Clueso, a San Francisco company that sells AI agents for producing business video, today said it had raised a $10 million Series A led by Dharana Capital. The company says more than 1,000 companies use its platform, and that its software can operate real products, capture the footage, write and voice the script, edit the result, and rerun the whole process when the product changes.

In Short

A startup called Clueso got $10 million from investors to build software that makes how-to and training videos for companies by actually clicking through their apps, recording what happens, and editing it into a finished video. This matters to marketing, training and customer support teams, because product videos usually go out of date every time the software changes and someone has to remake them by hand. Clueso says its agents can redo those videos automatically, but the savings it quotes come from the company itself and have not been checked by anyone else.

The round

The financing was announced on October 13, 2026, from San Francisco. According to Clueso, Dharana Capital led the $10 million Series A, with Founders Future and Neotribe Ventures joining as new investors and Y Combinator and f7 Ventures continuing to participate. The round brings total funding to $12 million.

One detail in the release says more about the deal than the headline amount. According to Clueso, more than 80% of the new capital came from existing investors "doubling down." On $10 million, that is more than $8 million from backers who were already on the cap table, which leaves less than $2 million for the two firms the company describes as new. The release does not say whether Dharana Capital, the lead, was already an investor. It names only Founders Future and Neotribe Ventures as newcomers, which implies that Dharana was, though the company does not state it. Arithmetic on the total also implies that Clueso had raised about $2 million before this round.

Insider-led rounds can be read two ways. Existing investors who see a company's numbers closely choosing to put in more money is a signal of conviction. A round in which outside capital is a small minority is also one in which fewer new parties set the price. Clueso did not disclose a valuation.

The amount is small against the capital flowing into AI companies that sell to marketers and sales teams. Crunchbase counted about $3.7 billion raised globally by sales, marketing and CRM startups in 2026 by early May, with five rounds, led by Sierra's $950 million, making up roughly 43% of the total. A $10 million Series A sits well outside that concentration.

Figures the company supplied

Every performance number in the announcement comes from Clueso, and none is independently verified. The company says more than 1,000 companies use the platform. It names Atlassian, Zendesk, Gainsight, GlobalPartners, Seismic and Ironclad as customers. In material sent to journalists alongside the release, Clueso added that customer numbers grew threefold over the past year and that revenue is approaching eight figures, meaning somewhere short of $10 million a year. It gave no exact revenue, no growth rate in revenue terms, and no definition of what counts as a customer.

The productivity claims need closer reading. The formal release says: "Teams have reported over 5x productivity gains and an 80% reduction in video production costs." It does not say which teams, how many, over what period, or what was measured. The accompanying pitch material attributes those same two figures to "customers like Atlassian and Zendesk." The quotes from Atlassian and Zendesk staff in the release itself do not mention either number. The Zendesk quote says the platform "has helped us reduce production time and expand our video output," with no percentage attached. Without a baseline, an 80% cost reduction could mean replacing an agency, replacing an in-house editor's hours, or replacing nothing that was previously budgeted at all.

What the platform does

Most of the attention on AI video in advertising has gone to generative models that turn a prompt into synthetic footage. Clueso describes itself as doing something else. Rather than generating imagery from text, the company says, it is given access to a customer's products, workflows and brand guidelines, and its agents operate the real software to record what a viewer needs to see.

According to Clueso, the platform covers five areas:

  • Capture. "Computer agents" navigate real software, perform workflows and record the footage required for a video.
  • Production and editing. The platform writes scripts and produces voiceovers, motion graphics, captions and edits automatically, with a browser-based editor for manual control.
  • Maintenance. When a product or process changes, Clueso says it can rerun the workflow and refresh the video.
  • Brand rules. Teams define reusable standards for voice, visual style, localization, motion graphics and music, which both agents and people then follow.
  • Integrations. Connections bring in context from other company systems, and a Clueso MCP server lets AI assistants such as Claude and ChatGPT create, edit and manage videos directly through the platform.

The maintenance claim is the technically distinctive one. A screen-recorded tutorial is a snapshot of an interface on a given day; once a button moves or a menu is renamed, the video is wrong. Re-recording depends on someone noticing the change, re-performing the steps and re-editing. Clueso's pitch is that if the agent performed the workflow once, it can perform it again against the new interface and regenerate the output. The release does not say how the system detects that a product has changed, whether refreshes are triggered automatically or on request, or how it handles a workflow that no longer exists in the same form.

Agents that operate software by looking at a screen are also a known weak point. When Google's Chrome team opened its WebMCP preview in February, the aim was to give agents structured tools instead of screenshot analysis, DOM parsing and simulated clicks, approaches that consume substantial computing resources and are often unreliable. Clueso does not explain how its computer agents navigate interfaces, how often a capture fails, or what human review sits between a recorded run and a published video.

Product details on the company's site

Clueso's website, checked on October 13, 2026, adds detail the release omits. It lists a free tier with 10 minutes of video exports per month, a Solo plan from $40 a month and a Business plan from $120 a month, both billed annually, and custom Enterprise pricing. The site says the company is SOC 2 Type II and ISO 27001 certified, supports SAML single sign-on and SCIM provisioning, and does not train models on customer data. It also lists AI avatars, a library of more than 1 million stock images and videos, reusable templates the company calls "Clueprints," and a feature that converts videos into written step-by-step guides. The MCP integration is listed as working with Claude, ChatGPT and Cursor.

The site is not consistent on language support. Its feature list and FAQ say translation covers more than 150 languages, while a statistics section on the same page says more than 80.

