CTERA embeds its own engineers at customers until their AI pilots reach production
CTERA, after 18 years selling software simple enough to skip professional services, now embeds engineers at customers until an AI use case runs in production. The engagement ends at go-live, so buyers who want proof of payback have to set that test themselves.
The Board Room · Leadership desk

What happened
- Chief executive Oded Nagel told the 70th IT Press Tour that enterprises are buying GPUs, hardware and storage, whatever they can get.
- Nagel blamed the gap between buying AI and getting value from it on data that is not AI-ready, security that cannot keep up, and too few people able to build agents.
- CTERA's September study of 16 petabytes of production data found 4.4% of capacity in active use and 88% of files untouched for over a year.
- Since last year's tour, CTERA has shipped five releases, including Microsoft 365 Copilot support in July and AI-assisted archiving in September.
Why it matters
- constraint The engineers stay until one specific use case runs, so every use case after that means either a new engagement or an internal build.
- exposure Customers that let CTERA build their first agent may need CTERA again to change or extend it, unless the engagement includes a handover.
- contradiction The skills case for embedding engineers is weaker than CTERA presents it: in the survey CTERA cites, about nine respondents in ten do not name skills as a challenge.
The engagement ends when a specific AI use case is running in production on the customer's own file data, according to iTWire's account of the launch [2]. Nagel set a harder test when he described the customer's problem. "If you use one million tokens, and now you need to pay one million dollars to OpenAI, how can you measure it and prove that it's actually going to improve your business?" he said [6]. A use case can run in production for months with that question still open.
The engagement also protects CTERA's own deployment, since its engineers stay on site until its own software is running. But it covers the step Nagel says customers cannot staff. Of in-house IT teams, he said: "They don't certainly know how to deal with the agentic layer." [9] The company's long-standing prescription puts that work where the data already sits. "Instead of moving the data to the AI, move the AI into your corporate data," Nagel said [11].
The outside evidence for a skills shortage is thinner than the pitch. Futurum research cited on CTERA's slides found 10.4% of respondents named insufficient skills as an organisational challenge, up from 4.8% six months earlier [10]. The share grew about 2.2 times [14], yet roughly 90% of respondents still left skills off their list [15]. The 16-petabyte readiness study is CTERA's own [1]. So the case that enterprise AI stalls between pilot and production rests mainly on the vendor's account and on one customer. Nagel said the IT storage manager at one of the largest banks in the US told him, "we are not there yet. It takes time." [7]
For a company with that history, iTWire wrote, "it's a turn that signals the changing times." [13] In my view the trade-off for a buyer is speed now against dependence later. This quarter, CTERA's engineers can get a first use case running sooner than an internal team that, on Nagel's account, lacks agent skills. Next quarter, that use case sits with the same internal team once the engineers leave, unless the contract covers the handover. Pricing, engagement length and the number of customers signed up were not disclosed in the launch coverage.
What to watch
- Whether any customer reports a payback figure, and not only a go-live date, for a use case CTERA's engineers put into production.
- Futurum's next reading on insufficient skills, after the rise from 4.8% to 10.4% of respondents in six months.
- Whether CTERA widens engagements beyond a single use case or adds formal handover terms for in-house teams.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence35
- Adoption
- Insufficient
- Hype gap+15
- Incentives70
- Confidence40
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
CTERA research published in September covering 16 petabytes of live production data found only 4.4% of capacity was actively used and 88% of files had not been touched in over a year.
- [2]
CTERA has launched Forward Deployed Engineering, a service that puts CTERA engineers inside customer organisations until a specific AI use case is running in production on the customer's own file data.
- [3]
CTERA spent 18 years making its software simple enough to skip professional services.
- [4]
CTERA CEO Oded Nagel and CTO Aron Brand presented on day one of the 70th IT Press Tour.
- [5]
"Everyone is buying AI," Nagel said. "They are buying GPUs, buying hardware, buying storage. Whatever they can, they buy."
- [6]
"If you use one million tokens, and now you need to pay one million dollars to OpenAI, how can you measure it and prove that it's actually going to improve your business?"
- [7]
Nagel said the IT storage manager at one of the largest banks in the US told him: "we are not there yet. It takes time."
- [8]
Nagel put the AI value gap down to three causes: data that isn't AI-ready, security that can't keep up, and too few people who know how to build agents properly.
- [9]
"They don't certainly know how to deal with the agentic layer."
- [10]
Futurum research cited on CTERA's slides found insufficient skills as an organisational challenge more than doubled from 4.8% to 10.4% of respondents in six months.
- [11]
CTERA's prescription is its long-standing philosophy: "Instead of moving the data to the AI, move the AI into your corporate data," Nagel said.
- [12]
Since its last IT Press Tour appearance in New York last year, CTERA has shipped Fusion Direct (March), InsightAI (May), an n8n integration (June), Microsoft 365 Copilot support (July) and AI-assisted archiving (September).
- [13]
iTWire wrote of the launch: "it's a turn that signals the changing times."
- [14]
The share of Futurum respondents naming insufficient skills rose about 2.2 times in six months.
- [15]
Roughly 90% (89.6%) of Futurum respondents did not name insufficient skills as an organisational challenge.
- [16]
CTERA shipped five releases since its last IT Press Tour appearance.
Get this ledger inside ChatGPT. Ask about any of our stories and the answer brings its figure, each claim’s verdict and the sources. Free, no account.
Sources
1 independent publisher whose own reporting we read for this story.
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.
Topics
- Enterprise AI AdoptionFollow
- Unstructured data managementFollow
- Forward-deployed engineeringFollow