Dayjob’s AI Agent Delivers £800K Revenue Boost in First Year

Dayjob's Agent Cuts Morning Planning to 60 Seconds

Dayjob, a London startup from Y Combinator's Spring 2026 batch, launched its autonomous scheduling agent in late April 2026 with a proposition that upends the morning ritual at every UK waste firm: plug into the ERP the hauler already runs, and the agent returns a full fleet schedule in about 60 seconds. It then keeps re‑optimising as the day unfolds, handling new jobs, driver changes, and exceptions, without a planner touching the board.

The company was founded by George Postlethwaite and Fred Fooks, who met at Oxford 13 years ago. Postlethwaite was an early employee at Deliveroo (acquired by DoorDash in 2025) and launched their UK grocery business; he previously led sales at Otta, the LocalGlobe‑backed job‑search platform. Fooks built Gaea, an ERP for e‑waste recyclers, from scratch; he holds an Engineering Science degree from Oxford and worked in AI and data science at Deloitte and Capgemini. They started in October 2022 selling ERP software to waste operators, reaching $150K in revenue before two hard lessons forced a pivot in January 2025: large operators would not migrate ERPs no matter how good the product, and better UI did not solve the core problem — manual decision‑making at scale. Dayjob was built to sit on top of existing systems instead of replacing them.

The agent went live with UK fleets in August 2025. Seven months later, annual recurring revenue reached $653K, Y Combinator's data shows. Early customers include some of the largest waste and recycling operators in the UK and US, though Dayjob does not name them publicly. VWS Software Solutions announced a partnership in January 2026; Coastal Recycling published a case study the same month. Backing comes from Paul Graham, YC's founder, and Harry Stebbings of 20VC.

Early UK customers report 5 to 10 percent fleet‑efficiency gains from day one, Founderland's data shows, with top accounts hitting 8 percent or better, Y Combinator's figures put. One customer recorded an £800K revenue uplift in the first year, Founderland reported. Time‑window adherence improved 50 percent, Y Combinator found. Admin time dropped 25 percent, roughly 65 planner‑hours saved per week. Dayjob's own calculator shows 100 jobs completed daily translating to £100K additional yearly revenue for a typical fleet; the company argues that adding 10 percent more jobs to a schedule is worth millions annually for larger operators. The agent handles skip sizes, driver schedules, travel times, and priority rules (distance, customer, time‑slot) out of the box. Integration with telematics feeds lets it adapt routes mid‑day when reality intrudes.

Three more agents covering the full operational workflow (pricing, booking, and dispatch) are slated to launch later in 2026. The US rollout is underway now. The question is whether the early UK traction, built on operators who were "true believers" in the founders' words, can survive the fragmented US market where regional haulers dominate and buying cycles move slower.

The Hiring Plan That Followed Proof

The launch forced a hiring plan that had been theoretical until the agent proved it could cut a morning's planning work to sixty seconds. Dayjob completed the Spring 2026 Y Combinator batch with a team of five in London, ~20 percent month‑on‑month growth, and a trajectory toward $1 million ARR. The founders posted two roles: a Founding Account Executive in Moorgate and a Founding Optimisation Engineer, also London‑based, on‑site.

The sales role is not a typical account‑executive slot. The job description makes that explicit: "This isn't a typical Account Executive role." The hire will generate their own pipeline through outbound prospecting, own the full cycle from first outreach to close, and spend time on customer sites — depots, not Zoom — to understand how waste operators actually run their mornings. They'll also feed directly into the sales playbook, qualification process, and messaging, working alongside the founders and the head of sales. Compensation runs ~£50–80k base plus OTE and meaningful equity; four days in the Moorgate office, occasional UK travel, 25 days holiday plus bank holidays and a birthday off. The requirement list is short: B2B software sales experience, comfort building pipeline from scratch, and a bias for action. New grads are eligible.

Postlethwaite, CEO, launched Deliveroo's UK grocery business and led sales at Otta. Fooks, CTO, who earned that Oxford engineering degree, built systems at Deloitte and Capgemini before co‑founding Dayjob. They know the buyer because they've sat in the depot at 7 a.m. watching planners drag jobs onto trucks for 60 to 90 minutes, only to see the plan collapse within the hour. They also know the hiring pain: customers say they literally cannot hire and train planners fast enough — most burn out within 18 months. The agent solves the planning bottleneck; the founding AE solves the distribution bottleneck.

The engineering hire signals where the product roadmap is heading. A Founding Optimisation Engineer, based on‑site in London, will work on the continuous re‑optimisation loop that handles new jobs, driver changes, and exceptions in real time. That loop is the differentiator: the agent doesn't just output a static schedule; it keeps rewriting it as the day unfolds. Scaling that capability across those operators (Dayjob's current customer set) means deeper ERP integrations, tighter data contracts, and lower latency on the optimisation engine. The founding team of five cannot carry that load alone.

