The DigiKat Learning Center

Where Your 10 Hours of Admin Actually Go | DigiKat

Written by Mick Goman | Apr 13, 2026, 9:30:00 PM

There's a gap between what LinkedIn says AI is doing for small businesses and what I see on the ground.

On LinkedIn, AI is replacing whole teams. In the businesses I work with, most owners have tried ChatGPT for an email or two and stopped there. They're sceptical, and they're right to be. Most of the hype doesn't apply to a team of 10 to 100 people.

What does apply is much less exciting. It's the admin. The hours your team loses every week to typing, chasing and checking. That's where AI earns its keep right now, and if you're on HubSpot, you're closer than you think.

Where the hours actually go

When we start working with a business, we map where the team's admin time goes each week. Every business is different, but the categories come up again and again. Here's what a fairly typical week looks like when I add it up with an owner. Treat the hours as an example, then do the same exercise with your own numbers.

Data entry: 2–3 hours

Typing contact details into the CRM. Updating deal records. Logging call notes. Copying information from one system to another. Every web enquiry has to be checked, tidied up, assigned and entered.

Follow-up emails: 2–3 hours

Recapping meetings. Nudging prospects who've gone quiet. Answering the same enquiry for the tenth time. The emails are slightly different each time, and that's why they never get templated.

Reporting and pipeline reviews: 1–2 hours

Pulling numbers together for the Monday meeting. Working out which deals moved and which stalled. Trying to tell which marketing is actually working.

Lead qualification and routing: 1–2 hours

Looking at new leads and deciding who should handle them. Checking whether they're a fit. Catching up on the enquiries that came in over the weekend.

General admin: 1–2 hours

Booking meetings, sending reminders, updating task lists, merging duplicate contacts. None of it is hard. All of it takes time.

Add it up and you're looking at something like 7–12 hours a week. Very little of it needs human judgement or relationship skills. That's the time worth taking back.

What AI can take off you

Here's where I'd start, mapped to the categories above. Most of this uses tools already in HubSpot, depending on your tier.

Data entry: let the system log it

Connect email and calendar so activity logs itself. Use HubSpot forms so enquiries land as structured records. If you're on Sales Hub or Service Hub Professional or Enterprise, HubSpot can record and transcribe calls and generate an AI summary, so the rep reviews a summary instead of writing notes from scratch.

Follow-ups: AI drafts, you send

HubSpot's built-in AI can draft follow-up emails, sequences and marketing emails. You still review and edit. Starting from a decent draft is a lot quicker than a blank screen. For routine messages like "thanks for your enquiry, here's what happens next", a workflow can send them instantly.

Reporting: build it once

Set up a dashboard with the handful of numbers you actually act on, and schedule it to land in your inbox each Monday. Fewer, better reports. I've written about why most dashboards don't help.

Lead routing: score, then assign

HubSpot's lead scoring tools can rank leads on fit and engagement, and workflows can assign them to the right person with a notification. On Enterprise, predictive scoring adds a likelihood-to-close estimate based on your historical data.

General admin: clean as you go

Use meeting links so prospects book themselves. Use HubSpot's duplicate management tools regularly. Set workflows to create tasks automatically when deals move stage.

For a practical walk-through of the workflows behind this, see The 5 HubSpot Workflows That Replace 20 Hours of Manual Work Per Week.

Where AI agents fit

Everything above is assisted automation. A person is still in the loop for most steps. The next layer is AI agents: software that handles a whole task, such as answering a routine support question or researching a new lead, and hands over to a human when it should.

HubSpot has its own Breeze agents for things like customer support and prospecting, and you can build custom agents that work alongside your CRM. We did this ourselves: I wrote up what happened when we replaced a hire with an AI workflow, including the parts that didn't work.

The rule we use: AI suggests, human decides for anything that matters, and AI handles it for routine, low-risk tasks. An agent can answer "what are your opening hours?" on its own. It shouldn't send a quote without someone checking it.

Where AI isn't ready for most SMBs

  • Strategy. AI can carry out a plan. It can't tell you what your business should focus on.
  • All your content. Fully AI-written content reads as generic. Use AI for drafts and outlines and keep a human edit.
  • Closing deals. AI can score and prioritise. Winning the deal is still about trust and relationships.
  • Fixing a broken process. If nobody knows how leads should be handled, AI won't work it out for you. It will just do the wrong thing faster.

"I've heard the AI hype before"

Fair. Here are the objections I hear most.

"It's only for big companies." Most of what I've described is built into HubSpot plans SMBs already use. The bigger cost is the time to set it up properly.

"My team won't use it." The best automation is close to invisible. Leads get assigned, notes get logged, tasks appear. Your team just sees less admin.

"AI makes mistakes." It does. That's why the human sign-off points matter, and why you start with low-risk tasks.

"I don't have time to implement it." Start with one thing. One process, one tool, one month.

How to start

  1. Measure your week. Ask each person to track admin time for one week, using the categories above. Rough is fine.
  2. Pick the biggest time drain. Just one process.
  3. Check your data. AI works on what's in your CRM. If records are incomplete or inconsistent, fix that first, or the automation will inherit the mess.
  4. Automate it, with a human checkpoint. Run it for 30 days and measure the hours again.
  5. Build on what works. Got a win? Expand. Didn't work? Find out why before adding more.

To put a dollar figure on it, multiply the hours you save by your team's loaded hourly cost, then by the weeks you work in a year. Use your own numbers. That's the figure that matters.

Talk it through

If you'd like help working out which parts of your week an AI agent could take on, and which should stay with your team, have a look at how we build AI agents for HubSpot users. We'll start with your actual week.