If I were an Account Executive today, I would build pipeline four times faster than I did at Google Cloud, by giving AI four jobs: account researcher, cold list builder, warm leads detector and copywriter. Not to replace the selling, but to delete the weeks of manual work that used to sit before it. Here is what each job looks like and how to run them in a week.

What selling at Google Cloud looked like without AI

I spent nearly five years in sales at Google Cloud as an Account Executive. The target went from 1.5M euros in year one to 10M euros in year four, a 100 to 120% quota increase every year. Miss it and welcome to pressure cooker mode.

And back then, everything was manual. Hours of research per account. Endless static list-building. Outreach written one email at a time. I was good at it, and it still ate most of the week before I got to a conversation.

In 2020, sales was about grind. Today, sales is about using AI as a lever. If you are in sales and not using AI daily, or at least learning how it can help you, good luck. The reps next to you are getting to the conversation four times faster.

The 4 jobs I would give AI today

1. The account researcher

Annual reports, case studies, earnings calls, competitor benchmarks, the CEO's last three interviews. At Google that was two hours per strategic account. Today an AI chat does the first pass in minutes: what the company said its priorities are, what changed since last year, which of my case studies matches.

The rule: research is preparation for a call, not a substitute for it. Be prepared for each call in minutes, then make the call.

2. The cold list builder

Tools like Clay can build and prioritize best-fit accounts based on tech stack, current job openings and website traffic. This is the right place to apply firmographic intelligence: deciding which 50 of the 500 accounts in your territory deserve attention this quarter. Focus on the best accounts, not all of them.

A warning from experience: this is powerful and it is also a rabbit hole. If the AE becomes the list engineer, the AE stops selling.

3. The warm leads detector

This is the job I did not have at Google and wish I had. Past clients. Website visitors. Prospects who engaged with a competitor's content. Contacts who just changed jobs. These are the warmest leads in any territory and in our experience they convert about 3x better than cold accounts.

They are also mostly already in the CRM. Reps use about 20% of the signals sitting there. At Sonik, 50% of our own pipeline comes from signals that were already in the CRM. The AI's job is to surface them every morning so the AE focuses on the right champions.

4. The copywriter

Not to write the email. To draft context-rich messages that tie a CEO's strategic objective directly to a case study, so that the one sentence I then rewrite by hand is the right one. No more "congrats on raising" emails. Relevance at scale, with a human finishing the sentence.

A weekly routine that puts the four together

1. Monday, 45 minutes: the territory list. Run the list builder on your accounts. Pick the 20 to 30 that matter this week.
2. Every morning, 10 minutes: warm leads. Review the detector's list of visitors, job changes and returning contacts. Those are your first calls of the day.
3. Before each call, 3 minutes: research. One prompt per account, one page of context, the trigger you will open with.
4. Calls at 10am and 4pm. Across the 1,000,000 B2B cold calls we analyzed with Katana, 10am connects at 14% and 4pm at 15.8%, against 9.8% at 8am. Founders pick up more around 10am.
5. After the calls: messages. Draft follow-ups and new-account messages with AI, rewrite the key sentence yourself, send.

The ratio to protect: AI does the hours, the AE does the conversations. If the week ends with more prompts written than calls made, the routine has inverted.

Common mistakes AEs make with AI

• Researching instead of calling. A perfect account brief with no conversation is a well-documented miss.
• Letting AI send. Fully automated outreach reads as fully automated outreach. Keep the human sentence.
• Building lists with no intent. Tech stack and headcount tell you fit. They do not tell you the moment. Warm signals do.
• Opening with the pitch anyway. 65% of reps open by explaining what they sell. With all that research, open on the trigger instead: 7-10% meeting rate versus 2-3%.
• Using AI for everything except the conversation prep. Discovery questions starting with "how" lift conversion by 190%. Prepare the question, not just the pitch.

How Sonik approaches AI for account executives

Sonik does job three, and part of job one, for the rep. Signal agents scan the CRM, website and social activity every morning and build a ranked list of warm leads: past clients, visitors, job changes, closed-lost, competitor engagers. AI call prep assembles the reason to reach out and the account context before each connect.

Then the parallel dialer turns that list into 5x more live conversations than manual dialing, so the research actually gets used in a conversation the same day. The AE's job stays the AE's job: listen, challenge, close. More on the signals and dialer pages.

Conclusion

AI is my researcher, list builder, leads detector and execution arm. It saves me weeks of manual work, and it gives those weeks back to conversations. If your sales process does not include AI today, the gap to the reps who use it is widening every quarter. This week, give AI one of the four jobs, the warm leads detector if you can, and count how many of this week's calls came from a signal rather than a cold list.