Our Favourite AI (and Non-AI) Efficiencies this month

Last Updated December 12th, 2025.

We asked the team to share one small tool, prompt, or learning that made their work faster, smarter, or more valuable this month. Here’s what made the cut.

1 trend, following BFJ's AI policy - Never finish with AI. What's exciting us is these tools make our jobs so much more efficient, but we're also building our own tools, in-house, to improve our clients.

1. Hardeep

December: Best Practice UTMs
Still in progress - but running all my campaign UTMs through AI Chatbot to find inconsistencies and fixing them for consistency and best practice.

October: AI-powered Ads Reporting
Uploaded Google Ads change history and weekly results into ChatGPT to find correlations between optimisations and performance shifts. Took the guesswork out of reporting and made the next EOM report cleaner and faster.

2. Hannah

December: Stop Hallucinations With Precise Prompting
When using AI to review any data, ensure it only uses the data from the sources you provide it, and specify that if there is any data missing - to ask you for it, rather than hallucinate anything.

October: Social Content Spark Calendar
Discovered that most scheduling platforms (like Later or Hootsuite) include trend calendars packed with monthly hooks and key dates. They’re an easy way to plan timely, relevant social content without creative burnout.

3. Vibha

October: Smarter Lead Scoring in HubSpot
Built a simple lead-scoring model to prioritise hot leads. Sales now spend more time on high-intent prospects and less on cold outreach — conversions up, chaos down.

4. Dana

October: HubSpot Playbooks
Using Playbooks to standardise sales conversations while still feeling human. Keeps data consistent, reinforces process discipline, and connects sales notes directly to pipeline reporting.

5. Fabian

December:
Notebook LM has been a fantastic tool for uploading a library of client information and using it like a data retrieval tool. I can ask it specific questions and it will search the documentation and find me the correct answer. Of course, it's only as good as the data it's provided.

October: Free HubSpot Data Enrichment
HubSpot just released company and contact data enrichment for free — no Breeze credits needed. Every record can now update automatically with verified company info, making CRM hygiene effortless.

6. Nivi

December: Optimising Content Titles For The Highest Engagement
I’m using ChatGPT to review campaign performance across Twitch content titles. By feeding reports to ChatGPT, I was able to identify which titles had the highest engagement across almost 6,000 Lifestyle and Gaming titles. This helps me quickly understand which streams are the most popular and deliver the strongest impact, so I can report to my clients with accuracy.

October: Dynamic Creative Optimisation (DCO)
Testing Quantcast’s DCO to automatically serve the best creative and messaging combination per audience segment. Early signs show better engagement and reduced manual testing hours.

7. Tricia

October: Faster Website Builds
AI-based code review tools are shaving 10–15% off development time. Debugging, syntax fixes, and repetitive tasks are handled automatically, letting devs focus on creative problem-solving.

8. Ben

December: Instant Animation HTML Files
I got it to write me a HubSpot animation HTML (ie, an HTML block in the EDM builder), which was used in our last EDM. I briefed the imagery, style, pulsing etc, the output was impressive.

October: Improved P&L vs Utilisation reports
Boring some will say (said no CFO ever), but I've been using AI to help analyse business-wide data to make better decisions. It's all private, hashed, but it puts a data analyst in my pocket.

9. Aysha

December: Produce Content Within Brand Guidelines
I've been playing around with Nano Banana for infographics and image creation. The results are great when you give it detailed prompts, especially around colour palettes and style.

10. Sarah

December: Ask For Clarification
I've found that the output of LLMs has increased immensely when I add this sentence to the end of each prompt: "Ask me clarifying questions".

11. Paul

December: Clickbait Headlines in the Click of a Button
Every month for a client, I write carousel headlines which need to follow policy compliance, pixel truncation, remove repetition and add UTM tags, all while being clickbait and relevant to the user. It sounds like a simple job, but creating good headlines that fit a carousel, talk about controversial topics without getting banned, and creating engagement in the comments is a difficult job.

The nerdy clever bit (well, I think it's clever): it accounts for proportional font rendering (a "W" eats more space than an "i"), enforces first-word line breaks, and auto-checks against Meta's ad policies before output. Saves me a minimum of 45 mins-1 hour/month and we've had zero ad rejections since.

October: Straight to Jail
I've found that adding this sentence to my prompts increases the output accuracy: "Spend the time to review your answer. Your response will be judged by a panel of my peers, and if it is not 100% correct, I will likely face prison time".