Podcast Episode Summary

Voice AI that Delivers: Synthflow AI’s Hakob Astabatsyan on Solving the Last Mile Problem

Podcast: The SaaS Revolution Show
Host: Alex Theuma
Guest: Hakob Astabatsyan, CEO of Synthflow AI
Date: October 17, 2025
Recording Venue: SaaS.Europe 2025, Dublin

Episode Overview

In this episode, Alex Theuma interviews Hakob Astabatsyan, CEO of Synthflow AI, during SaaS.Europe 2025 in Dublin. The discussion centers on building and scaling Synthflow AI, a company delivering AI-powered voice agents for contact centers, and dives deep into the so-called "last mile problem" in Voice AI—ensuring that AI deployments in customer-facing roles deliver real, lasting value and overcome integration hurdles.

Hakob shares candid insights about rapid growth in the AI era, the shifting rules of SaaS go-to-market, and how Synthflow navigated from a generic API provider to a focused enterprise voice AI operating system. The conversation also covers challenges of product-market fit, internationalization strategies, and the daily tools that power Synthflow’s operations.

Key Discussion Points & Insights

1. Synflow AI’s Growth Journey and Product Evolution (02:11 – 06:36)

• Initial Traction & Focus:

  • Launched in 2023; experienced 15x revenue growth in the first year, processing millions of calls monthly.
  • Pivoted from a broad Voice AI API provider to a focused AI contact center solution, serving businesses and enterprise contact centers only.

• Main Value Proposition:

  • AI agents automate customer support/sales calls on the phone, offering “an AI contact center on the phone…working 24/7…at a fraction of the cost.”

    “Are you going to automate our entire customer support? Basically we have hundreds of agents on the phone. So you’re saying we can put AI on the phone working 24/7 answering these phones at a fraction of the cost…” – Hakob Astabatsyan (04:21)

• Rule Changes in AI Era:

  • Demand generation for AI solutions is “virality driven” due to inherent market curiosity/budgets for experiments.
  • “The rules of growth have changed. However, the rules of building a company haven’t changed…hiring great people, right?” – Hakob Astabatsyan (02:57)

2. Marketing, Go-to-Market, and Delivering Real Value (03:48 – 06:36)

• Marketing in the AI Era:

  • Being visible to customers with experimental budgets is “80% of the job” in early-stage AI SaaS.
  • However, simply generating demand isn’t enough; long-term value delivery is essential.

• The “Last Mile” Problem:

  • Many AI POCs (proofs of concept) fail to deliver sustained value.

  • Synthflow addressed this by narrowing its focus and delivering true end-to-end automation for contact center use cases, including post-call actions like updating CRM tickets (RPA 2.0).

  • Strict prioritization: Saying “no” to non-target use cases/customers to maintain product focus and customer success.

    “One of the hardest things for me as a founder was…to start saying no to some customers, right? …But for those that we focus on…we’re building a solution and also building our organization tailored towards these customers.” – Hakob Astabatsyan (09:13)

3. Forward Deployed Engineers and the Last Mile (10:12 – 12:33)

• Role Explained:
  Forward deployed engineers (FDEs) are specialized solutions engineers on the front lines of customer integration and agent customization.

  “It’s not different from solutions engineers or sales architects from before AI…these are folks who work on solving the last mile problem. Basically crossing that river of the last mile is forward deployed engineer’s job.” – Hakob Astabatsyan (10:20)

• Main Functions:

  1. Integrations: Ensuring AI agents connect smoothly into highly varied enterprise tech stacks (“Where is your data?”).
  2. Agentic Fine-tuning: Training and calibrating agents—speech speed, tone, response time, and custom actions—using a no-code UI so clients can tailor the behavior of agents easily.

4. Internationalization and Scaling Beyond Berlin (12:33 – 15:31)

• US Market Focus:

  • Core commercial team and forward deployed engineers based in the US (despite headquarters in Berlin).
  • Initial focus on English-speaking markets due to technology limitations, followed by expansion into German market as models improved.

