Revenue operations has spent three years building the signal layer, and the work paid off. In customer research published by LeanData on 31 August 2026, 79% of RevOps, marketing operations and sales professionals reported that they are already scaling or deploying AI agents. Enterprise stacks have become more deliberate at the same time: the average large B2B organisation now runs 37 martech tools, down from 62 in earlier waves of the same benchmark. The intelligence is in place, the budgets are holding, and 79% of leaders expect to spend more next year rather than less.
Which makes the next move an unusually attractive one, because it is also the shortest on the list. The handoff, meaning the routing, assignment and ownership rules that decide who acts on a signal and how quickly, is the least crowded piece of work in revenue operations. In LeanData’s 2026 B2B State of Martech and Revenue Operations report, drawn from 201 senior leaders at organisations of 2,500 employees or more across seven countries, only 11% had applied AI to lead routing and assignment. That was the lowest of ten use cases surveyed. Content and campaign creation sat at 46%, data enrichment and hygiene at 42%, analytics at 39%. The handoff is where the signals a team already generates turn into conversations, and almost nobody has got there yet.
The hard part of RevOps is already behind you
The same report scores operational maturity across five pillars, and the pattern in those numbers is genuinely encouraging. People and Teams leads at 3.82 out of 5, up 0.43 over three years, the largest gain of any pillar. Platform and Technology follows at 3.81, and Pioneer and Pilot at 3.79 after accelerating sharply. Those are the pillars that needed hiring, procurement and executive sponsorship, and revenue teams have won those arguments. The specialist roles exist. The platforms are bought and consolidated.
Process and Operations sits at 3.66, having moved 0.13 in the same period. Read that alongside the rest and it is not a warning, it is a map. The remaining gap is the one pillar that does not require a purchase order or a new headcount. It requires a set of decisions about who owns what, written down in a form a system can execute. Teams that make those decisions get compounding value out of every platform and every specialist they have already paid for, which is a rare position to be in at this stage of a technology cycle.
What is lead routing, and why does it set the ceiling on everything upstream?
Lead routing is the set of rules that decides which person or team owns an inbound record and how quickly it reaches them. It covers matching a lead to an existing account, applying territory, segment and language rules, checking whether an open opportunity or an established owner already exists, and assigning the record with a service level attached. In practice it runs as assignment rules, queues and round-robin logic in Salesforce, as workflow-based rotation in HubSpot, and as assignment rules and sequences in Dynamics 365. Routing sits between the signal and the conversation, which is why its quality sets the ceiling on the return from scoring models, intent data and AI agents alike.
That ceiling is the useful thing to understand, because it works in the buyer’s favour. A scoring model that improves by ten points delivers nothing extra if the record it flags waits two days for an owner. Reverse it and the arithmetic is just as strong: sharpen the handoff and every signal already flowing through the system gets better at once. The intent data bought last year, the enrichment running today, the agent piloted next month all convert at a higher rate without a single change to any of them. Few investments in a revenue stack have that property.
What is speed to lead, and what is it worth in 2026?
Speed to lead is the elapsed time between an inbound signal, such as a demo request or a pricing enquiry, and the first meaningful response. It is best measured in two parts. Processing time covers everything that happens before a human sees the record: matching, enrichment, routing and assignment. Response time is how quickly the owner acts once it lands. 2026 benchmarks put the median B2B response at roughly 42 hours, with only around a quarter of companies replying inside five minutes. Teams that do reply within five minutes convert at approximately 21%, against about 2.3% for those who wait a day or longer. The difference is close to an order of magnitude.
Splitting the metric in two is what turns it into a project rather than a complaint. Response time depends on rep behaviour, capacity and coaching, and it improves slowly. Processing time is pure systems work, it is fully automatable, and on most inbound flows it is the larger half of the total. Organisations that instrument the two separately usually find several hours sitting in matching and assignment that nobody had ever measured, and that time can be removed in weeks. Best-in-class teams are now targeting sub-60-second processing on high intent signals such as demo and pricing requests, and they are hitting it with rules rather than heroics.
Why is CRM data quality important once agents can act?
CRM data quality is important because every downstream system inherits it. Scoring models, routing rules, forecasts and AI agents all read the same records, so a duplicate account or a stale owner field does not stay a small problem. It becomes a wrong assignment, a missed follow-up, or an agent acting outside the remit its designers intended. In LeanData’s August 2026 customer research, 70% of revenue professionals said data quality was undermining their AI go-to-market results, and 60% were concerned that agents would act on inaccurate data or breach ownership rules. The named root causes were inconsistent or incomplete CRM data at 45%, undocumented business processes at 37%, and siloed teams at 32%.
