How do you get your first 20 B2B customers from zero? In this recorded
session, Komplai co-founder Rishi Ayyalasomayajula breaks down the exact 0 to 1 GTM
playbook he used to grow an AI-native accounting firm, with no ad spend and a
one-person sales team. Hosted by Sayanta Ghosh, CEO and co-founder of nRev, it is a
working session on early-stage go-to-market: how to land your first customers, when to
switch from manual to automated outbound, why selling outcomes beats selling software,
and how to price when there is no playbook.
Key learnings
- Find the muscle before you automate. The first 10 customers came purely from
network at zero CAC, and their job was to reveal the vertical and persona that feel
the pain most. Automate before that and you are just praying at scale.
- A market that needs no education is an unfair advantage. When buyers already know
the problem and that AI can solve it, you skip awareness-building and compete only on
why you are different and how fast you prove value.
- Sell the outcome, not the tool. Running it as a services business on purpose,
where your team operates the product and the customer just consumes the output,
removes ramp-up and adoption risk and defuses the "will it replace my job" objection
that stalls AI deals in regulated functions.
- Engineer touches as one pipeline, not three campaigns. Seven points of recall
across email, LinkedIn, and calls only land when they reference each other and fire in
sequence, so it reads as organic follow-up instead of spam.
- Compress time to value, then move. A tangible outcome lets you close an NDA on the
first call and start a 30-day pilot the same week, instead of negotiating proof.
- Disqualify on business age. Companies older than 5 years have entrenched CA
relationships and rarely switch, so route them to network and point outbound at
younger businesses with shorter cycles.
- Price against the incumbent, not from scratch. Anchoring to what the buyer already
pays and coming in at roughly half gets a foot in the door, with room to raise later
once you are sticky.
- Do unscalable things early, on purpose. Founder-led calls set the baselines you
later automate against, and even a deal you cannot close often returns three
referrals.
- Know your number and protect your time. Working backwards from contract value to a
customer-count target tells you exactly when a stalling or ghosting lead is no longer
worth chasing.
Questions this session answers
How do you get your first 10 B2B customers with no budget?
Through network at zero CAC. Rishi closed his first 10 customers entirely through
his own network and network-of-network referrals, with no ad spend. The point of
those early deals was learning: identifying the vertical and persona that felt the
pain most before turning on any paid or automated outbound.
Should an early AI startup sell software or services?
Rishi argues for services early. He intentionally ran Komplai as a services
business so his own team operates the product and the customer just consumes the
output. That removes adoption risk and ramp-up time, and lets him sell outcomes
instead of teaching customers a new tool.
How many customers do you actually need to hit $1M ARR?
Work backwards from contract value. At roughly a $5,000 monthly contract, that is
about 200 customers to reach $1M. Knowing that number dictates how he spends every
sales hour and when to walk away from a stalling deal.
Free pilots or paid pilots: which one converts?
At his stage, Rishi runs free 30-day pilots tied to a tangible outcome, such as
closing the books in 30 days, rather than charging for the pilot. His view is to not
build a business around a $1,000 pilot when the goal is a $25,000 contract. Get the
foot in the door, prove value fast, then move to a service contract.
How do you build trust selling AI into finance and accounting?
By selling outcomes, not software. Instead of asking the buyer's team to adopt a
tool, his team delivers the result faster than the in-house team can, with zero
dependency on the buyer. Trust comes from a delivered outcome, books closed by month
end, rather than a product demo.
When should a founder stop running sales calls themselves?
Not until they have established baselines. Rishi did the early calls and manual
outreach himself to learn what works and set performance benchmarks, then automated
only the paths that proved repeatable. Founder-led selling first, automation after the
muscle is found.
What does a real founder-led GTM stack look like?
Consolidated. Rishi replaced a stack of point tools, including Clay and Sales
Navigator, with nRev for lead enrichment, email and LinkedIn sequencing, and
phone-number enrichment, keeping only a separate email-sending tool. One pipeline
across email, LinkedIn, and calls instead of three disconnected campaigns.
.png)