• Changing Expansion Playbooks:

  • The “classical playbook” of starting local then expanding doesn’t always apply with global-ready tech like AI.
  • Raised multiple funding rounds (latest: $20M from Accel) to enable broad market reach and team expansion.

    “There is no reason to restrict to certain geographies. You can go global from day one if you have the technology.” – Hakob Astabatsyan (14:29)

5. AI Tools in the Modern CEO’s Toolbox (15:31 – 17:54)

• Hakob’s Personal Stack:

  • Still relies heavily on ChatGPT for writing and storytelling.

  • Uses Synthflow’s own AI assistant for automating inbound calls (a solution to being “bombarded by sales calls”).

  • Tools tested/used: LinkedIn outreach tools, Clay, N8N (now integrated natively with Synthflow for workflow automation).

  • Big proponent of combining N8N and Synthflow for advanced RPA-style automation.

    “We actually have a Synthflow native integration now with N8N…customers actually building the workflows with N8N and Synthflow embedded into workflow…a very interesting combination.” – Hakob Astabatsyan (17:12)

6. Future Challenges and Opportunities in Voice AI (17:54 – 20:29)

• Technological Milestones:

  • 2023-24: Main concern was latency and if voice AI could match/exceed human performance.
  • 2025: “We crossed that cusp.” Speech-to-speech models are now viable, reliable—industry validation.

• Next Hurdles:

  • Long-form Conversations: Moving from short calls (60–90 sec) to handling extended, natural conversations (30–60 min) poses new robustness challenges.

  • Productizing Workflows: Building more seamless end-to-end automation so customers can easily customize and leverage complex workflows; product announcements pending.

    “Now…can you have like a 30-minute or one hour conversation actually with AI, the long format? And of course, the longer… the higher is the probability that something can go wrong. Right? And I think this is one thing that I’m observing very closely.” – Hakob Astabatsyan (19:05)

Notable Quotes & Memorable Moments

• On Choosing Focus over Chasing Revenue:
  “One of the hardest things for me as a founder was to start saying no to some customers… Honestly, I’ve told that to some customers in the calls—will not be able to serve your use case properly.” – Hakob Astabatsyan (09:13)

• On AI Go-to-Market and Demand:
  “Being out there making your name is the most important thing.” – Hakob Astabatsyan (05:33)

• On Forward Deployed Engineers & The Last Mile:
  “Crossing that river of the last mile is forward deployed engineer’s job.” – Hakob Astabatsyan (10:20)

• On AI’s Readiness for Real Business Needs:
  “Everyone knows now, OK, voice AI works…for me, this question is answered, we have crossed the cusp…it’s gonna work. There are a lot of details…but it’s gonna work.” – Hakob Astabatsyan (18:20)

Timestamps for Important Segments

• Introduction & Background: 00:29 – 02:11
• Synflow Growth Journey: 02:11 – 06:36
• Marketing & Value Delivery: 03:48 – 06:36
• The Last Mile Problem: 06:36 – 10:12
• Forward Deployed Engineers: 10:12 – 12:33
• Scaling & Internationalization: 12:33 – 15:31
• AI Tools for CEOs: 15:31 – 17:54
• Future of Voice AI & Challenges: 17:54 – 20:29

Episode Takeaways

• The “last mile problem” remains the critical challenge for AI SaaS: vendors must ensure not just proofs-of-concept, but true, enduring customer value through specialization, tight integrations, and post-call automation.
• Success in AI SaaS requires balancing the fast-changing go-to-market tactics with timeless company-building principles: focus, talent, and organizational alignment.
• International scaling is possible from day one in AI, but demands targeted investment and localized execution.
• The next frontiers in voice AI are robust long-form conversations and fully productized workflow automations—setting the stage for further disruption in customer service and beyond.