Build an AI Portfolio Without Being Technical

When someone says "portfolio" you probably think of designers with Behance pages or developers with GitHub repositories. Something technical. Something that requires creative or coding skills to build.

But here's a thought that took me embarrassingly long to arrive at: anyone who uses AI to do their job better has portfolio-worthy material. You just need to start capturing it.

i didn't document my AI work at my previous job. When i was made redundant and had to demonstrate my AI skills to potential employers and clients, i had nothing to show except my word. "Trust me, i used AI extensively in my previous role" is a weak pitch. Screenshots, metrics, and case studies are a strong one.

Don't make my mistake.

What an AI portfolio actually looks like

Forget fancy websites (unless you want one). An AI portfolio for a non-technical professional is simply a collection of documented examples that show: what problem you faced, how you used AI to solve it, and what the result was.

This can live in:

• A Google Doc or Notion page
• A section of your LinkedIn profile
• A simple one-page website (you can literally use AI to build one)
• A PDF you attach to job applications
• A folder on your desktop that you reference during performance reviews

The format doesn't matter. What matters is the content.

Each portfolio entry should have:

The problem. What was the task or challenge? Keep it concrete. "Needed to analyse 6 months of customer feedback data to identify trends for the quarterly review."

The approach. How did you use AI? Which tool? What was your process? "Used Claude to categorise 2,400 customer comments into themes, then had it generate a summary with supporting quotes for each theme."

The evidence. Screenshots if possible. Before/after comparisons. The actual outputs (sanitised of any confidential information, obviously). "Here's the analysis that was produced. Here's what the previous manual version looked like."

The result. Quantify wherever you can. Time saved, quality improvement, scope increase. "Reduced analysis time from 3 days to 4 hours. Identified two themes that had been missed in previous manual reviews."

That's it. Problem, approach, evidence, result. Repeat for each example.

What to document (starting today)

You're probably already using AI at work in ways that are portfolio-worthy. You're just not capturing them.

Here are the categories of AI work that make good portfolio entries:

Time savings. Any task where AI significantly reduced the time required. "Weekly report generation: 4 hours reduced to 45 minutes." Screenshot the AI conversation (with sensitive info removed), screenshot the output, note the time difference.

Quality improvements. Work where AI helped you produce something better, not just faster. "Used ChatGPT to identify gaps in our project plan that three experienced team members had missed." Show the gap analysis. Show that it was useful.

Process creation. Any workflow or system you built using AI tools. "Created an AI-assisted onboarding document generator for new client projects." Document the workflow steps, the tools used, and the before/after for the team.

Learning acceleration. Times when AI helped you get up to speed on something quickly. "Had to present on regulatory changes in the insurance sector with 48 hours notice. Used Perplexity and Claude to research and structure a 20-minute presentation." Show the presentation. Note that it was produced in a fraction of the usual time.

Creative solutions. Unusual or clever uses of AI that solved problems in unexpected ways. "Used AI to translate our technical product documentation into customer-friendly language, reducing support tickets by 30%." These are the entries that make people sit up and pay attention.

The screenshot habit

This is the single most important habit to develop: take screenshots of your AI interactions before you close the window.

Every time you have a particularly useful AI conversation — one that saved you time, produced something good, or solved a tricky problem — screenshot it. The whole conversation if it's short. The key parts if it's long.

Store these in a dedicated folder. Date them. Add a one-line note about what the context was.

You'll be amazed at how quickly this builds up. Within a month, you'll have dozens of examples. Within three months, you'll have more portfolio material than most "AI-certified" professionals can demonstrate.

i started doing this about six months ago and i now have over 200 screenshots. When someone asks me how i use AI, i don't talk in abstractions. i pull up specific examples. "Here, this is how i used Claude to restructure a client proposal last Tuesday. Here's the prompt i used. Here's what it produced. Here's the final version after my edits."

That level of specificity is incredibly persuasive. It shows real, practical competence, not theoretical knowledge.

Before/after comparisons

These are the most powerful portfolio entries. Show the old way and the new way, side by side.

Before: "Monthly competitor analysis took 8 hours. Manual Google searching, copying data into spreadsheets, writing up summaries. Here's what the old report looked like."

