Agentic AI in Financial Services: Scaling Trust and Efficiency
The non-scalable thing is what you do first, because you're never going to get to a more scalable motion, you're not going to have the power of the brand or the network or the capital to do something that's like truly scalable. So a lot of people outside in may look at this and say, why are you doing something that is super hard, you can't 10x or 100x, this is not like a scalable motion. It doesn't have to be, right? You could put 10, 20 advisory boards together and do this in like a much more bespoke, boutique-y type of way.
Just because you have it once, doesn't mean you'll retain it forever. And that's the big challenge that we've had, which is like, how do we continue to remain cycles ahead, not only in terms of the actual product, but being able to talk about the product in a way that makes it clear why we're cycles ahead of the other players in the space.
Having the best product and being cycles ahead is very important, don't get me wrong, and that is vital to have, but so is your ability to build a brand, rapport, a network that signifies trust and safety.
- Large language models (LLMs) can reason and understand in ways that people can, allowing complex operational processes, especially in financial services, to be redesigned for greater efficiency.
- For early-stage AI companies, focusing on recurring, complex, document and decision-heavy workflows that require human-level reasoning is a strong market entry strategy.
- Lending emerged as a prime target for AI automation because it involves numerous documents, significantly impacts 'time to quote' (a critical factor for winning business), and drives top-line revenue growth for financial institutions.
- In financial services, C-suite leaders are primarily concerned with growing their deposit base and the volume/quality of their lending, making these areas key drivers for top-line revenue.
- Pricing AI products on a volume-based subscription (e.g., per page of documents processed) at a fraction of the cost of human operators creates a clear and compelling value proposition for customers.
- Initial customer interviews can generate 'false signals' of interest; it's crucial for startups to differentiate genuine intent to buy from mere curiosity to avoid diluting focus and effort.
- In regulated industries like financial services, market adoption of new technology, especially AI, happens slowly, requiring a focused approach that builds trust and relationships.
- Non-scalable activities like forming advisory boards and fostering deep relationship-driven partnerships are often essential early-stage strategies to build trust and network, especially in industries where outbound channels are less effective.
- 50 buyers (The number of potential buyers at financial institutions Multimodal initially interviewed to identify their target niche.)
- 30-40% (The percentage of initial customer interviews where 'lending' emerged as a primary area of interest.)
- 30 days (The duration of the proof of concept (PoC) that led to Multimodal's first paying customer.)
- 10x (The target increase in growth rate within a year by leveraging both direct and channel sales.)
RevBots.ai View:
- AI Sprinkler stage: Multimodal's AI bolted onto lending workflows shows potential but lacks full integration.
- Focus on complex workflows aligns with ARM's emphasis on AI orchestration replacing legacy processes.
- Non-scalable trust-building tactics are crucial for AI adoption in regulated industries like financial services.
- Volume-based pricing models can drive adoption but may complicate scaling in the ARM stage.
Join The RevBots ARMy
The insider daily for Autonomous Revenue Masters.