Sunil Karkera, Founder Soul of the Machine, is leveraging agentic AI to outpace much larger companies. "We solve boring problems and it's exciting," said Karkera.
Soul of the Machine has migrated SAP in 90 days and implemented a voice-based LLM augmented factory and production planning system in days. Karkera's services are completely agentic based with engineers doing the work up front.
"Everybody is in the US and we are forward deployed engineers. We work directly with the customers. Engineers, strategists and designers are totally vertically integrated," said Karkera, speaking at Constellation Research’s AI Forum in Washington DC.
- AI is your new co-founder, core and creative and engineering muse
- AI Exponentials Will Rule the World
- Exponential Efficiency in the Age of AI
- The road to AI Exponential will be bumpy
Karkera also said AI has flattened the services model. He also doesn't believe in proofs of concepts and pilots. Prototypes can be created in that first customer meeting and can rapidly go to production. "We are using an entirely end to end AI toolchain," said Karkera. "Vibe coding is about 10% to 20% in the prototyping phase. Then it's basically deep architecture. Engineering AI is really hard because most of the work is context engineering and it's not straightforward."
More from AI Forum Washington DC:
- Google DeepMind’s Danenberg on emerging LLM trends to watch
- For AI agents to work, focus on business outcomes, ROI not technology
In other words, it's hard to keep it simple. Karkera said Soul of the Machine tries to avoid multi-agent orchestration to keep tools limited. "Once you use more than three tools, it goes all over the place. Ideally, it's one tool per agent," said Karkera. "If we do multi agent orchestration we do it handmade. There's no choice at this point."
According to Karkera, enterprises are going down the wrong route with proof of concepts.
"We have a rule that we don't do any POCs. We have left money on the table by saying no to POCs, because we want to embrace the problem and do it all the way, rather than explain how hard it is. One cultural thing is to go after a full problem, segment that problem, solve it all the way and put it in production. Don't dwell on fancy problems to solve instead of the real problems with ROI."
