Datadog is expanding its Digital Experience Monitoring platform with two new tools designed to connect what users experience with what engineering teams see behind the scenes. The additions, called Journey Monitoring and Bits Testing, combine production telemetry with autonomous synthetic testing to help organisations identify broken customer journeys and understand why important conversion paths are failing.
Both capabilities are currently entering technical preview for Datadog customers already using tools such as Real User Monitoring, synthetic testing and product analytics. The broader objective is to close a familiar gap inside many organisations: business teams may notice that conversion rates have suddenly fallen, while engineering teams are separately looking at service availability, browser errors and infrastructure performance. Datadog wants those groups to work from the same view of the customer journey.
Connecting User Behaviour With Technical Performance
A typical digital service generates several different versions of the truth. Marketing teams may see that fewer customers are completing checkout, product teams may notice increased abandonment on a particular screen, and site reliability engineers may be watching latency and availability dashboards. Each team has useful information, but connecting those signals often requires considerable manual investigation.
Journey Monitoring attempts to bring those perspectives together. It maps actual user activity alongside technical service-level objectives, showing where journeys begin, how long they take, where people abandon them and whether availability or synthetic uptime is deteriorating at the same time.
This means an organisation can look at something like a purchase journey as a complete path rather than a collection of unrelated pages and services. Instead of simply knowing that checkout conversion has fallen, teams can begin examining whether the problem started during login, payment validation, account creation or another dependency earlier in the journey.
Journey Monitoring Maps the Blast Radius of Failures
One of the more useful aspects of Journey Monitoring is its ability to show how one failure can affect several connected workflows. A problem with authentication, for example, may not simply break the login page. It may also prevent customers from checking out, managing their accounts or accessing support features.
Datadog visualises these upstream and downstream relationships through an experience map, helping operators understand the wider impact of a fault. Service-level objective breaches can be highlighted directly on the journey, making it easier to identify which part of the user experience is falling outside acceptable thresholds.
This kind of visibility becomes particularly important in complex applications where many seemingly independent features depend on the same underlying service. A technical failure that looks small from an infrastructure perspective may actually affect several high-value customer journeys at once.
Teams Can Move From Funnels Into Individual Sessions
Journey Monitoring is also designed to let engineers move from high-level experience data into more detailed investigation without leaving the context of the journey. If an availability target is breached on a purchase form, for example, operators can move directly into Session Replay or Datadog's RUM Sessions Explorer to examine what happened on the user's device.
That connection is important because conversion data can tell teams that something went wrong without necessarily explaining why. A user may have abandoned checkout because of a JavaScript error, a broken form field, an unavailable API or simply because the interface became confusing.
Being able to move from the funnel view into individual sessions helps connect those business and technical perspectives. Instead of one team reporting that conversion is down and another team searching through logs for possible causes, both can investigate the same affected journey.
Datadog Can Discover Emerging Journeys From Production Traffic
Journey Monitoring is not limited to workflows that teams have manually defined in advance. Datadog says the platform continuously analyses production traffic to identify new user paths that begin appearing over time.
That could be useful when products evolve quickly. Users do not always follow the exact paths product teams originally expected, especially after new features or navigation changes are introduced. Real-world behaviour can create important journeys that nobody formally documented.
By recognising those patterns automatically, teams can register new transactional flows as managed journeys before they become significant sources of customer complaints. It effectively allows the monitoring system to learn which parts of the product are becoming important based on actual usage.
Bits Testing Takes a Different Approach to Synthetic Monitoring
The second new capability, Bits Testing, focuses on one of the long-standing weaknesses of synthetic testing: maintenance.
Traditional synthetic tests usually depend on predefined sequences. Click this button, enter this value, select that menu item and wait for a particular response. These tests are excellent when the interface remains predictable, but they can become fragile when developers change the layout or introduce dynamic content.
Move a button, rename an element or change how recommendations are rendered and a synthetic test may fail even though the application itself still works perfectly. QA teams then spend time repairing the test rather than investigating an actual product problem.
Bits Testing attempts to reduce that maintenance by giving an autonomous agent more flexibility when navigating the application.
Developers Can Describe What They Want Tested
Instead of requiring every interaction to be manually scripted, quality teams can provide Bits Testing with a target website or describe the intended workflow using natural language. The agent then explores the application, identifies possible routes through it and generates runnable tests.
Datadog says those tests can cover browser interactions, network requests and APIs, allowing the agent to examine more than just what appears visually on the screen.
This approach is particularly useful during early test creation. Rather than beginning with an empty automation project and manually documenting every step, teams can allow the agent to discover the workflow and then build testing coverage around what it finds.
The human engineer still defines the outcome that matters, but the AI handles more of the mechanical exploration needed to reach it.
Deterministic and Goal-Based Tests Can Work Together
Datadog is not replacing traditional scripted testing entirely. Bits Testing supports both deterministic verification and goal-based autonomous testing, because the two approaches are useful for different types of validation.
Some workflows need exact behaviour. Payment-processing endpoints, authentication contracts and other critical functions may need to follow strict sequences every time. Those are good candidates for deterministic tests where every expected step is explicitly defined.
Other workflows are more dynamic. A travel site may display different hotels or flights depending on availability, while a recommendation engine may change the position of elements between sessions. A rigid test can become difficult to maintain in those environments.
Goal-based testing allows the test to focus on the intended outcome instead of the exact route taken to achieve it.
The Agent Can Adapt When the Interface Changes
Instead of storing every click as an unchangeable instruction, a goal-based test can contain an objective such as complete a round-trip flight reservation or retrieve an answer from the support assistant.
During execution, the agent decides how to accomplish that objective using the interface currently available. If a submission button has moved, the page structure has changed or different recommendations are displayed, the test can potentially adapt and continue rather than failing immediately.
