Palo Alto's Sycamore opens its tool for office workers to build AI agents without coding

Sycamore Labs, a Palo Alto startup, on Tuesday opened Sycamore Build to all businesses and individuals, the company said in a news release.  The service lets employees who don't code create their own work software and AI agents.

What an AI agent is

A chatbot like ChatGPT answers questions.  An AI agent goes further:  it is given a goal, plans the steps and uses tools, such as searching, reading company records or sending data to other software, to finish the job on its own, IBM explains in a guide to the technology.  People set the goal and the rules, and IBM recommends keeping a log of everything an agent does and requiring a human sign-off before high-stakes actions.  In one example IBM gives, a legal-research agent cut contract review from 90 minutes to 45.

How Sycamore Build works

An employee describes a job they do in plain language, shares documents or process maps, and grants access to the company systems involved, such as customer records, files, ticketing and messaging tools, according to Sycamore's release and product page.  Sycamore's AI asks about missing details and writes a plan for the employee to approve.  Then a team of specialized AI agents researches, writes, connects and tests the software, and the employee can watch a live preview, try it and send feedback.

The result can be an app, an agent or a team of agents, a report, a dashboard, a website or a presentation, Sycamore says.  Agents built this way ask a person when a decision needs human judgment, every action is logged, and agents earn more freedom only gradually, according to the company's website.  When a process changes, workers can describe the change and keep building instead of starting over, the company said.

What it could mean for office workers

Founder and CEO Sri Viswanath, writing on Sycamore's blog, described the target as the everyday snags office workers know:  a customer request that gets stuck, an invoice that is hard to match with records, onboarding steps that waste time.  Sample requests on Sycamore's homepage include an agent to evaluate vendor risk, a board meeting slide deck, a retail demand forecast and a round of feedback for new hires.

Sycamore pitches this as an alternative to off-the-shelf software that forces businesses to bend their processes, and to custom projects that take months or years.  Companies can set shared rules for logins, permissions, testing and review that apply to every project, it said.

The people who understand a business best should be able to build the software that runs it.

Sycamore's other product, Sycamore Govern, oversees agents once they are running.  The company raised a $65 million seed round in March led by Coatue and Lightspeed, it says.  It did not give prices; people can sign up at sycamore.so.

Source:  Sycamore Labs news release, Nov. 3, 2026; Sycamore website, product page and blog posts of Sept. 9 and Nov. 3, 2026; IBM, "What are AI agents?"