Getting Started with AI in 2026: A Practical Guide
Welcome to AI-Proof.
If you’ve been watching AI from the sidelines, curious but cautious, interested but unsure where to begin, this newsletter is for you.
No hype (we’re AI users but not sponsored by products), no jargon, no pressure. Just practical guidance about what works, aimed at professionals and businesses who want to understand what’s actually possible and where to start.
The truth is, you haven’t missed the boat. While the headlines have been breathless, most organisations are still figuring this out. The difference between those making progress and those still watching isn’t technical sophistication, it’s simply that some have started experimenting while others are still planning to plan.
So if you’re ready to take a first step, or looking for your next one, here’s what we’ve learned works.
Four Principles Before You Begin
Before diving into use cases, a few guiding principles that will save you time and frustration.
Start with a business problem, not the technology. The question isn’t ‘how can we use AI?’ It’s ‘what’s slowing us down, costing too much, or not getting done?’ AI is a tool. Like any tool, it’s only useful when applied to a real problem. There are lots of “cool” AI tools on the market, retrofitting cool tech into your workflows while it might generate “cool” stuff, in the long run it won’t create lasting value.
Begin internally. Your first AI projects don’t need to touch customers. Internal processes, research, operations, administration, are perfect proving grounds. Lower stakes, faster learning, room to iterate.
Measure something. Decide upfront how you’ll know if it’s working. Hours saved? Faster turnaround? Fewer errors? Without a baseline and a target, you’re just tinkering.
Experiment. Don’t be afraid to try things. Sign up for free trials, test tools in real workflows, and see what actually works. But always anchor experimentation back to the principles above, and never implement AI just because it looks impressive. Value comes from solving real problems, not chasing novelty.
With those in mind, here are four areas where organisations are seeing genuine results today.
Research and Market Intelligence
The problem: Good decisions need good information, but gathering it takes time. Competitor analysis, market research, industry trends, someone has to research, read, articulate, and summarise. That someone is usually expensive or stretched thin.
What’s now possible: AI research tools or AI powered workflows can now read dozens of sources, rank the content relevant to your needs, pull out key information, and compile findings with citations - in minutes rather than days. You describe what you need to know, and the tool goes and finds out.
A starting point: Pick a recurring research task your team does manually. Weekly competitor monitoring, perhaps, or pre-meeting briefings on prospective clients. Try running it through a deep research tool like Perplexity or Claude. Compare the output to what you’d normally produce. You might be surprised.
One professional services firm we spoke with now has overnight briefings prepared automatically for their partners, ready and waiting each morning. The junior analyst who used to compile them now spends that time on actual analysis.
Sales and Business Development
The problem: Sales teams spend too much time on activities that don’t directly involve selling. Researching prospects, finding contact details, preparing for calls, chasing confirmations. The actual conversations, where value gets created, are a fraction of their day.
What’s now possible: Much of the preparation and follow-up work can be automated. Lead enrichment tools can compile prospect profiles before your team makes contact. AI assistants can handle initial qualification calls, appointment confirmations, and basic discovery questions - then hand off to humans for the substantive conversations.
A starting point: Map out your current sales process and identify where time disappears into administration rather than conversation. Lead research and appointment confirmation are often good first candidates - visible impact, manageable complexity.
The goal isn’t to remove humans from the sales process; it’s to elevate them. AI should take care of the repetitive, administrative work so your salespeople can focus on what only humans do well: building relationships, exercising judgement, and closing complex deals.
Marketing and Communications
The problem: Staying visible requires consistent content across multiple channels. But quality content takes time, and most organisations either sacrifice consistency or stretch their teams too thin trying to maintain it.
What’s now possible: AI can handle much of the mechanical work around content, such as monitoring industry news, drafting initial responses, reformatting for different platforms, scheduling distribution. The creative and editorial judgment stays with your team; the assembly line work doesn’t have to.
A starting point: Try using AI to draft a first version of your next piece of thought leadership, based on bullet points or a rough outline from your subject expert. Edit it, shape it, make it yours, but let the tool do the initial heavy lifting of getting words on the page.
Because AI isn’t perfect, the most effective approach is a human-in-the-loop model. AI does the preparation (removes the starting from a blank page); humans refine, shape, and approve the final output. You retain control and quality, while dramatically increasing the volume and speed of what your team can produce.
Simple Internal Tools
The problem: Every organisation has a wish-list of small tools that would make life easier, trackers, dashboards, simple databases, but they’re never quite important enough to justify the IT investment. So they stay as clunky spreadsheets or don’t exist at all.
What’s now possible: AI coding assistants have reached the point where you can describe what you need in plain English and get a working prototype. These aren’t enterprise-grade systems, and they don’t need to be. They’re lightweight, functional tools designed to solve immediate, everyday problems. We’ve used this approach to build simple but highly effective solutions such as timesheet trackers, mileage logs, and other internal tools that would never have justified a full IT project, but deliver real value day to day.
A starting point: Think of a simple tool your team has wished for but never got. A project tracker, a client database, a booking system. Describe it to Claude or a similar assistant (e.g. Loveable) and see what it produces. You might be surprised how quickly ‘wouldn’t it be nice if...’ becomes ‘actually, we have that now.’
The barrier to trying something has never been lower. The cost of a failed experiment is now an afternoon, not a budget request.
Where to Begin
Our advice is simple: pick one thing and try it. Despite all the hype, AI is not yet a fully autonomous solution for most use cases. But by starting with a single, well-defined task, organisations can achieve meaningful results quickly. Focus and simplicity are what turn experimentation into real progress.
Not the most ambitious application you can imagine. Not a transformation programme. Just one process, one task, one problem, something contained enough to experiment with safely and concrete enough to measure.
Give yourself permission to learn. The organisations making progress aren’t the ones with the best strategy documents. They’re the ones who started.
That’s what AI-Proof is here for: practical guidance, real examples, honest assessments of what works and what doesn’t. No hype. No pressure. Just useful information for leaders who want to make informed decisions about AI in their organisations.
More to come. Welcome aboard.
—Les