InstaLILY Inc. said today it has raised $25 million exclusively from Insight Partners to fuel the growth of industry vertical-specific “AI Teammates” that are designed to live inside legacy software systems and perform tasks that were previously impossible to automate.

The artificial intelligence startup says it’s pioneering a new way to fulfill the potential of AI in a range of industries where automation has been slow to catch on. Rather than building automation flows or stitching together various tools, companies can employ InstanLILY’s InstaWorkers, which are AI agents that integrate with enterprise resource planning, customer relationship management and other platforms.

According to InstaLILY, industries such as physical goods manufacturing, insurance and healthcare services have struggled to take advantage of AI because they depend on specialized knowledge, large catalogs and fragmented tools to get their work done. These legacy tools result in lots of high-volume, multistep work that can only be performed laboriously by human experts – but InstaWorkers changes this.

InstaLILY founder and Chief Executive Amit Shah told SiliconANGLE the InstaWorkers have been trained on domain-specific processes unique to each industry, allowing them to navigate complex software environments to execute complex workflows, without any need to replace those legacy tools.

The startup has developed a sophisticated vertical stack to power its InstaWorkers, made up of a data layer, composable layer, learning layer and customer layer. ”Our four-layer architecture is what lets InstaWorkers act like real teammates, not just chatbots in complex, enterprise systems,” Shah explained.

Shah said the data layer continuously pulls information from systems such as CRMs, ERPs and email so that the InstaWorkers always have the richest, cleanest and most up-to-date data and context to work with. The composable layer is like “LEGO-style building blocks of workflows and integrations that we snap together to create new AI teammates.”

The learning layer helps to teach these AI teammates, using human feedback and reinforcement signals so that the InstaWorkers learn which actions to implement, improving outcomes the more they’re used. Finally, the customer layer is the “front line,” where work gets done directly inside customer’s systems, based on the rules they establish, Shah said.

Together, the four layers help InstaLILY’s AI agents to read and understand context, make decisions, learn over time and take actions. This unique architecture makes it possible for companies to train InstaWorkers on their unique, domain-specific business processes, documentation and systems, and armed with this specialized knowledge, they can then navigate through the customer’s fragmented software environment to automate business processes and tasks with full human-in-the-loop oversight.

“The data, composable, and learning layers are connected in a way that every action is informed by full business context and every outcome feeds straight back into learning,” Shah explained. “That tight loop means the agent doesn’t just get smarter — it gets more effective in the work it’s allowed to do.”

Shah believes InstaWorkers will surpass existing AI copilots, which he says are useful in some situations but generally struggle to automate most industries. “I’m talking about the toolbars, pop-ups and side panels that give advice but leave all of the clicking and updating to you,” he clarified. “They are useful because they can surface insights, draft context and suggest next steps, but they cannot close the loop.”

By this, Shah explained that copilots are unable to take the final steps required to act truly autonomously. For instance, they can’t actually send out an email or update a system of record. The human has to take these actions. It means that such work cannot be fully automated, and remains in a queue until a human finds the time to finish it.

“The difference is that copilots suggest, and we execute,” Shah said. “InstaWorkers log in like a real teammate, and they finish the job inside your systems.”

InstaLILY didn’t name any customers, but it claims to have made a significant impact already in a number of industries. For instance, one of its clients is a $10 billion-plus construction supply distributor, which is using InstaWorkers to enhance its operations by transforming sales data into actionable follow-ups and increasing its overall sales.

The startup also claims to serve one of the world’s largest original equipment manufacturers with its AI service specialists, which support field technicians, helping them to analyze complex fault descriptions and predict which parts need replacing. Another of its customers is a large insurance firm, which uses InstaWorkers to assist with its claims operations, extracting policy and claim data and comparing it with its coverage rules, reducing manual review times by 70%.

Insight Partners Principal Crissy Behrens said InstaLILY’s domain-trained AI agents are both intuitive and built to scale and execute workflows where traditional horizontal AI tools fail. “With InstaWorkers, InstaLILY is helping to build a better future of work, grounded in actions not suggestions,” she said.

Moving forward, InstaLILY plans to expand its catalog of pre-trained InstaWorkers to cover more industry verticals and increase its support for common enterprise platforms. In addition, it also plans to expand the multimodal capabilities of its AI teammates, so they can process voice and video data to unlock new use cases in contact centers, field services and human-robot collaboration.

Image: InstaLILY