Both roles are tied to the US launch already underway. The founding AE walks into a conversation with a $45 billion short‑haul market still run on spreadsheets, phone calls, and 1990s software, armed with a product that plugs into existing ERPs without rip‑and‑replace. The optimisation engineer ensures the agent keeps its sixty‑second promise as the fleet size and exception volume grow.

The hiring pace reflects a deliberate choice: stay lean until the agent proves itself in production, then add the two functions that turn a working prototype into a repeatable revenue engine. Sales and optimisation. Distribution and depth. The next six months will show whether that sequence holds.

Can Incumbents Bolt On AI Fast Enough?

The waste-hauling operators Dayjob is courting have spent decades buying transportation management systems from a short list of incumbents. Those vendors, including Carrier Logistics Inc. (CLI) for less-than-truckload and the telematics giants Samsara, Geotab, and Verizon Connect, are not sitting still. Since Dayjob's April launch, each has accelerated AI dispatch features that look less like roadmap items and more like defensive fortifications.

The clearest signal came April 13, 2026, when STG, a Menlo Park private-equity firm, announced its acquisition of CLI, the premier TMS for LTL carriers. STG's stated plan: use CLI's installed base as a launchpad for an "AI-agentic platform" that does the work the software currently just tracks. The distinction matters. Today's TMS shows a dispatcher a screen of unassigned loads; an agentic layer assigns them, reassigns them when a driver calls out sick, and re-optimizes the whole network before the coffee cools. STG's press release framed the move as maintaining CLI's "market-leading" position: code for keeping startups like Dayjob from wedging in at the scheduling layer.

Geotab, the telematics leader with roughly six million connected vehicles, has taken the most visible swing. In March 2026 it made ACE (its generative AI assistant) generally available to customers after a year in which over 3,000 fleets asked it questions such as "who is driving at this intersection in North Las Vegas between 1 and 2 p.m. on January 15th" and "which cities have the highest vehicle downtime?" Two new modules dropped alongside: "suggest and assist," which recommends the right technician for a job based on certification, proximity, and customer lifetime value, and "capacity planning," which prevents overloads by matching vehicle load to route constraints. A Geotab demo illustrated the logic: a dispatcher sees Tom three miles away but uncertified; the agent picks Sarah twelve miles out because she holds the cert and the account spent $15,000 last year. Wide Open West, a long-standing customer, saw its seat-utilization score climb from 8.2 to 9.5 after adopting Geotab's routing and optimization platform.

The voice agent, now in closed beta with Telus, lets a driver say "re-route me around the accident on I-90" and watch the schedule rewrite in real time. Geotab built it with its own AI division and Google's DeepMind team (the same group behind Gemini) signaling that the telematics stack is becoming an agent runtime, not just a data pipe.

Samsara and Verizon Connect have been quieter publicly, but the pattern is consistent across the category. Dayjob sits in the AI-dispatch platform group; the incumbents are bolting agentic features onto telematics-led products. McKinsey notes that incumbents "can build on their scale, data, and customer relationships while rewiring how they work." BCG adds that companies embracing generative AI report "substantially improved productivity, customer responsiveness, and data-driven decision-making capabilities": metrics waste operators cite when they evaluate Dayjob.

The defensive logic is straightforward. A waste hauler running CLI or Geotab already pays for the TMS and the telematics. If the vendor ships a "good enough" scheduling agent inside that contract, the switching cost for a standalone tool like Dayjob jumps. The incumbents don't need to beat Dayjob on algorithmic purity; they need to clear the "good enough" bar before the procurement cycle renews.

Early evidence suggests they're clearing it for basic dispatch. Geotab's collision-risk users saw a 30 percent cumulative drop in collisions after two quarters; fleets using its safe-driving coaching logged a 40 percent improvement in driving scores by day 60, translating to roughly $200,000 a year in avoided crashes, fuel waste, and driver turnover. Those are operational wins, not scheduling wins — but they deepen the moat. The next frontier, which Geotab explicitly targets, is making ACE "capable of doing any task within my Geotab and more importantly integrating with all the other AIs that exist out there." That is a platform play: own the orchestration layer, and the specialist agents become plugins.

For Dayjob, the implication is clear. The wedge — 5–10 percent efficiency gains from day one — must widen before the incumbents' agentic layers mature from beta to default. The waste operators Dayjob talks to are evaluating both paths in parallel: a standalone agent that plugs into their ERP, and the AI dispatch module their TMS vendor just added to the renewal quote. The next twelve months will decide whether a specialist agent can hold ground against a platform that already owns the data, the contracts, and the fleet.

Fuel, Carbon Rules, and a $45 Billion Wedge

Dayjob's job posting for its founding account executive in London describes short‑haul trucking as a "$45bn market still dependent on those same legacy systems." That figure is a wedge: the company's serviceable slice of a much larger pie.

