI, many are unsure or cautious.
Implication: Don’t market PPC as “AI-driven magic.” Instead, position AI as a support tool guided by human expertise → “AI precision, human judgment.”
I have spent approximately five hours from exporting the raw data, working with ChatGPT on the five steps of the marketing plan creation, to documenting the process as we go. What we now have (towards the top of the post) is a data and AI-driven one-page marketing plan.
I’d be interested to know your thoughts? Please provide your feedback in the comments section below. Thank you.
