Earlier this year, we worked with a Lagos-based B2B services firm that was growing fast — but drowning in manual admin. Their sales team was spending four hours every day on tasks that had nothing to do with selling: copying data between spreadsheets, manually sending follow-up emails, updating a CRM that nobody trusted because it was always out of date.
The problem in numbers
Before we started, we asked the team to track every manual task for one week. The results: 28 hours per week on data entry across three systems that didn't talk to each other. Average time from a new lead arriving to first follow-up contact: 18 hours. Leads that fell through the cracks because no one noticed them in the inbox: roughly 15% per month.
What we built
The automation we designed had four main components. First, a web form connected directly to their CRM (HubSpot) via API — no more copying emails into spreadsheets. Second, an automated email sequence triggered the moment a lead was created: an immediate acknowledgement, a follow-up with relevant case studies at 24 hours, and a check-in at 72 hours if no reply. Third, a Slack notification to the assigned sales rep the instant a new lead came in, with all context already pulled from the CRM. Fourth, a weekly pipeline report generated automatically and emailed to the management team every Monday morning.
The results after 60 days
- Manual data entry time: from 28 hours/week to under 2 hours/week.
- Average lead response time: from 18 hours to under 12 minutes.
- Leads falling through the cracks: from 15% to under 1%.
- Sales team capacity freed up for actual selling: estimated 26 hours per week across the team.
What we learned
The most valuable lesson from this project was that the technology was the easy part. The hard part was process design — understanding exactly what "a lead" meant in their context, what information needed to be captured, what different lead sources looked like, and what the sales team's actual workflow was (which differed significantly from what the management thought it was).
We always spend at least a week on process mapping before we write any automation code. Rushing that phase is how you automate a broken process and make it run faster — which is worse than doing nothing.