That makes synthetic testing behave somewhat closer to a human user. People generally do not stop functioning because a button moved twenty pixels to the right. They look at the page, understand what changed and find the new path.
Autonomous testing attempts to bring some of that flexibility into automated validation.
This Could Reduce Test Maintenance Considerably
Test maintenance is one of the hidden costs of automation. Building a large suite of synthetic checks can initially save enormous amounts of manual QA work, but the tests themselves eventually become another software system that needs maintenance.
Every redesign can create broken selectors. New workflows require new scripts. Dynamic interfaces introduce more edge cases. Over time, engineering teams can spend a surprising amount of effort keeping their tests aligned with the application.
If autonomous testing can reliably adapt to harmless interface changes while still detecting genuine failures, that could significantly improve the economics of synthetic testing. QA engineers could spend more time deciding what needs to be validated and less time fixing automation that broke for purely cosmetic reasons.
The challenge, naturally, is making sure the agent does not become so flexible that it accidentally navigates around a real defect and reports success.
AI Sits Underneath the Monitoring Workflow
Datadog Chief Product Officer Yanbing Li describes the company's approach to AI as something that should operate underneath the platform rather than appearing as an isolated feature. Journey Monitoring uses AI to discover and observe important customer paths, while Bits Testing uses agentic capabilities to continually validate those paths.
That creates an interesting feedback loop. Journey Monitoring can identify which workflows matter most based on production behaviour, while Bits Testing can then continuously exercise those workflows and check whether they remain functional.
In theory, the platform becomes increasingly aware of both what customers are actually doing and whether those experiences still work correctly.
That is a more useful application of AI than simply adding a chatbot to another dashboard. The intelligence becomes part of how monitoring and testing operate rather than another interface layered on top.
Marketing and Engineering Get a Shared View
One of the less technical but potentially valuable outcomes is better collaboration between business and engineering teams.
When a conversion funnel suddenly deteriorates, marketing often sees the business impact first. Engineering may not see an obvious outage because every individual service still appears technically available. The disagreement becomes one of perspective rather than data quality.
Journey Monitoring gives both groups a shared representation. Marketing can see which customer path is failing, while engineers can immediately see the corresponding service health, availability and client-side errors.
That could shorten the time between detecting a business problem and identifying the technical reason behind it.
Production Monitoring and Testing Begin to Converge
Traditionally, synthetic testing and production monitoring have been treated as related but separate disciplines. Synthetic tests ask whether a predefined workflow still works, while real-user monitoring shows what actual customers are experiencing.
Datadog's new approach brings those two closer together.
Production traffic can reveal which journeys have become important. Those journeys can then become managed assets, and autonomous tests can continuously validate them. When synthetic checks begin failing, teams can compare the results with real-user behaviour to understand whether customers are experiencing the same problem.
This creates a much tighter connection between testing before something goes wrong and observing what happens when real users encounter the application.
AI-Generated Software Makes This More Relevant
There is also a broader industry context. Software itself is changing more quickly as development teams adopt AI coding assistants and increasingly autonomous engineering agents. If code can be created and modified faster, testing and monitoring need to keep pace.
Rigid synthetic scripts become increasingly difficult to maintain when interfaces and application logic are changing constantly. Autonomous tests that adapt around those changes could become more valuable as release frequency increases.
The same applies to journey discovery. Product teams may introduce features faster than monitoring teams can manually document every resulting user flow.
Using production telemetry to automatically discover important pathways gives observability systems a chance to keep up with applications that are evolving much more quickly than before.
The Human Still Defines What Success Means
Autonomous testing does not completely remove humans from QA. Someone still needs to decide which outcomes matter, what constitutes acceptable behaviour and which workflows require strict deterministic validation.
A payment transaction, for example, may need very precise checks around amounts, authentication and API responses. Allowing an AI agent too much freedom in that workflow could hide subtle problems.
A browsing or discovery workflow, meanwhile, may tolerate many possible routes as long as the user eventually reaches the correct result.
The most practical approach therefore appears to be a combination: deterministic tests for critical contracts and goal-based agents for workflows where flexibility is useful.
Both Features Are Currently in Technical Preview
Journey Monitoring and Bits Testing are currently being offered in technical preview to existing Datadog customers using the company's broader Digital Experience Monitoring tools.
That preview status matters because autonomous testing is still a relatively new field. Datadog will need to demonstrate that its agents can distinguish harmless interface changes from genuine problems consistently enough for teams to trust them.
Journey Monitoring will similarly need to prove that automatically identified workflows actually reflect meaningful customer behaviour rather than simply producing another layer of telemetry.
Still, the direction is interesting because it shifts monitoring from observing individual services toward understanding complete experiences.
Final Thoughts
Datadog's new Journey Monitoring and Bits Testing capabilities show how observability and QA are starting to move beyond individual pages, services and scripted tests toward something much closer to the complete customer journey.
Journey Monitoring connects conversion funnels with technical performance, allowing teams to understand how infrastructure problems actually affect users. Bits Testing approaches the problem from the opposite direction, using autonomous agents to continuously validate the workflows that matter even as interfaces change.
Together, the two features create a useful cycle: discover the journeys customers actually use, monitor how well they perform and continuously test whether they still work.
The idea becomes particularly relevant as software development accelerates through AI-assisted coding. Faster development creates more changes, and more changes create more opportunities for traditional test scripts and monitoring configurations to fall behind.
Datadog's answer is to make the monitoring and testing layers more adaptive as well.
If that works reliably, teams may spend less time rebuilding synthetic tests and manually connecting conversion drops to infrastructure incidents—and more time actually fixing the problems affecting their users.
Comments 0