Task automation is counterproductive when you have less than 100 active clients

The importance of direct contact with early users

When you're in pre-PMF with 10 beta testers, the main challenge is to talk to them directly, conduct unbiased interviews, test hypotheses very quickly, spend maximum time with users and potential prospects to really understand the problem you want to solve - not to automate tasks that aren't even taking up much time yet.

A concrete example: B2B Accounting SaaS

To illustrate this, I'll take the example of a B2B accounting and financial management SaaS for small and medium businesses.

In January 2024, we launched the MVP after 1 month of development. In January, we onboarded our first 15 clients, everything was done manually, the founder and product manager spent 2 hours per day talking with users and onboarding others. In January, no automation was created for NPS, we had no support, let alone a chatbot, or onboarding screens.

"In our MVP, no automation was created for NPS, we had no support, let alone a chatbot, or onboarding screens."

Progressive Evolution

In February and March, we onboarded 1 client every 2 days (that's 40 in 2 months); still manually. But this allowed us to perfect our onboarding, what to say and not say; what's important and what's not; the key elements for good onboarding.

"50 clients and still nothing automated"

First Automations

It wasn't until April that we started suffering from manual onboardings, follow-ups, and managing our community of 50 people. We started to see a decrease in our service quality, ⅔ clients began to complain.

We took 2 very simple actions:

1. Add a modal after Sign up to guide clients on the first steps to take when they arrive in the application
2. Send our step-by-step guide via email automation through Brevo rather than sending it manually

Communication Channels

In May, June, July, and August, we continued to discuss and exchange with our users through three main channels:

• Crisp support: to quickly respond to blocking issues
• The WhatsApp group for our early users
• Continuous interviews

Implementing NPS

That's how in September, with 180 paying clients, we decided to implement a very simple NPS.

"10 months later, in November, we still don't have a chatbot for support. It's still too early."

Conclusion

A startup works very well, and even better without automation of its support, NPS, or onboarding at the beginning because it forces us to talk to users in the field and maximize our touchpoints with them.

This is crucial for launching solutions and improvements that are relevant at the start.