2025 is off to a flying start. I usually publish a reflective end-of-year blog about key milestones and trends going forward. On the personal side, I decided some years back to do away with resolutions, largely because they were too hard to keep and by the end of January, new distractions took over.
I now commit to one word and I channel that throughout the year. It works! That word becomes a focal point. I breathe life into it and it’s surprising how you can find multiple meanings across different contexts, with one word or phrase. I meditate on it, lean on it when things get wobbly, I even use it as a conversation starter with strangers.
Thanks for reading! Subscribe for free to receive new posts and support my work.
So I’m applying this singular approach to describe business and tech expectations in 2025. The phrase is: AI needs HI. And HI means Human Intelligence. So let me explain what I see, as I gaze into my ‘25 crystal ball coming from a place as a user of AI tech and advisor to some interesting startups.
Less AI “Dabble” and More “Do”
We love our new BFF, ChatGPT and probably lean on it more than Google search for a host of questions. Quick and free (well sort-of). Who doesn’t love that? Its value goes farther than a single prompt.
ChatGPT or any LLM model that automates content creation or gets answers quickly, is a capability baked into many tools. Its now-commodity power is often a step in a series to get work done.
Clay.com, a GTM darling has integrated such models into their templates for personalization, based on the target persona. This is a gem for targeted messaging to yield higher responses.
All types of GTM tools are taking this approach, treating ChatGPT as built-in to a workflow or step-series. Copy.ai does this and many run-the-business platforms such as Hubspot, Salesforce, and Zendesk have integrated ChatGPT. You can find Reddit threads where teams are getting it deployed, fast and easy. No coding. Asana, a fave for busy teams, integrates with ChatGPT for productivity gains. The list goes on.
So there’s less dabbling as more vendors have AI baked into their latest offerings and as more teams take AI seriously and adopt it into their day-to-day practices. Experimentation has moved into the early adopter phase. By early adopter, I mean seriously executing programs and practices using AI tech, not dabbling.
AI Needs a Purpose with Realistic Timelines
As soon as the GTM team shifts into AI gears, the natural next question becomes “why are we doing this, how long will it take and what outcomes are we measuring”. AI should never be a distraction from the core business initiatives. The promise of high returns has caught GTM’s attention in 2024 and now teams are assessing the amount of effort for that beautiful payback. Set expectations early before diving in.
Marketers have the biggest FOMO of any group, so coming up to speed with latest-greatest is high on their list. Make sure you clearly define your AI use-cases and set timelines with KPIs, just like any other new tech investment.
I recently talked with a team at a series A startup to find out that implementing Clay took the best part of 5 months before bearing fruit. In fact most freelancers doing this peg it at 6 weeks to get it off the ground.
Now they didn’t spend a consecutive 40 hours a week for five months as there were other things to get done. However, it’s surprisingly long and not for the faint-of-heart.
The challenge described was more to do with actual adoption. Sound familiar? AI or not, all tech needs to be implemented correctly and then adopted or you will never realize gains.
With powerful capabilities from vendors like Clay, the thinking often starts with “let's do these easy first steps” which then turns into “oh it can also do these other steps in an automated way” to “oh we must onboard team-members to see real results so we need enablement and change-management too”. There’s always too much to do, not to mention the regular quarterly pressures, so you can see how this type of initiative can consume many months.
So what are we learning from this? Be realistic, know what resources you need to commit and then be patient taking the time to learn and adjust.
AI without a goal is really just an experiment
Good or bad, ChatGPT’s immediate value may mislead some into thinking all AI yields instant results.
When new tech forces some adjusting of core processes, you need to apply rigor. Everyone on the team must know how to adjust practices and of course avoid collisions with the older way.
The good news is that AI’s flexibility with use-cases allows you to parallelize lots of tasks to automate steps that previously took manual hours and a predefined sequence. So what might this look like in reality?
Before that, let’s break down the types of capabilities. If it answers yes to the following, you should investigate using AI:
Is it a repetitive task or set of steps?
Is this an ongoing need for multiple team-members?
Does this consume a ton of manual effort and resources today?
Does this need to scale over time for productivity gains?
If most of the answers are “yes!”AI may be the ticket.
Now to deploy an AI solution, it will likely falls into the following use-case types:
Data insights or KPI/Metrics answers
Anything content or written copy-related
Workflow tasks or steps that integrate with multiple apps, databases, tools and systems.
Of course the solution will lean on an underlying LLM model and that’s awesome but you still need to ‘stitch things together’. For that you need a human to drive it, create and shape it and then execute. We love AI but it won’t run without human intelligence (HI).
Let’s take one example from that outline and drill down a bit.
Pipeline KPIs
Every GTM team watches their metrics, a lot. It informs “what happened and how we need to adjust”. The more that data is available and reliable, the faster you move. A metric is any number from across the entire funnel such as new pipeline created to conversion rates from inbound, to CT rates and CPC for paid channels. Every data point is critical to understanding the full picture.
Now you must prepare for an important meeting to lay out plans and budgets. There’s so much data and oh so many dashboards, charts and reports.
How can AI help in this case? You need access to the right data and then you need to understand it.
Pull data reports from your core apps, e.g. Hubspot, Salesforce etc.
Use AI to merge the data files into one readable format, error free and clean.
Upload merged structured data files into an AI platform for analytics and insights.
Ask English questions (or SQL) about the data for immediate answers.
Share reports before the meeting and a summary to get everyone on the same page.
Run the meeting. Get fast alignment and faster decisions (hopefully).
For future meetings, run the same AI workflow and share in advance of the meeting.
Now you avoid hours of data wrangling, staring at multiple sheets and cross-checking. With AI You experiment with the questions you need answers to and so much faster. No coding. Simple text. No debating anymore as everyone is on the same data page. Reach decisions faster. The same process steps actually exist like before but you’re having AI do all the legwork. Brillant.
A similar set of steps can be automated for other non-metrics tasks such as:
Create landing page copy for a new product feature and related copy for social posts.
Personalize copy for specific target accounts and personas for SDR to execute.
Analyze the sentiment of a top competitor's social posts over the past six months.
The first step is understanding what needs to be done and the second step is can it be done with AI and automation. The most important step is “does the human on the team believe that this is the best approach going forward and can it add value, save time and increase ROI?”
Thehuman intervention is an important part of this. I would argue the most. Just as you do today. You need to review all KPI result output. Yes, you need to validate any copy that the models create, avoid plagiarism and boring vanilla. AI is wonderful and speeds up so many things relieving us to do high-value work. However, it doesn’t happen without our human intelligence.
Looking ahead 2025
If AI was so dang powerful, wouldn't it predict what’s about to happen during 2025? OK, maybe that’s a bit too stretchy. I do know AI is changing the game for GTM teams. There’s no choice but to embrace it. We are moving out of the hazy fog and hopefully ‘25 will bring clarity and an easier path forward.
Finally teams can get ahead of tasks and execute faster. Finally they have the power of data insights at their fingertips which means freedom, fewer headaches, less fog and more revenue.
Wishing you a healthy, safe and prosperous year ahead.
Thanks for reading! Subscribe for free to receive new posts and support my work.