Trust · Disclosure
How we use AI
DigeeSell uses AI in its own content and in client work. This page says where, who stays accountable for the output, and which parts of this disclosure are still incomplete.
Short answer
Does DigeeSell use AI in its work?
Yes. AI is used to draft content, to assist research and analysis, and inside the automation workflows we build — lead routing, CRM hygiene, reporting pipelines and campaign alerting. It is not used to decide anything that carries consequences. Every published piece is edited by a person and finally reviewed by Nivesh Babbar, every figure we publish carries a measurement note, and regulated material is approved by the client's compliance function before it goes live.
Updated
How AI is used in the content on this site
Drafting is automated. A content writer then reviews and edits the draft, and Nivesh Babbar carries out the final review before anything is published. That is the full process, and it is set out on the editorial policy page as well as here.
We state this rather than implying the alternative. A great deal of published marketing content is drafted the same way and presented as though it were not, and the claim rarely survives scrutiny. What matters for a reader is not whether a machine produced the first draft but whether a person is accountable for what was published.
How AI is used in client work
- Drafting — first drafts of pages, articles, ad copy and meta content, which a person then edits against the brief and the client's constraints.
- Research and analysis — summarising competitive and keyword data, grouping queries by intent, and surfacing patterns in large result sets that would be slow to read by hand.
- Automation workflows — the automation service we build for clients uses AI inside lead routing, CRM hygiene, reporting pipelines, content workflows and campaign alerting. This is a service we sell, so it is disclosed here rather than left implicit.
In each case the output is an input to someone's judgement, not a decision. A draft is edited, an analysis is checked against the source data, and an automation routes work to a person rather than replacing them.
What people decide
- Whether a claim is true, and whether we can evidence it. Statements on this site are labelled as documented, reported or our own observation, and that labelling is a judgement rather than a generated one.
- Whether a lead counts. A qualified lead requires a phone conversation confirming interest in the specific offer advertised, which is a human conversation by definition. The standard is on how we measure.
- Anything regulated. Healthcare, property, legal and financial material goes to the client's compliance function before publication, and no automated step substitutes for that approval.
- What gets published at all. Final editorial review sits with a named person, not a process.
What we do not do with it
We do not publish people who do not exist. Where this site needs a team profile it has not been given, it shows an explicitly labelled placeholder rather than a generated name, photograph or biography — which is visible on the team page today.
We do not publish figures we cannot attribute. Every performance number on this site carries a note saying how it was measured, and where a claim comes from our own experience rather than a published source, it says so in those words.
We do not publish generated regulatory detail as fact. Regulatory summaries carry a review date and send the reader to the regulator, because requirements change between reviews and a confident-sounding summary is the easiest thing in the world to generate and the hardest for a reader to check.
We do not generate review or rating markup for ourselves. The site strips Review and AggregateRating nodes automatically at every level of its structured data, so self-serving review markup cannot be introduced by accident.
What happens to campaign and client data
Client data is governed by the engagement agreement, and nothing on this page overrides what has been agreed with an individual client. The practices below are what we apply by default.
- Campaign and conversion data is fed back to the advertising platforms. This is how Google and Meta optimisation works: the platforms use conversion signals to decide who sees an ad and at what bid. A campaign that withholds those signals performs worse, so this is part of running the account rather than an additional use of the data.
- Performance patterns, aggregated and stripped of client and personal identifiers, inform how we work across accounts — which is what the published benchmarks are built from. Aggregate means aggregate: no single client's figures are identifiable in them.
- AI tools are used in delivery — drafting, research and the automation workflows described above — and the output is reviewed by a person before it reaches a client or is published. Nothing generated is delivered unread.
- Client work is not published without permission. Case studies here are anonymised by sector and market, every figure carries a measurement note, and no client is named without agreeing to it first.
- Account access — ad accounts, analytics, Search Console, CRM — is used for the agreed work and nothing else, and is handed back or revoked when an engagement ends.
Personal data in leads
Lead data is different from campaign data and is treated differently. Names, phone numbers and email addresses submitted by a client's prospective customers are personal data under the UAE's Personal Data Protection Law and, for Indian clients, under the Digital Personal Data Protection Act. It was collected for one purpose — so the client could respond to an enquiry — and we process it for that purpose and the measurement that goes with it.
The client decides why that data is collected and what happens to it. We act on the client's instructions, which means we do not acquire our own right to reuse it by virtue of holding it. Repurposing it for an aim of our own would make us responsible for that aim in our own right and would require its own lawful basis and its own notice to the people concerned — so we do not do it.
- Identifiable lead data is not used to train our own models, and not used for any purpose outside the engagement it came from.
- Where AI tools are used on client material, they are configured so the vendor does not retain the content or use it to improve their models. Business and enterprise tiers of the major providers offer this; consumer tiers generally do not, which is why the tier matters more than the tool.
- What does inform our wider practice is aggregate: performance patterns with client and personal identifiers removed, at a level where no individual and no single client is identifiable. Aggregation is a judgement about re-identification risk rather than a label, so the test applied is whether a figure could be traced back, not whether a name was deleted from it.
- Special categories get stricter handling. Health-adjacent data for clinic clients and anything touching financial-services clients is kept inside the engagement regardless of what any general policy here says.
Where a client's own agreement with us says something narrower than this page, the agreement governs.
What do buyers ask before starting?
Is the content on this site written by AI?
Drafted with AI, then edited by a content writer and finally reviewed by Nivesh Babbar before publication. We say so rather than implying otherwise, because the claim that no AI was involved is one most agencies could not defend.
Does AI decide anything about my campaign?
No. AI drafts, summarises and routes; people decide. Budget changes, creative approval, what counts as a qualified lead and anything regulated are human decisions with a named person accountable for them.
Will AI-drafted content hurt our rankings?
Not inherently. Google's published position is that it rewards helpful, reliable content made for people whatever the production method, and treats automation used primarily to manipulate rankings as spam. The failure mode is content that restates what already ranks and carries no first-hand detail — which is a quality problem, not a tooling one.
Do you tell clients when AI is used?
This page is that disclosure, and the editorial policy states the drafting process. For questions specific to your account and your data, ask us directly and we will answer for your engagement.
Is our data used to train your AI models?
No. Identifiable client or lead data is not used to train our models and is not used outside the engagement it came from. Campaign and conversion data does go back to the advertising platforms, because that is how their optimisation works and withholding it makes the campaign perform worse. Aggregate performance patterns with all identifiers removed inform our published benchmarks and how we work generally.
Do the AI tools you use train on our content?
They are configured not to. Business and enterprise tiers of the major providers let you turn off retention and model training on submitted content; consumer tiers generally do not, which is why the tier matters more than which tool it is. If this is material to your procurement, ask us which tools apply to your account and we will tell you.
What happens to the personal data of our leads?
It is processed for the purpose it was collected for — so you can respond to the enquiry — and the measurement that goes with it. It is personal data under the UAE's data protection law and India's DPDP Act, and we do not repurpose identifiable lead data outside the engagement it came from. Anything used more widely is aggregated and de-identified first.
Do you use AI to generate reviews, testimonials or case study figures?
No. Case studies are anonymised real client work and every figure carries a measurement note. The site also strips Review and AggregateRating structured data automatically, so self-serving review markup cannot be introduced accidentally.
Free audit
Judgement is the part you are paying for.
The audit establishes where search visibility and site performance are failing, and a person decides what is worth fixing first.
Request the free audit