Who is quoted

Akash Anand, co-founder and chief executive, described how the product widened from its original use. "We originally started by helping customer-facing teams create product tutorials faster," Anand said, according to the release. "We quickly saw the same problem across marketing, sales, training and internal documentation. That made us realize the opportunity was much bigger. Today, Clueso acts as an AI video production team that anyone or any team in the organization can rely on, handling the work of creating, editing and maintaining videos as the business moves."

The company was founded by Anand, Prajwal Prakash and Neel Balar, who, according to Clueso, encountered the problem at a previous company. The release does not name that company or give a founding year.

Garrett Marttinen, head of customer learning at Atlassian, is quoted on the customer side. "Clueso has fundamentally simplified how we produce learning videos at scale," he said. "They help us turn rough, scrappy input into polished, on-brand output, streamlining our workflows and dramatically reducing the burden of creation and maintenance for my team."

Kevin Gross, senior manager of instructional design at Zendesk, said: "Clueso's agent makes it easy to automatically transform product knowledge into polished, engaging learning experiences with far less manual effort." According to Clueso, Zendesk uses the platform for both internal and customer-facing content.

Both customer quotes come from learning and instructional design functions, not from marketing or advertising teams. That is consistent with Anand's account of where the product began.

Vamsi Duvvuri, founder and managing partner at Dharana Capital, framed the investment around volume. "Video is one of the most effective formats enterprises use to train teams, educate customers, and sell products. However, traditional video production was never designed for enterprise volume and velocity," he said. "With AI, video is evolving from static media files into living, context-aware assets that continuously adapt as business processes and products change."

Where the money goes

According to Clueso, the funding will be used to scale enterprise sales, deepen the agentic AI layer behind the platform, expand editing capabilities, and connect video creation more closely to the systems where product and company knowledge already sits. The company describes a typical adoption pattern in which a single team starts using the platform and others follow, until Clueso becomes "the centralized video creation platform across the organization."

No hiring numbers, geographic expansion or product release dates were given.

Why this matters for marketers

The marketing relevance is less about ad creative than about the volume of explanatory video that sits around a product: demos, onboarding sequences, feature announcements, help-centre clips and sales enablement material. These assets are expensive mainly because they decay. Every interface change creates a backlog of outdated footage, and for software companies shipping weekly that backlog never clears.

The ad platforms have been building in a different direction. Google added Gemini Omni video creation to Asset Studio at Google Marketing Live in May. TikTok raised the length of AI-generated ads in Symphony from 15 to 30 seconds with its Seedance 2.5 upgrade in August. Meta told Cannes Lions in June that more than 8 million advertisers use its generative AI creative tools, and extended video voiceover translation to at least 11 languages. Amazon took its Creative Agent into Streaming TV and Sponsored TV formats in November 2025. All of these produce new media for paid placements inside the platform that sells the placement.

Independent agent vendors aimed at agencies have taken a broader route. Luma said in March that its agents would handle full creative projects across text, image, video and audio, routing tasks to models including Veo 3, Sora 2 and Kling, with Publicis Groupe and Serviceplan Group deploying them at launch. Clueso's approach differs from both camps in its raw material. Its footage is a recording of software that exists, which removes one category of risk in generated video - showing a product doing something it cannot do - while adding another: whatever the agent recorded is only as current as its last run.

The MCP server is the other point of interest. Marketing software vendors have spent the past year exposing their products to general-purpose AI assistants. Adform said in September that assistants could reach more than 800 functions in its FLOW platform through an MCP server. Meta's Muse agent connected to ad accounts and 15 third-party business tools in the US and Canada. And Ads Uploader reported that 17.5% of 36,289 Meta ad batches in September 2026 came from AI agents such as Claude or ChatGPT, using its MCP server or command-line interface. If video production can be called the same way, an assistant asked to prepare a campaign could in principle request the matching product walkthrough as one more step. Clueso has not said how many of its customers use the MCP server, or whether any use it in production.

What remains open is evidence. A company that says 1,000 organisations use its product, that customer numbers tripled in a year and that costs fell 80% has supplied claims, not measurements. The release offers no method, no sample and no period for any of them, and the two customer quotes it does include stop short of the numbers. For a category whose central promise is that video will stay accurate without human effort, the absence of any failure rate or review process is the largest gap.

Timeline

Summary

Who: Clueso, a San Francisco company selling an agent-based video production platform, founded by Akash Anand, Prajwal Prakash and Neel Balar. Investors are Dharana Capital (lead), Founders Future, Neotribe Ventures, Y Combinator and f7 Ventures. Named customers include Atlassian, Zendesk, Gainsight, GlobalPartners, Seismic and Ironclad.

What: A $10 million Series A, bringing total funding to $12 million, with more than 80% of the new capital from existing investors, according to the company. Clueso's agents operate real software to capture footage, then script, voice, edit and refresh videos when products change, and an MCP server lets assistants such as Claude and ChatGPT drive the platform. The company says more than 1,000 companies use it and reports 5x productivity gains and 80% cost reductions, figures it did not source or explain.

When: The round was announced on October 13, 2026.

Where: San Francisco, with the platform sold to businesses using it for marketing, customer education, onboarding, training, sales enablement and internal operations.

Why: Business video goes out of date whenever software changes, and remaking it is manual and expensive. Clueso says recording real workflows with agents, and rerunning them after changes, keeps that video current. The funding is meant to expand enterprise sales and the agent layer, while the central claims on cost and accuracy remain unverified.