SMB AI Adoption Is Stuck at 9%. Tech Data’s Answer Is Free Skills.
The distributor is giving away prebuilt AI skills to sell the licences they consume. Its own commissioned research says the barrier was never skills.
Tech Data, the distribution arm of TD SYNNEX, added four capabilities to its Destination AI framework this week, aimed at partners selling to small and medium businesses across Asia Pacific and Japan. The most consequential is an AI Skills Marketplace inside StreamOne that offers prebuilt, reusable skills for sales, marketing, customer support, operations and knowledge management at no cost. The skills are free. What they consume is not. Tech Data states plainly that the free skills create revenue by driving licence adoption, increasing consumption and expanding demand for AI services.
That is a distributor conceding the application layer to win the consumption layer, and it is the most interesting thing in the announcement. It is also the part no trade outlet covered. Every Indian outlet ran the release verbatim.
What was actually announced
The four additions are SMB-specific experiences in AI Game Plan, the workshop format Tech Data uses to identify and prioritise use cases; SMB-focused Cloud Labs, where partners and customers validate a use case hands-on before deployment; the AI Skills Marketplace in StreamOne; and an AI Skills Foundations training course. They sit inside a five-pillar structure, enablement, discovery, proof of value, delivery and adoption, that Tech Data has been building since it launched Destination AI in APJ in September 2024.
The framework is not new, and the SMB tier is an extension rather than a launch. What is new is the marketplace, because a marketplace changes the unit of sale. A partner no longer has to build an AI capability from scratch for each customer; it can start from a prebuilt skill and adapt. That is the difference between selling a project and selling a repeatable practice.
A note on what a skill is, because the term carries weight here. In the agentic vocabulary Tech Data is using, a skill is a packaged instruction set: a defined task, the tools it may call, the data it may read and the output format it produces. A sales skill might draft a proposal from a CRM record; a support skill might classify and route a ticket. Packaging these is useful because the same task recurs across customers with small variations, and rebuilding it each time is where partner hours disappear. The open question is what happens when a skill meets a customer’s actual systems, which is the part no marketplace can standardise.
The economics of giving it away
The pricing logic deserves to be stated as the strategy it is. A distributor’s margin comes from moving volume, and AI has a volume problem in the SMB segment: the skills to implement it are scarce, the use cases are unclear and the willingness to pay for an experiment is low. Tech Data’s answer is to remove the price of the first step. The skill costs nothing, so the partner’s cost of trying is time rather than money. If the skill works, the customer buys the licence it runs on, consumes more of it, and eventually needs services around it.
That is a razor-and-blades structure, and it is reasonable for a distributor, because Tech Data does not need to monetise the skill itself. It needs the skill to exist so that the licence behind it sells. The risk is the mirror image: if the skills are good enough to be reusable, they commoditise the implementation work partners currently charge for. Tech Data is trading partner services margin for licence volume, which works if the volume arrives and fails if partners conclude the marketplace competes with them.
What TD SYNNEX’s own research says
The company has better evidence about this market than most vendors, because it commissioned some of it. In June 2026 it published, with SAS and IDC, a study of 1,600 SMB leaders across 28 countries called AI for SMBs: Closing the Readiness-Reality Gap. The findings are blunt. Nearly 70 per cent of SMBs remain in the early stages of AI maturity. 37 per cent are running isolated, disconnected experiments. 33 per cent have put some structure around their efforts. Just 9 per cent have fully embedded AI into how they make decisions and run operations, and only 22 per cent are aligning people, priorities and execution.
A TD SYNNEX executive quoted in that study put the conclusion more directly than the release does: the gap between aspiration and execution stands out, and what partners and customers lack is not interest but a repeatable path. Nine per cent is the number that justifies the marketplace, and it is the central fact of SMB AI adoption today: a market where nine in a hundred customers have operationalised AI is a market where the constraint is repeatability, and a catalogue of prebuilt skills is a plausible answer to repeatability.
The barrier the marketplace does not address
IDC’s own SMB research, published the same month across 2,700 IT decision-makers in 23 countries, identifies a different constraint. Asked in 2026 what stands between them and their business priorities, SMBs named implementing new technology securely as the number one challenge, ahead of lack of budget, ahead of user adoption and ahead of lack of IT staff. The same research names unpredictable pricing as the first barrier to AI specifically: an SMB implements a capability, sees it work, and then watches the bill triple because it crossed an invisible usage tier.
Neither problem is solved by a free skill. A prebuilt skill that drives consumption is, from the customer’s side, a mechanism for exactly the pricing unpredictability IDC warns about. Tech Data’s model depends on consumption rising; the customer’s interest lies in it being legible and capped. Those are not the same objective, and a marketplace that accelerates deployment without addressing the meter does not remove the objection, it postpones it.
The 40 per cent with no IT staff
The structural fact underneath SMB AI adoption is that 40 per cent of SMBs in IDC’s sample have no full-time IT employee in the building, a figure that has held steady for years even as technology budgets grew. That is why embedded AI has been winning against standalone AI: an SMB with no IT staff does not want another console to administer, it wants a capability switched on inside the CRM or the accounting system it already runs.
A skills marketplace is, in that light, an awkward fit. It is a new surface, even a well-stocked one, and it assumes a partner who can evaluate, adapt and support what is inside it. For the 60 per cent with some IT capacity that is fine. For the 40 per cent without, the marketplace helps only if the partner does the work, which returns the story to the partner’s own capacity.
Why India looks different
The Indian market adds a wrinkle worth noting for SMB AI adoption, because the release was made in India and the local evidence is unusual. In the CPA Australia Asia-Pacific Small Business Survey for 2025-26, covering 4,166 small businesses across 11 markets, India reported AI as its most invested-in technology at 35.6 per cent, ahead of cloud computing at 34.9. Across the region the share of small businesses naming AI as their most significant technology investment reached 32 per cent in 2025, up from 9 per cent in 2022.
More striking is where Indian small businesses go for advice. AI tools were the most popular source at 40.5 per cent, ahead of IT consultants at 39.2 per cent, and only 5.8 per cent sought no external advice at all. Indian SMBs are consulting the model directly rather than the channel. That is a market adopting AI fast and bypassing the intermediary, which makes the partner-led, marketplace-mediated model a harder sell here than the investment figures suggest.
What is unproven
The release supplies none of the numbers that would test any of this. There is no count of partners active on Destination AI, no number of skills in the marketplace, no pricing for what sits behind them, no adoption or consumption data, and no named customer. It cites IDC for the proposition that SMBs favour easy-to-implement solutions without naming a report, a date or a figure, which is a weaker citation than the research the company itself published four months earlier. And the central quote about moving beyond productivity use cases appears in the India release under Anand Chakravarthy’s name and in the global release the same day under Sergio Farache’s.
What to watch
Three markers will show whether the marketplace changes anything. The first is skill count and reuse: a catalogue of twenty skills used once each is a demo, and a catalogue of two hundred used repeatedly is a practice. The second is whether consumption rises without pricing complaints, since the razor model fails at the meter rather than at the blade. The third is partner economics, because if partners conclude the marketplace competes with their services revenue they will route around it.
What it signals
The announcement marks a stage in SMB AI adoption: not through a new product but through the distribution layer deciding which layer it will own. Tech Data is betting that the skills will commoditise anyway and that the durable position is the licence and the consumption behind it. That is probably right about the skills. Whether it is right about the customer is a separate question, because the SMB that cannot predict its AI bill has not been helped by getting the skill for free.