The spread reflects different boundary choices (trip distance, vehicle class, domestic versus cross‑border) but all three agree the segment is growing faster than long‑haul freight.

Fuel is the single largest variable cost. Mordor Intelligence reports fuel represents up to 30 % of operating costs on urban and regional runs; MarkWideResearch narrows the band to 25‑35 % for Class 8 fleets. In 2024 the national average on‑highway diesel price swung between $3.60 and $4.75 a gallon. That volatility hits smaller fleets hardest: roughly 350,000 owner‑operators lack the scale to hedge or secure bulk discounts. Surcharge mechanisms typically lag price moves by 30‑60 days, compressing margins on contracted lanes when diesel spikes. California's expanding renewable‑diesel mandate adds another compliance cost layer that will ripple nationwide through 2030.

Regulatory pressure is tightening in parallel. The EPA's Phase 3 greenhouse‑gas standards for heavy‑duty vehicles take effect with interim targets in 2027 and run through model year 2032, forcing fleets to accelerate electrification or alternative‑fuel adoption. California's CARB Advanced Clean Trucks rule layers on escalating zero‑emission sales quotas, creating a bifurcated compliance burden for national carriers that run interstate routes. The Inflation Reduction Act's commercial‑vehicle tax credits blunt some capital outlay for depot‑charging operators, but megawatt‑scale charging infrastructure still requires utility coordination, transformer upgrades and land acquisition that stretch payback periods beyond traditional equipment cycles.

The driver shortage compounds the cost squeeze. The American Trucking Associations counted a 78,000‑driver shortfall in 2024 with turnover above 90 % at large truckload carriers. Entry‑level training rules, stricter drug‑testing enforcement and a workforce skewed toward drivers over 48 constrain inflow. In July 2026 the FMCSA finalized new non‑domiciled CDL standards that could remove an estimated 200,000 drivers over a five‑year phase‑out (roughly 40,000 per year through 2031). Simultaneously, ELD certification revocations are grounding carriers whose devices fail new safety‑integrity checks. Insurance carriers are beginning to refuse coverage for fleets employing non‑domicile drivers.

Capacity is already reacting. In June 2026, for the first time since early 2022, the national average van truckload spot rate exceeded the contract rate. Drive‑van rates hit $2.49 a mile in the first week of July (nearly 50 % above year‑ago levels) while line‑haul rates excluding fuel surged 45 %. Tender rejections climbed above 15 %, and shippers are raising contract rates 10‑11 % just to lock in capacity. The Southeast is a particular hotspot: origin volumes there rose more than 57 % year‑over‑year. Flatbed logged a 17‑week rate increase streak, clearing the 2021 weekly record by 24 cents.

E‑commerce continues to rewire freight flows. U.S. online sales reached $1.1 trillion in 2024, 16 % of total retail, driving high‑frequency, low‑volume deliveries that favor short‑haul networks. Verified Market Research identifies this as the single most significant driver reshaping the segment. Manufacturing still holds the largest end‑user share at 36.6 %, but wholesale and retail trade is forecast to grow at a 4.22 % CAGR through 2031. Domestic movements command 92‑95 % of revenue; international short‑haul is growing at 4.11 %.

Against this backdrop, AI routing moves from nice‑to‑have to margin defense. When diesel is 30 % of your cost base and capacity is tightening, a 5‑10 % efficiency gain from tighter routing (the range Dayjob's early UK customers report) translates directly to surviving the next rate spike.

Integration Is Where Pilots Stall

Dayjob's pitch is clean: plug the agent into whatever ERP a hauler already runs, no rip-and-replace, and get a full fleet schedule in sixty seconds. The reality inside the depot is messier. Waste operators don't run one system — they run a stack. Municipal solid waste integration typically touches three core surfaces: a work-order management platform such as Infor EAM or Tyler FleetFocus, a citizen request portal (often a 311 system or CRM), and a billing and revenue engine like Munis or SAP utility billing. Layer on fleet telematics from Samsara, Geotab, or Verizon Connect, routing engines like RouteSmart or TruckLogic, weigh-scale systems, and sometimes SCADA feeds from material recovery facilities, and the "single ERP" becomes a federation of databases that rarely share a common key.

Data quality is the primary constraint. Successful implementations often begin with a data-hygiene project to standardize address, material_code, and service_frequency fields across billing, CRM, and field systems. Without that cleanup, the agent ingests duplicate service points, mismatched container types, and stale frequency codes — garbage in, garbage out. One integration guide notes that the delay in AI-ERP deployments is "not often due to technological challenges but rather data quality and the readiness of the organization to implement the solution." Pilot projects for a single workflow (say, dynamic routing for commercial collections) typically take three to six months; full deployment across the stack runs 12 to 18 months.

