Physical Address
Crimson Lynx Media Ltd
Scottish Provident House
76-80 College Road
London
HA1 1BQ
Physical Address
Crimson Lynx Media Ltd
Scottish Provident House
76-80 College Road
London
HA1 1BQ

AI is transforming secondary-market forecasting, but inconsistent device grading and unpredictable consumer behaviour continue to threaten margins and customer confidence.
So said executives from Apkudo, Phonecheck and Dipli at Crccular Summit in London this week. The discussion featured Josh Matthews, Co-Founder and CEO of Apkudo, Harby Garchay, President MEA & APAC at Phonecheck, and Simon Deakin, UK Country Manager at Dipli.
Artificial intelligence is becoming increasingly effective at predicting refurbished smartphone prices and supply volumes, but the industry still faces major challenges in accurately assessing the condition of devices entering the secondary market.
That was the message from a Circular Summit 2026 panel examining how AI is changing trade-in forecasting, pricing, demand planning and inventory management.

The panellists agreed that AI could significantly improve commercial decision-making across the secondary market, but warned that unreliable grading, poor-quality data and unexpected changes in consumer behaviour could undermine its effectiveness.
Deakin rated the effectiveness of AI in predicting used-device prices and volumes at eight or nine out of ten, citing the substantial amount of historical market data available.
However, he gave AI-driven device grading a score of just three or four, reflecting continuing inconsistencies between assessments made by consumers, retail staff and refurbishment operations.

He said smartphone launches and their impact on used-device prices had become increasingly predictable, allowing businesses to forecast supply and demand with greater confidence.
Deakin cited an example in which a predicted iPhone price reduction was within 98 per cent of the pattern observed during comparable previous launches.
He also reported forecasting accuracy of between 90 and 100 per cent when comparing historical patterns with actual outcomes.
However, he acknowledged that unexpected consumer decisions, including switching brands or delaying purchases because of price, could disrupt otherwise reliable predictions.
Matthews was similarly optimistic about AI’s potential but cautioned against assuming that all forecasting systems delivered comparable results.
He rated the technology’s overall potential at ten out of ten and its current performance against capable human forecasters at around eight.
He said sophisticated models using extensive historical data, current market conditions and risk analysis could deliver strong results, while less carefully developed applications of general-purpose AI tools could produce unreliable forecasts.
Matthews argued that businesses needed to understand the limitations of their forecasting systems and establish safeguards against incorrect predictions.
Garchay said AI was already capable of delivering highly accurate pricing information, but assessing the physical condition of a smartphone before it was received remained considerably more difficult.
Unlike price forecasting, which can draw on large volumes of historical transaction data, grading must account for variations in wear, damage and device condition.
He said the industry needed better photographic evidence, diagnostic information and supporting documentation to make automated grading more reliable.
The objective should be to provide a defensible assessment of a device’s condition rather than simply assigning it a grade.
Deakin agreed that discrepancies between initial trade-in assessments and subsequent warehouse inspections remained a significant source of commercial friction.
A consumer might receive one valuation when submitting a device, only for that valuation to change after a more detailed inspection.
Such inconsistencies can reduce margins, increase processing costs and damage confidence in trade-in programmes.
Garchay argued that technology providers needed to focus on practical tools that helped customers make more accurate assessments under real-world conditions.
The panel also examined how the growing tendency for consumers to retain smartphones for longer could affect the availability and quality of refurbished inventory.
Garchay warned that devices entering the secondary market after extended periods of use could require more extensive refurbishment to meet customer expectations.
Replacing batteries with original manufacturer components and carrying out thorough refurbishment would become increasingly important in maintaining quality.
Deakin said the industry needed to encourage consumers to return devices earlier and make trade-in a routine part of the smartphone ownership cycle.
He suggested that guaranteed future trade-in values could help achieve this by giving consumers greater certainty about the residual value of their phones.
Such arrangements could also encourage customers to take better care of devices, improving their eventual resale condition.
Matthews said financing and leasing programmes could provide a more predictable flow of devices into the secondary market.
He expected manufacturers to expand direct leasing arrangements, creating opportunities to establish relationships with customers and secure devices when contracts ended.
The panel also considered how higher refurbished-device prices might affect consumer behaviour.
Rather than automatically switching to new handsets, some consumers could choose to retain their existing phones or wait for prices and availability to improve.
That uncertainty creates an additional challenge for forecasting models built around historical replacement patterns.
Deakin argued that improving trade-in participation was one of the most effective ways of addressing future supply shortages.
He called for greater incentives for retail staff to promote trade-in services and for simpler processes across online and physical retail channels.
Reducing friction in the customer journey would help increase the number of devices returned and provide refurbishers with a more consistent supply of stock.
The panel also highlighted the commercial opportunities created by expanding refurbishment beyond smartphones.
Laptops, tablets and other consumer electronics, including household appliances, were identified as categories where established mobile refurbishment and resale practices could be applied.
Matthews warned that rapid technological development without consistent industry standards could damage confidence in the refurbished market.
He said consumer trust needed to remain the industry’s guiding priority, supported by accurate battery-health information, reliable grading and transparent device histories.
The panellists also discussed the potential benefits of sharing market intelligence.
Deakin said Dipli already provided partners with data and dashboards to support commercial decisions.
Matthews argued that industry collaboration could improve forecasting and market understanding, provided commercially sensitive information was protected.
He pointed to data-sharing arrangements that allow businesses to benefit from aggregated industry intelligence without disclosing confidential trading information.
The panel also considered the impact of regulatory pressure on international device flows.
Matthews said global trading had historically helped match surplus devices with markets where demand was strongest.
However, increasing requirements for more localised processing could make regional refurbishment capacity more important.
Asked to identify the single most important improvement for the industry, Garchay highlighted technology that could reduce grading discrepancies and provide more reliable assessments.
Deakin called for greater trade-in participation, supported by better retail incentives and simpler customer journeys.
Matthews urged the industry to establish stronger common standards before the widespread adoption of advanced technology created further inconsistencies.
The discussion highlighted a growing distinction between two areas of AI development in the secondary market.
While forecasting the future price and availability of used smartphones is becoming increasingly sophisticated, accurately determining the condition of individual devices remains a much less predictable process.
For networks, retailers, distributors and refurbishers, closing that gap could prove just as important as improving the accuracy of supply forecasts.