What do AI consultants actually use AI for every day?

Most people outside the industry still think AI is about generating strange images and uncanny video. The AI a working consultant actually uses every day is almost invisible: it reads faster than any human, remembers everything it is told, drafts at the speed of thought, and talks back in plain English. It does not make weird fingers. It makes decisions cheaper, research faster, and small teams genuinely dangerous.

According to McKinsey's February 2026 State of AI report, 88% of organisations now use AI in at least one business function, yet only around 6% report material financial returns. The gap is not the technology. It is knowing what to actually point it at. Here are five things I use AI for every single week as an AI strategy consultant, and none of them involve generating a single image.

1. Competitive monitoring that runs while I sleep

The single highest leverage workflow in my week is competitive intelligence. Not "I googled my rival", but a structured weekly rhythm that reads a competitor's pricing page, their last ten blog posts, their LinkedIn activity, their job ads, and recent reviews, then tells me what changed, what it signals, and what I should do about it.

This is exactly what sits inside the Market Radar engine of our Pulse platform. The twist most people miss is that Pulse has a voice agent, Lexi, sat directly on top of the data. I do not open a dashboard. I ask Lexi, out loud: "What have our top three competitors done this week that actually matters?" and she answers in seconds, with the source links if I want them.

This is not a customer service bot. Lexi is a voice interface to your own business data. That distinction matters, because it is the difference between a toy and an unfair advantage.

For a five-person firm, this is not about replacing a junior analyst. It is about giving every person on the team an extra hour back every day to do the work only they can do. That compounds. Across a year, it is the productivity multiplier nobody is talking about loudly enough. It is also the first time in my career the word "dashboard" has stopped mattering, because the data comes to you, in the car, on the school run, at 7am before anyone else is awake.

2. Prospect prep that used to take three hours, now takes four minutes

Before any sales call or partnership conversation, I run the same routine. Company website, last twelve months of LinkedIn posts from the founder, a high-level scan of their top competitors, recent press mentions, and any podcast interviews they have done. Ten years ago that was a three hour job for a junior. Today it is four minutes, and the output is sharper than the junior ever produced, because nothing gets forgotten.

The output I ask for is not a summary. Summaries are useless. The output I ask for is: "Ask me clarifying questions before reporting. Then: three weaknesses in this business's systems or online presence that their competitors are already exploiting, what I would do about each in the next ninety days, and the one opening line that would make them realise I have already done the work." That is a prompt a good consultant writes. It is not a prompt the average user thinks to write, which is most of the gap.

The shift is not that AI is smarter than it used to be. It is that the person driving it finally is.

3. Meeting prep that used to swallow Sunday night

Picture a department manager on a Sunday evening. Their leadership meeting is at 9am on Monday. They are expected to report on current status across three or four projects, flag risks, and recommend where the team should focus the week. Historically that meant opening a dozen documents, chasing colleagues on Slack, re-reading minutes from meetings two weeks ago, and trying to remember what was actually agreed. By the time the deck was done, the story was average and the energy to tell it well was gone.

The AI version looks very different. Drop in the last four weeks of relevant material: project notes, status emails, meeting transcripts, chat threads, whatever exists. Ask not for a summary, but for a position: "Ask me clarifying questions first. Then tell me what has genuinely moved, what has stalled, the biggest risk this leadership team has not flagged yet, and the three decisions I should push them to make tomorrow morning."

What comes back is not a recap. It is a point of view. The manager walks into the room already ahead of it. They are not reporting status; they are steering it toward decisions that matter. That is the difference between a middle manager and the one their boss privately relies on.

Reporting status is a junior job. Steering a room toward the right decision is a leadership job. AI is the thing that compresses the junior work so the leadership work actually gets done.

4. Reports, decks and PDF plans without the design team

Three deliverables eat most professional service businesses: client reports, boardroom presentations, and PDF strategy plans. Historically each was a two-to-five day job split across a strategist, a junior analyst, and someone who actually knew Figma. Most SMEs skip the polish entirely, send a Word document, and wonder why the competitor who walked in with a beautifully designed twenty-page plan won the retainer.

My own workflow for all three now sits inside AI. Raw thinking goes in: voice notes, call transcripts, spreadsheets, meeting minutes. What comes out is a structured draft with an executive summary, clean hierarchy, evidence-backed diagnosis, a ninety-day plan with owners and measurable outcomes, and a properly referenced appendix. I refine it, run it through a design layer, and export to PDF. The finished deliverable looks like the kind of document that used to cost a client five figures to commission.

The prompt that works is not "write me a report". It is deliberate: "Produce a twelve-page strategic review in this structure: executive summary, three diagnosis sections with direct quotes from the transcripts I gave you, a ninety-day plan with specific decisions, owners and measurable outcomes, and an appendix with the data sources. Ask me clarifying questions first. Flag any claim you cannot back up from the material I provided."

The quality of your deliverables is read as the quality of your thinking. Until now, smaller firms were punished for the gap between the two. AI closes it.

5. Content multiplier: one idea, six outputs, in an hour

I write a weekly article like this one. That single piece of thinking gets turned into a LinkedIn post, a newsletter section, an outline for a YouTube video, three short social clips, and a talking points doc for client calls that month. The multiplier is not ten. It is at least six, if I am being honest about the quality.

The common mistake here is asking AI to "write me a LinkedIn post". That produces AI slop, and everyone reading this knows that smell from a hundred paces. The version that works is: "Here is my article. Pull out the three least obvious arguments. Write a LinkedIn post in my voice that picks one of them and ends on a question that invites genuine disagreement." Different job, different output, different world.

I have written more in the last twelve months than in the previous five years combined. The quality has gone up, not down, because AI handles the mechanics and I spend my thinking time on the thinking.

So why haven't you seen this side of AI?

Two reasons, and neither of them is the technology.

• The loudest AI use cases are the most visual ones. Videos of six-fingered humans get forty million views. A consultant asking a voice agent to summarise a competitor's quarterly results gets ignored, because it does not photograph well.
• The professionals quietly using AI to run circles around their rivals do not post about it. Publishing your workflow is publishing your edge. Most of the best operators I know are deliberately quiet about exactly what they have automated.

The result is a sector-wide distortion. The public sees the gimmicks. The people actually using AI are using it to do real work, quietly, and compounding the lead every week they do.

If any of the five examples above sounded more useful than the AI content you usually see, that is not an accident. The bottleneck is not AI. It is what you ask it. The firms that figure that out in 2026 will spend the rest of the decade looking like geniuses.

If you want to see a weekly competitive intelligence rhythm actually running on your business, book a Pulse Check. No pitch, no pressure, just a clear picture of what is possible.