Telefónica - Success Story
Telefónica partners with Sagittal AI to transform their software development processes with Neo
About Telefónica
Telefónica operates one of the world's largest telecommunications networks, serving 383 million customers across critical infrastructure in Europe and Latin America. Through Telefónica Innovación Digital TID, Telefónica drives the company's exploration of emerging technologies and next-generation development practices. In early 2024, the division began exploring AI-powered approaches to software development that could accelerate delivery, without compromising the security and reliability standards required from critical infrastructure providers.
The Challenge
Telefónica was proactively seeking AI solutions that could fundamentally change how their teams worked. The existing AI development tools in the market primarily focused on individual developer productivity, such as autocomplete and code suggestions that make typing faster. While useful, such tools don't address the broader challenge of accelerating team-level delivery across the entire software development lifecycle.
In the QA team we can't work in isolation. AI tools that run on our laptops are useful to a certain degree, but with Neo we collaborate with developers and product managers in Jira or GitHub, where hand-offs are immediate and transparent. It's very satisfying to see my team report a problem and have a developer say, - Neo, fix it!
The division required that any solution works within their existing tools rather than requiring engineers to adopt to entirely new workflows. Telefónica has a multi-vendor stack with services like Jira, Confluence, GitHub, Jenkins or Figma and a very large codebase across hundreds of repositories, making it challenging to adapt any one AI tool to their development environments.
Security
As a telecommunications provider managing critical infrastructure, maintaining security and control is non-negotiable. Neo achieves this control and security because:
- The IT and security teams set up Neo to their requirements, they understand scopes and permissions and know exactly how they want tools to be integrated, authenticated and monitored.
- Developers don't have to install anything on their laptops, and most importantly, they don't need to allow any tool to act on their behalf, a common problem with current AI tools having autonomy with as much power as the users that drive them.
- Neo is accountable, as it acts with its own account. Every line of code written, pull request reviewed, ticket modified or document edited; all users can transparently see what Neo has done within the tools themselves. Traceability is not lost or buried in logs.
Neo avoids the security issues other AI solutions fall into. We control its access centrally, no one has to install anything and it doesn't act on behalf of our users with their credentials, all of Neo's actions are traceable.
The Solution
Telefónica engaged with Sagittal AI and its Neo platform in early 2024. Ivan Izaguirre, Chief Technical Officer at Telefónica Innovation, led the evaluation.
"Neo offers a very different approach that complements and often supersedes other AI tools. We already use AI tools to help developers write code, but there is so much more than code in the lifetime of a feature. Neo helps our product and engineering teams from specification to deployment, and I can actually measure how much more my teams deliver. It was eye opening to see the before and after".
The pilot started small with Neo available to just a few engineers on a single team. Word spread quickly through the division as engineers experienced the difference firsthand. "We use Neo on all types of projects: backend, frontend, embedded. It finds the right context, and it's easy to customise using human-readable onboarding docs," Izaguirre said. "What began as a limited pilot grew organically across the entire Innovation Division as teams saw the results and new teams requested access."
Results
The partnership has delivered both technical and organisational outcomes that have exceeded Telefónica's initial expectations.
Teams using Neo have seen measurable improvements in both velocity and quality. More than 300 engineers have over 2000 pull requests reviewed every month, and about 100 tasks get completed directly from Jira to Github.
The security and control model has proven equally valuable. "Having AI in the SDLC with version control, tracking, etc. - means that we can be more aggressive with AI without the downsides. Other AI solutions don't allow this." Izaguirre explained.
Organisational outcomes
"We consider Neo just another team member", said Gonzalo Fernandez, Head of Platform Engineering. "When we open an issue, we don't even know yet that it will be the one to handle it. Neo can take it from JIRA, implement the functionality, and push it to the repository for review, or review our changes if needed".
Beyond the technical metrics, Telefónica has observed a cultural shift in how teams approach development. By using Neo, engineers learn to delegate to AI, so it's as if every team member has someone to ask to do work for them, and check on their output later.
Like other agentic tools, Neo requires iteration to do things the way I prefer, but what I get from Neo that others cannot provide is the ability to assign it two or three tasks in the morning and continue with my day. I can return later and review its code the same way I review my colleagues'. When Neo gets something wrong, I have not wasted my time.
Looking Forward
"Sagittal has been an excellent design partner", Izaguirre said. "They have not only sped us up as we hoped, they've pushed us further into AI and really showing AI's potential to speed up teams. And they've quickly made customisations to our processes. Upon reflection, this isn't too surprising, since Sagittal uses Neo to implement such changes!"
For Telefónica, this partnership represents just the beginning of their AI-powered development journey. "We see so much potential here, and it feels like we're only starting this journey with Sagittal", Izaguirre said. "By being a design partner we have access to the alpha features as they're implemented, and we can't wait to see the impact they'll have in the teams."