After: "Same analysis, AI-assisted, takes 90 minutes. Research compiled by Perplexity, data structured by Claude, first draft generated and then refined by me. Here's what the new report looks like."

Show both versions. Let the improvement speak for itself.

The most compelling before/after comparisons aren't just about speed. They show that the AI-assisted version is also better. More thorough. Better structured. More insights. Because when you spend less time on the mechanical parts, you can spend more time on the thinking parts.

Measuring what matters

You need numbers. Not because the portfolio has to be a spreadsheet, but because managers, recruiters, and clients respond to quantified impact.

Track these metrics for your portfolio entries:

• Hours saved per week/month. "This workflow saves approximately 6 hours per week."
• Quality metrics. Error rates before and after. Completeness improvements. Customer satisfaction scores.
• Adoption numbers. "Three other team members now use this workflow."
• Financial impact. If you can estimate it. "6 hours per week x £50/hour equivalent = £15,600 per year in time savings for the team."
• Scope expansion. "Previously we could only analyse 100 customer responses per quarter. With AI, we now analyse all 2,400."

You don't need all of these for every entry. But having even one or two quantified metrics per example makes the portfolio significantly more credible.

Building the portfolio (practically)

Here's my recommended approach. Takes about 30 minutes to set up and then 5 minutes per entry going forward.

Step 1: Create a document. Google Doc, Notion page, whatever you prefer. Title it something professional like "AI Applications Portfolio" or even just "AI Work Examples."

Step 2: Create a template entry. Each entry gets: Date, Task/Problem, AI Tool Used, Approach, Result, Evidence (screenshots/links), Metrics.

Step 3: Backfill. Think about the last month. What have you used AI for? Document what you can remember, even without screenshots. These entries won't be as strong, but they start building the collection.

Step 4: Capture going forward. Every time you use AI for something worth noting, spend 5 minutes documenting it. Screenshot, brief description, metrics. This is the habit that matters most.

Step 5: Curate quarterly. Every three months, review your entries. Pick the 5-10 strongest ones and polish them. These become your "highlight reel" for LinkedIn, interviews, or performance reviews.

Using your portfolio

A portfolio is only useful if you actually use it. Here's where:

Performance reviews. Walk into your review with a printed (or screen-shared) portfolio. "Here are 12 examples of how I've used AI to improve my work this quarter, with a combined time saving of approximately 40 hours." Your manager will not have seen anything like this from your colleagues.

Job interviews. "Can you give me an example of how you've used AI?" Most candidates will waffle about ChatGPT. You'll pull up specific, documented, quantified examples. The difference is enormous.

LinkedIn and professional visibility. Turn portfolio entries into posts. "Used AI to reduce our quarterly analysis from 3 days to 4 hours. Here's how." These posts perform extremely well because they're specific and practical.

Client pitches. If you're a freelancer or consultant, portfolio examples demonstrate capability better than any proposal. "Here's a similar project I completed, here's the AI-assisted approach I used, here's the result."

Internal advocacy. When you're pitching AI tools or projects to your boss, a portfolio of proven examples is your most powerful evidence.

Common objections (and why they're wrong)

"My AI use is too basic to document." No, it isn't. If you used AI to write a better email, that's worth documenting. Basic use that's well-documented beats impressive use that nobody can verify.

"I can't show work examples because of confidentiality." Sanitise them. Remove names, companies, specific figures. Show the structure and approach without the confidential details. "Analysed customer feedback (details redacted) and identified three previously unrecognised trends."

"This feels like showing off." It's not showing off. It's professional documentation. Developers have GitHub profiles. Designers have portfolios. You're doing the same thing for a new skill category.

"I don't have time." Five minutes per entry. If you're too busy for five minutes of career investment, we should have a different conversation about priorities.

Start now

Here's what to do today:

1. Create your portfolio document.
2. Write down three times you used AI at work this month, from memory.
3. Take a screenshot of your next useful AI interaction.
4. Set a reminder to add one entry per week.

In three months, you'll have a portfolio that puts you ahead of 90% of professionals who "use AI." Because they use it, but they can't prove it. You can.

The gap between using AI and being able to demonstrate that you use AI well is massive. Close it. Start using the tools, and start documenting the results.