Legacy architecture compounds the problem. Many waste ERPs still speak SOAP or proprietary batch interfaces; the agent expects REST, GraphQL, and webhooks. Middleware or an iPaaS layer becomes mandatory: an orchestration tier that authenticates into each system, normalizes payloads, and surfaces a clean event stream the agent can consume. That layer must also handle fallback: if the AI service goes dark, dispatch reverts to standard operating procedures so collection never halts. Role-based access controls from the host ERP must govern who can approve an AI-suggested route change or tonnage forecast, and every agent action needs an immutable audit trail tied to source data and model version.

Change management is the silent killer. Planners who have spent 60 to 90 minutes each morning dragging jobs onto trucks don't hand over the wheel because a dashboard says "8 percent efficiency gain." They need co-designed workflows (dispatchers, drivers, and MRF supervisors shaping the UI prompts and approval thresholds) and wins that matter to them: fewer missed pickups, less overtime, predictable end-of-shift times. Big-bang rollouts fail; the pattern that works starts with a narrow, high-impact pilot (dynamic routing on one commercial route or automated citizen-inquiry triage), measures baseline versus uplift, then expands by adjacency: predictive tonnage forecasting next, then facility maintenance alerts.

Security and compliance cannot be afterthoughts. Agents touch citizen PII, billing records, and hazardous-waste manifests. A security-first architecture means agents run as microservices calling core systems via scoped APIs, never storing sensitive data in third-party models, with OAuth 2.0 least-privilege tokens, rotation, and human-in-the-loop gates for any action that moves money or changes safety-critical routes. Governance boards review model drift against real-world outcomes (predicted versus actual fuel burn, missed-pickup rates) and feed dispatcher overrides back into retraining loops.

Dayjob's "no rip-and-replace" promise holds at the contract level; at the keyboard level, the integration tax is paid in data cleansing, middleware engineering, and months of phased rollout. The operators who clear those hurdles see payback in three to nine months. The ones who don't stall at pilot.

Will Agentic AI Cross the Chasm in Five Years?

Gartner published its 2026 Hype Cycle for Agentic AI on April 2. The verdict: "rapid progress in agentic AI is exceeded by hype and confusion." The firm places the technology squarely in the phase where extreme interest and substantial investment outpace proven enterprise readiness. Gartner describes the current enterprise focus as "incremental automation, not yet transformative change." That assessment lands hard for a sector like short‑haul trucking, where Dayjob's agent promises to replace morning spreadsheet sessions with a 60‑second schedule — but only if the surrounding data, governance, and change management hold up.

Adoption numbers tell a similar story. Only 17 percent of organizations have deployed AI agents to date. Another 42 percent expect to do so within 12 months, and 22 percent within the following year, pushing the two‑year deployment intent above 60 percent. Intent is not installation. The same research flags "agent washing" as the biggest risk: vendors rebranding traditional automation, robotic process automation, and legacy workflow tools as "AI agents" to capture budget. For waste‑haul operators evaluating Dayjob against incumbent TMS add‑ons, that distinction matters. A rebranded rules engine does not re‑optimize routes when a driver calls in sick.

ROI evidence remains thin. BCG surveyed more than 180 logistics providers and found only 13 percent achieved measurable AI returns. Across industries, just 10 percent report measurable financial impact so far. Uncertain ROI and a lack of internal capabilities rank as the top barriers. Gartner adds that fully autonomous agents are not yet ready for most enterprise use cases; human oversight remains essential, and semi‑autonomous deployments will dominate for the foreseeable future. That aligns with what Dayjob's early UK customers experience: planners still in the loop for exceptions.

The readiness gap cuts both ways. "Organizations cannot automate what they do not fully understand themselves," the Gartner analysis notes. Many firms still treat AI as a technology investment rather than a systems transformation. Without AI‑ready data, strong engineering foundations, and scalable governance, they risk accelerating fragility faster than value. Waste fleets running on 1990s ERPs and fragmented data know this problem intimately: it is the integration hurdle detailed in the previous section.

Looking five years out, Gartner's Plateau of Productivity — where agentic AI delivers consistent, measurable value — remains the target. The organizations best positioned to succeed are those investing now in pilot implementations and internal expertise. Meanwhile, the labor impact looms: Gartner projects more than 32 million roles transformed significantly each year starting around 2028‑2029, a "job chaos" scenario rather than an apocalypse. For short‑haul trucking, that timeline coincides with the freight market's projected growth and with rising shipper expectations that logistics providers offer AI‑enabled services.

The hype cycle turns on proof, not promises. Dayjob's 5‑10 percent gains in UK waste fleets are a data point. The next 24 months will show whether agentic scheduling becomes standard infrastructure or another layer of software that planners work around.

Working in frontier tech? Zero G Talent tracks the openings: see every open ASML role, browse frontier tech jobs, openings at Stripe, and the people building the field.