Erik Kuld, CEO
June 24, 2024
Warning: sports analogy ahead, proceed at your own risk.
First: Formula 1
Check out the F1 Racing team leaderboard—what do you notice?
👀 The same manufacturer names show up multiple times—that’s because the vast majority of teams don’t build their own engines: Red Bull is powered by Honda, McLaren is powered by Mercedes, Haas is powered by Ferrari… etc.
It’s really hard to build a great F1 race engine that is performant and reliable. The obviously better ROI for most teams is to partner with engine builders than build their own. Partnering lets teams focus engineering investment where it’ll make the biggest impact: car design, race strategy, etc.
The same principle applies for SaaS Product leaders building with AI in 2024.
You need a great AI “engine” powering your product and customer outcomes … and you need it ASAP. But if you’re like most companies, building your own AI is not your core competency. Instead you want to focus on your “car” and to partner closely on specialized components rather than build every part from scratch.
That’s where we come in…
Why we started Inventive
Next-Gen Embedded Analytics
We’re now in the early days of a massive technology shift: conversational AI.
Previous major technology shifts—like the internet, cloud, and mobile—catalyzed the emergence of better, faster, and cheaper solutions to more problems thanks to the capabilities unlocked by the new transformational technology infrastructure.
Source: Coatue EMW/2024, Bg2 Pod
The same is happening now with conversational AI. Every company is starting their “AI transformation”, needing to leverage new AI super powers in its products or operations, and using next-generation AI-native tools to do so. Naturally, the vendor ecosystem has to evolve.
Embedded Analytics—the legacy market of solutions for embedding charts, dashboards, and reports into software—exists because building in-product analytics from scratch is deceptively hard, and not the core competency of most companies. Companies need to focus their finite technical capacity on sharpening their competitive edge, not re-inventing the analytics wheel.
The problem with Embedded Analytics is that it's never been good enough. Your customers always want more data in more custom dashboards, reports, alerts, etc. A great self-service experience is always just out of reach because of the gap between the semantic understanding of the domain, and the technical understanding of the data.
However, now for the first time ever, we can bridge this gap thanks to AI agents—like our embedded AI Analyst—and usher in Next-Gen Embedded Analytics for the AI era. Always available, flexible AI agent systems complete the job of Embedded Analytics by enabling any user to answer the long tail of questions via true self-service experiences that keep getting smarter and more integrated over time.
A powerful AI Analyst agent solves the problems of legacy tools
AI agents are flexible software systems that can reason, decide, and take action toward goals. Unlike traditional software tools that help you get your job done, AI agents can actually do the job for you.
For example, let's say you run a software company, and your customer needs a report that requires a custom filter not available in your product. They email their Account Manager who submits a request to your data team, who manually produces a custom report for the customer and passes it back down the line.
Instead, an AI Analyst "agent" system can understand the customer's request, ask for clarification, reason about the best way to solve the problem, run the query, generate the custom report, schedule it to send weekly on Mondays at 9 AM to the individual customer, ask for feedback, and continue to iterate on custom requests with the customer directly.
The manual work to complete the job and resolve the customer requests is handled by the AI Analyst agent system, rather than a team of humans manually communicating and working via traditional software tools.
To build or to partner: that is the question
If AI agents will transform the customer experience in such ways, shouldn’t we all build our own?
Depends. If your company is the rare Ferrari or Mercedes F1 racing team in its market, and has spent years refining its expertise in AI, then vertically integrating your entire AI stack might make a lot of sense. But most companies are not like this.
When deciding on your company's AI strategy for your customer-facing products, it’s important to consider: how fast you can launch your AI experience, how likely you'll meet user expectations, and how much effort is required to build and maintain it—all in the context of your core product and roadmap.
Everyone wants to build with AI, but getting a great user experience into production is much harder than most product teams think. Most engineering effort in building with LLMs today goes toward trying to wrangle non-deterministic LLMs via complex prompting, retrieval, fine-tuning, testing, evaluation, debugging, monitoring, and so on. This is because LLMs are so new and the software development lifecycle tools and best practices for building with them are still emerging.
As a result, it's usually very ineffective for companies to build their own AI agent-powered product experiences in-house. Especially at this early stage of this AI era when it requires a team of top-notch engineers to discover and invent some of these tools and best practices. All that work is outside the core competency of these companies.
It's usually more strategic to invest product effort on unique and valuable data, intuitive design, and deep integrations that automate valuable workflows. Picking a specialized partner can help you build better AI agents faster.
This is why today we’re launching our AI Agent Platform: a suite of tools that make it easier for companies to build, launch and manage custom AI agents throughout the development lifecycle. For example, we help product teams manage, monitor, and monetize an embedded AI Analyst agent for customers with tools for testing, evaluating, optimizing, monitoring, personalizing, integrating, and more.
Racing to the future
Just like in F1, most great companies don’t build everything in-house.
They pick strategic partners that help them build better and faster, while focusing in-house engineering efforts on where they’ll generate the most value.
We’re honored to partner with large enterprises in their race to to build the future of the customer experience.
“As an AI-native company, Inventive enables us to innovate at the leading edge of the ecosystem and execute faster on our AI product needs.”
Pavan Bedadala
Senior Director of Product Management
We can’t wait to share more on what we’ve built and where we’re heading next.
If you’re a product or company leader curious to learn more, schedule time with us. Or follow us on LinkedIn and X/Twitter for more in the weeks ahead.
About Inventive
Inventive helps product teams at large companies who struggle with endless custom data requests from customers and lack the capacity to serve them well. Our AI-native custom report builder embeds seamlessly into your SaaS product to let your customers use natural language to explore, save, and operationalize custom insights from their trusted data. Our customers focus on their core, and we extend their data to reach the long tail of their customers' reporting and analytics use cases. Our co-founders bring decades of expertise building data platforms at Google, Microsoft and Meta, and we're well funded by leading Silicon Valley investors.
