AI Operations Consulting for founder-led businesses

Turn friction
into flow.

Flow is your team at their best: doing challenging, meaningful work. I build the AI systems that take the rest off their plate. For some clients that means hours back every week. For others it has changed what their business is able to promise.

50+ Workflows automated

Two ways to use AI

Most businesses use AI the cautious way: same process, same steps, a bit quicker. That gets you 10%. Anyone can get 10%.

The bigger gains come from rebuilding the process around what AI and automation make possible. One of my clients used to take two working days to produce a customer pricing update. Speeding up the old process might have got them to a day and a half. Rebuilding it got them to 40 seconds.

I can do both. Quick wins pay for themselves and earn trust. But the deeper work, that's where the real advantage is.

Hi, I'm Fergus.

I built OnBrief.ai after 25 years leading digital agencies and technology businesses (including Future Platforms, Dare and Engine). I see the same thing in almost every business I work with: talented, expensive people bogged down in work that's not their core talent.

AI can now do that work: the processing, the repetition, the data-heavy grind. That frees your team for the strategy, the creativity, and the decisions that move the business. And once the grind is gone, there's a second question worth asking: what could the business deliver now? That's where the interesting answers live.

I'm a commercial strategist who's worked with organisations like Aviva, Barclays, BMW, First Bus, Virgin Active and Waitrose.

25+
Years in commercial strategy
50+
Workflows automated
Fergus Adam

What I do

Most businesses know they could be more efficient. The hard part is knowing where to start.

The Audit

We look across your operations to find exactly where busy work is undermining performance. You get a specific diagnosis: where your team's time is going, where it shouldn't be, and what becomes possible once it's freed.

The Strategy

We prioritise. The quick wins are the tasks draining your best people and are the fastest to fix. They pay for themselves and prove the approach. The bigger opportunity: the parts of your delivery worth rebuilding, that's where the gains stop being incremental.

The Build

I build the engine, using automation and AI appropriately for your business. Your team stays central. The system handles the grind, and where it makes sense we redesign the process around what's now possible, rather than bolting AI onto how things have always been done.

Embedded Partnership

For businesses that want to keep going. The first build removes friction. The ongoing work compounds your benefits: extending the system, rebuilding delivery step by step, turning time saved into capability your competitors don't have. I work alongside you on retainer. Think fractional Head of AI Ops.

How I run my own business

OnBrief is built on the principle it sells: work out what genuinely needs a person, and build the system to do everything else.

In practice, my business runs on a single master agent orchestrating a library of custom skills: meeting analysis, account management, finance, outbound, scope writing. Each skill knows its job, the tools it needs, and how I like things done. I ask for what I want in plain English and the system does the rest.

One agent orchestrates a library of custom skills and tools. Click into the diagram to explore it, and hover any skill to see what it depends on.

If the system is built right, the size of the team stops being the limit.

Selected projects

The numbers tell the story. What took two working days now takes 40 seconds. Days of processing became near instant synthesis. Twelve million messy data points became instantly queryable. None of these teams got smaller. All of them can now deliver something they couldn't before.

PR & Communications Multi-Agent System

Hello Tomorrow: raw research to ready-to-brief

Days of processing to near instant synthesis

The problem

Hello Tomorrow is one of the sharpest comms agencies in the UK. PR Week's Agency to Watch 2026, founded by Fenella Grey and Frankie Cory, built around a simple belief: the best work comes from experienced human judgment, not headcount.

But every new brief still arrived front-loaded with hours of reading. Documents to digest, research to cross-reference, context to surface; before any real thinking could begin. For a team built around strategic edge, that was an expensive way to spend the most valuable thing they have. Their time.

What I built

This wasn't the old process with AI bolted on. We rebuilt how a brief moves through the agency, phase by phase: audit and current-state analysis, competitive positioning, workshop preparation, strategic narrative, content execution. A specialist agent for each stage. Each one does one thing well.

The design principle was deliberate: AI handles what AI is good at. The team controls the inputs, reviews what comes back, and brings their judgment, insight, experience and creativity to bear where it matters. The more creative the work, the more that layer counts. The agents clear the path. The strategists decide where to go.

Results

  • Strategists start from synthesis, and the thinking starts on day one
  • Hours of document processing eliminated per brief
  • Creative and strategic thinking gets the time it actually deserves
  • Agent instructions updatable in plain English, no developer needed
  • Built inside Google Workspace, no new tools required

Technologies used

Claude Google Drive Google Docs RAG Make.com
"With the system Fergus built, AI takes care of the groundwork - and we get the space back for the strategic and creative thinking our clients value most."
— Fenella Grey & Frankie Cory, Co-Founders, Hello Tomorrow
← Back to all work
FMCG / Brand Data Pipeline BigQuery

The Jolly Hog: from manual data prep to real answers

12m data points, instantly queryable

The problem

The Jolly Hog runs regular brand tracking surveys. Every wave produces a detailed picture of awareness, consideration, and purchase intent across different audiences. Exactly the kind of data a growing brand needs to make confident decisions.

