Tavus
30 Qualified Meetings in 25 Days for Tavus
Tavus is a developer platform for real-time conversational video AI. Their Conversational Video Interface lets a developer drop a photorealistic digital human into a product and have it hold a live, face-to-face video conversation with a user. It handles speech, vision and turn-taking in one pipeline, so the agent can read facial cues, know when to pause and when to speak, and respond fast enough that the conversation feels like a call rather than a chatbot with a face on it. It is genuinely new technology, which is the interesting part of this engagement, and also the hard part.
The Challenge
Nobody has a mental model for the product
Most outbound sells a better version of something the buyer already understands. Faster monitoring. Cheaper storage. Fewer support tickets. Tavus does not have that advantage: a conversational video agent is not a better version of anything the prospect currently owns. Ask someone to give up thirty minutes for a demo of a category they cannot picture and the honest answer is no.
This is the core outbound problem for any genuinely new category, and it is where email struggles most. An email gets one shot at explanation, in a format the reader skims, with no ability to answer the obvious follow-up question. A phone call has a different shape: a rep can explain what it is in one sentence, hear the confusion, and clear it up immediately. For a new category, the ability to answer "wait, so what does it actually do?" in real time is worth more than any amount of copy testing.
Wide, not deep
Tavus did not want one territory covered properly. They wanted reach across North America, the UK, Europe and beyond at the same time. That is unusual and it breaks the standard playbook: most outbound programmes pick a market, build the list, learn the market, then expand. Running eleven countries in the first month means no single market gets the benefit of accumulated local learning, and the calling day never really ends.
The window was short
Tavus needed meetings quickly, and that constraint alone decides the channel. Cold email cannot start on day one: sending domains need warmup, small and gradually increasing volumes over two to three weeks, before any cold volume goes out. So an email-led programme spends the first month producing nothing by design. Calling has no warmup. A researched list and a trained caller produce a booked meeting on day one.
“A conversational video agent is not a better version of anything the prospect already owns. Email gets one shot at explaining that. A call gets a conversation.”
The Approach
Phone first, because the product needs a conversation
A-Sales put a calling team on Tavus from the first week. No infrastructure build, no warmup wait, no setup period billed as progress. The channel choice was not just about speed: for a product this new, the phone does something email structurally cannot, which is answer a question the moment it forms.
One team, eleven countries
Rather than splitting into country teams, A-Sales ran a single calling operation across the full target footprint, with calling windows set to each prospect's local working day. In the first 25 days that produced meetings in Canada, the United States, the United Kingdom, Italy, Spain, Poland, Denmark, Romania, the Czech Republic, Saudi Arabia and the United Arab Emirates.
Volume first, then quality, then show rate
This is how A-Sales sequences every new account, and it is worth stating plainly because most agencies pretend the first month is something it is not. Volume first: until an account is producing meetings daily, nothing else can be measured, because there is no quality signal from four data points. Then quality: once meetings are flowing, the client's feedback on each one becomes the most valuable input in the account. Which titles converted into a real conversation, which company profiles wasted the AE's time, which segments to cut entirely. Then show rate: with the right people booked consistently, the last lever is making sure they turn up, and turn up warm.
Tavus is currently in the second phase. The first weeks produced volume across a wide list, the client reviewed every meeting, and the targeting was tightened in response: headcount floors were raised, seniority requirements moved up to VP and above, and whole categories of account that looked plausible on paper were cut once the calls showed they were not. That is not a problem in the engagement. That is the engagement working.
The client's feedback loop is the product
Every meeting booked for Tavus is reviewed by the client, and notes come back on each one. Those notes go straight into the targeting brief and the calling script, weekly rather than quarterly. A-Sales asks for rejections: a meeting the client did not value is more useful than one they did, because it tells you where the boundary of the ICP actually sits rather than where you assumed it was.
“Volume first, then quality, then show rate. Any agency that tells you month one is about quality has not run enough month ones.”
The Process
Week 1: live
- Target lists built across North America, the UK, Europe and the Gulf
- Qualification standard agreed in writing with Tavus
- Calling team assigned and dialling in the first week
- No warmup dependency, no setup period
Weeks 1 to 4: volume
- Calling as the primary channel across all markets
- Calling windows set to each prospect's local working day
- Script rebuilt weekly against live objection data
- Daily booking rate established
Weeks 3 onward: calibration
- Every meeting reviewed by Tavus, with notes on each
- Headcount floors raised and seniority requirements moved to VP and above
- Account categories that underperformed removed from the list entirely
- Targeting brief updated weekly rather than quarterly
Ongoing
- Pre-meeting context added so prospects arrive understanding how they would use the product
- Callback list worked daily to recover no-shows
- Continuous tightening of ICP against live results
How We Define a Qualified Meeting
A meeting only counts when it meets criteria agreed with the client before calling starts. Right company profile, right seniority, genuine interest, confirmed slot, and a prospect who knows who they are meeting and why.
If a meeting does not meet the standard, it does not get billed. Tavus can reject a meeting, and rejections are reviewed rather than argued about.
On a young account this matters more than on a mature one. The standard moves in the first weeks as both sides learn what a good meeting actually looks like, and a performance-based model is the only structure where that is a shared interest rather than a fight. A-Sales earns the overwhelming majority of its revenue from the per qualified sales meeting fee, so a meeting the client does not value is a loss for us too.
Results
Current as of the first 25 days of a live, ongoing engagement.
- 30 qualified sales meetings booked in the first 25 days
- 29 of 30 booked by phone
- Prospects across 11 countries in North America, Europe and the Gulf
- Live and booking within the first week of the engagement
- More than one meeting per working day from a standing start
- Engagement ongoing
Countries reached: Canada, United States, United Kingdom, Italy, Spain, Poland, Denmark, Romania, Czech Republic, Saudi Arabia, United Arab Emirates.
What Made the Difference
New categories need a conversation, not a pitch
When the buyer has no reference point for the product, the ability to answer their first question in real time is the whole game. Email cannot do that. This is the strongest argument for calling that exists, and almost nobody makes it.
Speed comes from the channel, not from effort
Being live in week one is not about working harder. It is about picking the one channel that has no warmup requirement. Every agency that promises a fast start and then leads with email is promising something the infrastructure will not allow.
Breadth is a choice with a cost, and it can be the right one
Eleven countries in a month means no market gets deep local learning. For a client who needs to know where the demand actually is, that trade is worth making.
Client feedback beats upfront research
Four weeks of real meetings, reviewed honestly by the client, teaches you more about the ICP than four weeks of desk research. The agencies that resist that feedback are the ones that stay wrong for longer.



