Automated prospecting turned an eight-hour job into a ten-minute one. Verifying and enriching 200 contacts used to take me a full working day. The same work now runs in the background while I think about something else.

The speed is the obvious part. The change that mattered more was what it freed up. Here is what the manual version actually cost, and what replaced it.

What the manual version cost

From early 2021 to mid-2022 I worked as a market research analyst, back when research automation meant having Excel formulas that worked.

The job was to build prospect datasets, research companies and contacts, and hand the sales team something usable. In practice that meant hours building and cleaning lists, then examining every company website, reading every contact’s LinkedIn profile for an angle, and cross-referencing news, funding announcements, and hiring activity looking for a hook.

On a good day I could fully validate and enrich 200 to 300 contacts. That is roughly the number of emails it takes to produce one conversion in an average campaign. Reaching 1,000 contacts a day needed four or five analysts. The role existed because sales teams were drowning in research instead of talking to anyone.

Where the time actually went

Verification was the worst of it. Copy an address, paste it into a verifier, wait, copy “safe to send” back into the sheet, repeat 299 more times. Run out of credits on one tool and start the whole dance again on another.

The bigger sink was the hunting. Sometimes I found a genuine angle, or spotted that a company was in expansion. More often I came up empty after hours of digging, and the sheer volume meant I was constantly missing things that would have made the outreach better.

It was a permanent trade between volume and depth. We needed the numbers, going deeper would have made the campaigns better, and there was never enough day for both.

What replaced it

Tools like Clay do that work in minutes, at a scale I could not have run by hand.

It is also more thorough than I was. While I checked one LinkedIn profile at a time, these tools pull from dozens of sources at once, cross-reference them, and build a richer profile than I could have assembled with a full day spent on a single contact.

Verification happens in the background now. It is not a task anyone schedules.

What the automated version actually is

It is worth being concrete, because “automation” gets used to mean anything. The pipeline that replaced four analysts has four stages, and none of them are clever on their own.

1. Apollo builds the list. Filters on industry, seniority, company size, and location, run until the set is large enough to be worth processing.

2. Clay cleans it. Deduplication, company name normalisation, and email verification, with anything that is not safe to send dropped rather than risked.

3. Enrichment runs against whatever the campaign needs. Company research, a summary of what the business actually does, and any signal worth writing about.

4. A model drafts the personalised line from that research, and a person reviews every variation before the campaign sends.

The important part is that stages two and three used to be the entire job. Now they are infrastructure. The analyst work that survived automation is stage one and the review at stage four, which are the two places where a judgement gets made.

That is the honest shape of it. Automation did not remove the research. It removed the retrieval, which was never the part that needed a person.

The part that mattered more than speed

What I did not expect was getting the attention back.

When you are not collecting, cleaning, and verifying, you can think about what the data means. Which patterns are showing up, what they say about the ideal customer, and which insight is worth building a campaign around.

Experimenting also got cheap, and that is the real shift. Testing a new segment or a fresh personalisation angle no longer costs half a day. When an experiment costs four hours you run one a week. When it costs ten minutes you run six before lunch, and you learn faster than the market moves.

The repetitive work is handled. The energy goes into working out why a message lands with one segment and dies with another.

Looking back

I liked the analyst job, grunt work included. There was something satisfying about the detective part of it. I am not nostalgic for the inefficiency.

The tools did not just save time. They moved entire teams off tactical execution and onto the work that decides whether outreach succeeds at all. Instead of a room of analysts copy-pasting through a prospect list, two people can focus on the thinking.

And this is early. If the jump from 2021 to now feels sharp, the next couple of years will not be gentler.