Start With the Most Annoying Thing: How SMBs Should Adopt AI

We did not build a contact management AI agent because we had a vision for the future of CRM. We built it because keeping track of contacts was genuinely maddening — and we got tired of losing to the chaos.

Emails, WhatsApp threads, business cards photographed and forgotten, event notes in a notebook, phone calls with no follow-up. Every person we met lived in a different silo. Nothing connected. Nothing updated itself. And every time we needed to find something, we wasted twenty minutes we did not have.

That frustration was the brief. Not a strategy document. Not a boardroom transformation plan. Just: this is broken, let us fix it.

What the agent actually does

The agent works through text, screenshot, or audio. You share information about a person — a note from a call, a photo of a business card, a voice memo after a meeting — and it does the rest.

It links that person to related contacts, companies, events, and threads it already knows about. It fills in gaps, cleans duplicates, and connects dots across sources you would never manually cross-reference. Every morning you get a summary: who you should follow up with, what you said you would do, what has changed.

You can draft and send emails from inside the same interface. You can ask it who you met at a specific event, or which companies are connected to a contact you are about to call. No spreadsheet. No form. No copy-paste.

The result is not a productivity dashboard with impressive metrics. It is the quiet relief of not dreading a task anymore. That feeling is underrated as a business outcome.

Why boring wins over bold

There is a version of AI adoption that starts with the biggest possible question: where can AI transform our business? That question almost always leads to a six-month pilot, a consultant engagement, and a deliverable nobody uses.

The version that actually works starts smaller. What is the single most repetitive, zero-thinking task your team does every week? Find that. Automate it. Then move to the next one.

Research from Gusto's 2025 SMB study backs this up in a specific way: the median small business using generative AI reports only two regular AI use cases — and four out of five of those businesses still report productivity gains of 20% or more. You do not need ten AI initiatives to see real returns. You need two that actually stick.

80% of SMBs using AI say the number-one benefit is saving time — not saving money. Most save 11 to 20 hours per month, and over half reinvest that time into getting new customers. (Thryv, 2025)

Time is what most teams are actually short on. Not ideas, not ambition. Time.

A Thryv survey of small business AI users found that 78% believe AI is necessary to keep up with customer expectations. But the businesses making that math work are not doing it with grand deployments. They are doing it by removing the friction that was quietly eating their weeks.

The morale argument nobody makes loudly enough

Admin debt is not just a time problem. It is a morale problem.

When your team spends hours on work that requires no judgment — data entry, contact cleanup, scheduling follow-ups, formatting reports — they are not just losing time. They are losing the sense that their work matters. That accumulates. People start protecting their energy by doing less of everything, including the high-value work you actually hired them for.

Automate the annoying work and something shifts. The team has capacity again. They come to meetings less depleted. They think more clearly about customers, about product, about the relationships that require real human attention.

We have seen this pattern with clients: the first AI win is always operational. The second and third wins, the ones that actually change the trajectory of the business, only become visible once the team has bandwidth to see them.

How we advise clients to start

We do not open with "where is AI most transformative for your business?" That question invites speculation and delays action.

We open with: "What is the most frustrating, repetitive thing your team does every week — the thing everyone silently dreads?" That question gets a real answer, fast. Usually in the first five minutes of the conversation.

Common answers: manual CRM updates, chasing internal approvals, reformatting data between systems, summarizing meeting notes, writing the same emails over and over. These are not glamorous problems. They are also not hard to solve with a focused AI implementation.

The Gusto research makes a useful point here: most SMBs do not have a formal AI strategy, but the ones that do see the largest productivity and revenue gains. A strategy does not have to mean a 40-page document. It can mean a clear decision: we will fix the three most painful admin tasks before we touch anything else.

Start with a specific process. Define what done looks like. Ship something that works in the real workflow, not in a demo. Then measure the time recovered and ask what the team wants to do with it.

Start with the frustration, not the ambition

AI adoption among SMBs jumped 41% from 2024 to 2025, according to Thryv's research — from 39% to 55% of small businesses using AI in some form. That is a real shift. But only 22% of SMBs are using generative AI regularly, per Gusto's data. The majority are still experimenting or using it occasionally.

The gap between experimenting and regular use is almost always a targeting problem. Teams try AI on a task that is too vague, or too high-stakes, or too far from the daily workflow — and when the result is mediocre, the whole thing stalls.

Pick the most annoying thing. The one that causes sighs in team meetings. The one everyone has a workaround for. The one that piles up on Friday afternoons.

Solve that. Get the relief. Then ask what is next.

The compound effect of eliminating three or four genuinely painful admin tasks is not three or four small wins. It is a team that has real capacity again — and a business that can finally direct that capacity toward the work that actually moves things forward.

We started with contacts because contacts were maddening. That choice was not strategic. It was honest. And it turned out to be exactly right.