How Sintra AI uses Wordware to build AI agents

Who is Sintra?

Sintra provides AI-powered employees (AI agents) to numerous small- to medium- sized businesses. Their digital assistants include data analysts, business operations agents, customer support and customer satisfaction agents, sales managers, marketing and SEO specialists, recruiters, and dozens of other. Their behavior follows step-by-step prompts provided by domain experts that have business experience in certain areas in order to ensure the actions are happening in the way human would do them.

Why they create them and what's the idea behind Sintra AI?

Many small businesses face a dilemma: should we hire full time employees who can help us with growing our core business, eg. through marketing or should we pay premium for part-time freelancers? Sintra's cute AI agents are a cheaper, reliable equivalent for a large selection of these roles, allowing businesses to “hire” digital experts, delegate task to them and reduce manual work done by humans.

Customers of the company are mostly small and medium businesses that want to have access to highly-specialized roles like Google Ads specialists, Customer Experience specialists or Twitter Management specialists from day one, maximizing automation at a tiny fraction of the cost of hiring full-time.

What problem did their team face?

Complex development process

Sintra's AI agents need to achieve top performance across a wide variety of different tasks in order to strengthen their position in the market. However, the orchestration of different LLMs required to perform those tasks was hard, and length of iteration cycles using code-based LLM orchestration tools, for example LangChain or LlamaIndex, was considerable.

Lack of Streamlined Collaboration

Moreover, because of the nature of Sintra's business, behavior of AI agents needs to be consulted with domain experts that advise on the personas and capabilities of these digital assistants. These experts can give ideas on how to automate certain workflows with step-by-step guide what to follow, but the aspect of actually building a working AI agent is usually out of the scope of their skills. That's why this process would require an engineer, acting as a catalyst between domain expert and AI agents, translating expert feedback and prompts into the code required to prompt an LLM and hopefully achieve the desired output from an agent.

How to Solve That?

With dozens of personal, digital assistants on the roadmap, each requiring individual optimization, Sintra's team searched for advice on how to improve the productivity of the team, cut the waiting time between domain expert response and implementing prompt into a codebase and reducing the engineering resources required to meet their ambitious timelines.

Working with Wordware: How Sintra AI uses Wordware to build AI Agents

During our initial conversation, Sintra's team outlined the ideas for the agents, their company goals, and the future they envision for the product. With strong demand and sales, they needed to show their business as the best choice for any company that wants to hire AI agents.

New Way of Agents Development

To answer those questions, Wordware and Sintra developed strategies that were to drive the success and fulfilling their business needs by cutting cost of development and improving the output of AI agents. Our solutions focused on: creating a new system of collaboration for building AI agents and providing a platform that will enable to boost productivity of a whole development team.

With our tool Sintra redefined how they collaborate with domain experts on agent building: no need to write and send prompts in Slack messages back and forth, no hours spent to develop one of the AI employees, much more control over the outputs and fast iterations by domain experts to optimize the performance of agents.

Right now domain experts create prompts that are fed to AI directly on Wordware platform. Experts then can instantly check AI answers and compare that with the ground truth or quality bar they have. The point is to have employees with domain expertise to focus on things they do best, without the need of involving the rest of the team into the iteration process.

Later on, once experts create AI assistant, developers can plug it into their product using Wordware's API.

Results of Building on Wordware

All in all, Sintra built capabilities for their agents on Wordware, leveraging the platform to quickly translate expert input into functional AI agents without the need for constant engineering intervention. This allowed them to focus on refining agent personas and capabilities directly, significantly accelerating their development timeline.

Utilizing Wordware's platform, Sintra manages the complexity of dozens of AI agents simultaneously, each tailored to specific business needs. Future plans include broadening the coverage of their AI employees. The team wants tools for each business function, utilizing prompts in a way that will enable social media managers, sales representatives, and any employees to augment their work. Wordware's multimodality reduced the time and resources typically required for such sophisticated orchestration, enabling Sintra to deliver industry-leading performance across their suite of AI employees, hence moving closer towards being a one of the leading organizations in AI employees space.

Sintra AI and other digital assistants solutions

Many businesses face the challenge of needing specialized expertise without the capacity to employ full-time professionals in those roles. Sintra AI agents propose a solution to this problem, providing digital assistants that can perform tasks such as social media management, marketing, sales, video generation, data analysis and abilities to automate other personal tasks, boosting your and your business productivity.

The problem however lays in the prompts and it compounds when businesses scale, market demand changes and the quality bar for AI assistant rises, as AI agents providers they often lack the in-house engineering talent to fully leverage these tools.

What the future holds?

Until now there was no easy way to address cross-functional team collaboration. Wordware's mission is to be the most beneficial tool that companies can use to build their AI agents. Wordware addresses this by providing an end-to-end solution that enables businesses to deploy AI in both internal tools and customer-facing products quickly and effectively, even without deep technical knowledge.

Sintra addresses this by providing agents capable of nearly every task, providing agentic personalities to meet countless critical functions at small businesses and early startups, automating their tasks, boosting teams productivity and the company performance.

If you're interested in learning more about Sintra, you can see their combined offering including all available agents on their website here. If you'd like to see how to build like Sintra, explore prompts with Wordware here.