All three of those are addressable without buying anything, which is the reason to be optimistic about them. More usefully, the scope is finite. Routing does not read every field on every object. It reads account matching keys, owner and territory fields, country and language, lifecycle stage and a handful of firmographics. Cleaning that specific set is a bounded piece of work measured in weeks, not a data programme measured in years. It also pays twice, because the same fields drive territory planning, pipeline reporting and forecast accuracy whether or not a single agent ever goes live. That is the kind of effort worth doing on its own merits, with the AI readiness arriving as a bonus.
What an AI-ready execution layer actually contains
Three things, and none of them is exotic. The first is a written definition of the rules: what qualifies, who owns it, what the fallback is when the primary owner is unavailable, and how long each step is allowed to take. This is also where sales and marketing alignment stops being a slogan and becomes an artefact. The most reliable way to align the two functions is to agree a single definition of a qualified record, attach an owner and a response time to it, and report on the pipeline that definition produces rather than on the volume each side hands over. Once the rule is written, the disagreement becomes a specific, solvable question about a threshold.
The second is guardrails. Fewer than one in three organisations in the benchmark have enforcement mechanisms behind their AI policies, and the ones that do are getting more from every agent they add, because an agent operating inside clear ownership boundaries can be given more latitude rather than less. Agentic AI in a CRM means software that takes actions against records rather than only summarising them, so the boundaries are the feature that makes autonomy safe to grant.
The third is measurement, and it doubles as the adoption strategy. The most dependable way to improve CRM adoption rates is to make the system visibly useful to the person being asked to use it, and fast, accurate assignment does exactly that. A seller who consistently receives the right account, already matched, already enriched, within a minute of the buyer raising a hand, starts treating the CRM as an advantage rather than an obligation. Adoption follows utility, and the handoff is the most visible utility a revenue system has.
How to sequence this over a single quarter
Start by measuring, because the first two weeks usually produce the business case on their own. Instrument processing time and response time as separate numbers on your three highest-value inbound flows. Most teams discover that the half they assumed was a people problem is largely a systems problem, and that reframing is worth the fortnight by itself.
Spend the next fortnight writing the current rules down, including the exceptions that live in people’s heads. Undocumented process is the second most cited root cause in the research, and the documentation is a deliverable in its own right: it is the input to any future migration, any onboarding pack and any audit conversation. Weeks five to eight go on the bounded field clean-up described above, prioritised by what the rules actually read. Weeks nine to twelve are for automation and instrumentation, moving the documented rules into the platform and putting a dashboard on processing time so the gain stays visible.
The sequence is deliberately arranged so that each stage stands up alone. If the quarter is interrupted after week four you still have measurement and documentation, both of which have independent value. That is a comfortable risk profile for a project with this much upside, and it is a large part of why the handoff is the easiest RevOps initiative to get funded right now.
The Sirocco perspective
We work across Salesforce, HubSpot and Microsoft Dynamics 365, and the handoff is one of the few areas where our advice barely changes between them. Routing logic, ownership definitions and service levels are written in business language, not vendor syntax. A rule that says an inbound enquiry from a manufacturing account above a revenue threshold in the Nordics goes to the named account owner within sixty seconds, with a regional queue as fallback, is expressed the same way in every platform we implement. Only the configuration underneath differs.
That portability is the practical case for an independent CRM partner. Work done on the execution layer belongs to the client rather than to a licence. It survives a platform change, it shortens any future migration because the rules are already documented, and it improves reporting and forecasting from the day it lands. When we are asked where a revenue team should spend its next quarter of operations capacity, this is currently the easiest recommendation to make, because the return does not depend on any prediction about which agent or which vendor wins.
The signal layer is built and the platforms are in place. The remaining move is short, well-understood, and almost entirely uncontested, which is a good position for any revenue team to be starting from. If you would like to look at where your own processing time is going and what the handoff could be worth, we are glad to walk through it with you. Schedule a consultation and we will take a look together.
Get in Touch
If you are weighing up where routing, assignment and speed to lead sit on your roadmap for Salesforce, HubSpot or Dynamics 365, tell us a little about your current setup and we will come back with a view.