The problem was what happened between each new wave arriving and anyone being able to use it. The raw files needed significant manual preparation before any analysis could start. With 12 million-plus data points, the insight was always there. Getting to it was the busy work.

What I built

An automated pipeline built to handle the messy reality of real-world data: inconsistent column names, format changes between waves, slight differences in how questions are asked. The mapping rules live in a simple schema; when the survey changes, the pipeline updates itself. No developer required.

When a new file arrives, it triggers the process. Everything loads into a structured BigQuery database, ready to query the moment it completes. On top of the pipeline sits an AI query layer. The team types a question in plain English (how awareness has shifted in the last two waves, how consideration compares across audiences, which metrics have moved most) and a routing agent writes the SQL in the background and returns a clear answer.

The practical effect: brand data stopped being something you schedule time with and became something you simply ask.

Results

  • Manual data preparation eliminated entirely
  • 12m+ data points queryable in plain English
  • Questions answered in the meeting where they come up, rather than weeks later
  • Pipeline updates itself when survey structure changes
  • Live Looker Studio dashboard for always-on brand tracking

Technologies used

Python BigQuery Looker Studio Claude Make.com
"On my (and everyone's?) to do list: explore how A.I can make me more efficient. Fergus has created a platform that allows me to look at my brand tracking data instantly.... I'm not going to be attempting the back end of the A.I systems (what a minefield) so thank goodness for people like Fergus!"
— Christine Everett, Marketing Director, The Jolly Hog | B Corp
← Back to all work
Fire & Security Excel Automation

Specialised Fire & Security: from 2 days' work to 40 seconds

40s used to take two working days

The problem

Every month, someone at Specialised Fire & Security sat down with a pricing spreadsheet of over 40,000 rows. The job was straightforward in theory: update the prices, check the figures, get it ready for the team. In practice, it was two days of slow, careful, manual work: easy to get wrong, hard to delegate, and difficult to justify as the best use of a specialist team's time.

This is what we call busy work. It is not the reason anyone joins a specialist business. And it accumulates, month after month, in ways that are easy to accept and hard to fix.

What I built

The solution didn't require new software or a replacement system. Using Power Query in Excel, a built-in feature most teams never reach, all the manual handling could be rebuilt to run automatically at the click of a button.

AI played a supporting role here, as the way to reach the right solution faster rather than the engine doing the work. The result is a properly built, reliable process running inside the tool the team already knows.

A monthly job that was hard to delegate and easy to get wrong is now a button anyone can press. That is a small change on paper and a different working month in practice.

Results

  • What took two working days now takes 40 seconds
  • Risk of manual errors eliminated
  • A job that depended on one careful person now runs at the press of a button
  • No new tools to learn or manage
  • Time freed for the work a specialist business actually needs its people doing

Technologies used

Excel Power Query Claude
"If you are looking for advice on how you can get the most out of AI, you would do well to speak with Fergus, who's been brilliant."
— Albert Hall, Director, Specialised Fire & Security
← Back to all work
Client Skills Training Outbound Research Data Automation

Team Sterka: from 60 hours of research to minutes

60hrs of research, replaced by minutes

The problem

Team Sterka is a people skills training programme that builds commercial confidence across agencies and consultancies, founded and run by leadership coach Trenton Moss. I know the programme from the client side, so I knew exactly how good the product was. The bottleneck was in front of it.

A high-touch training programme lives or dies on reaching the right businesses. Too big and the programme doesn't fit. Too small and it isn't the right moment. And within each right-fit company, only a handful of people are genuinely worth a conversation. Getting that targeting right meant hours of manual research per campaign: finding companies, checking profiles, cleaning lists. Around 60 hours of it, before a single message could go out.

What I built

Two connected workflows. The first handles research: a direct link between Claude and Clay, built on MCP, so Trenton can research potential target companies conversationally and get enriched, reliable data back. Clay's LinkedIn integrations surface the prospects within them.

The second handles quality: an automation that cleans and checks every record before a campaign starts. Trenton's targeting criteria are built in as skills, written in plain English. Companies that are too big, too small or the wrong profile are weeded out automatically, and the right people within each qualifying business are identified and verified. When his ideal client profile shifts, he changes a few sentences rather than a system.

Results

  • Around 60 hours of manual research replaced by automation that runs in minutes
  • Wrong-fit companies filtered out before any outreach begins
  • The right people identified within every qualifying business
  • Campaigns start from clean, checked data
  • Targeting criteria updatable in plain English, no developer needed

Technologies used

Claude Clay MCP LinkedIn Make.com Google Sheets
"You've basically replaced 60 hours of work with an automation that works in minutes."
— Trenton Moss, Leadership Coach and Founder, Team Sterka
← Back to all work

Let's talk.

Is there a process in your business that your team has accepted as "just how it works"? That's usually a good place to start.

Book a free online meeting

30 mins: where the friction is, and what removing it makes possible.

We'll work out where the friction is, and what removing it would make possible. If the honest answer is that it's not worth doing, I'll tell you that too. No PowerPoint. Just a useful conversation.

Find a time →