Attribution engineering
A protected deal-level source property populated by your own logic, so reporting can answer why rather than only what.
AI for Marketing
We build the machine. Your team still does the marketing.
We are engineers, not an agency. We build the attribution, the grounding layer and the approval workflow your marketing team uses. What they say with it is theirs.
.png?height=28&name=Gunnebo%20Logo%20(1).png){kind=small}
.png?height=28&name=Closer%20Still%20Media%20(1).png){kind=small}
.png?height=28&name=Ecommpay%20Logo%20(1).png){kind=small}
.webp?height=28&name=Tussell%20Logo%20(1).webp){kind=small}
-4npmo6jxod.png?height=28&name=iconsolutionsrgb%20(1)-4npmo6jxod.png){kind=small}
What we usually find, and what we build instead
Four things we find in almost every marketing team whose reporting cannot answer why.
Without SpotDev
What we usually find
With SpotDev
What we build instead
Without SpotDev
Reporting describes, it does not explain
You can see that pipeline fell. Nothing in the dashboard tells you which decision caused it.
With SpotDev
Reporting built on the deal record
Source held at deal level and protected, so the question survives contact with a buying committee.
Without SpotDev
Four systems, four definitions
Marketing counts leads, sales counts opportunities, finance counts invoices, and none of the three reconcile.
With SpotDev
One pipeline, rules written down
Scheduled, matched on agreed keys, reconciled to invoiced revenue rather than closed-won optimism.
Without SpotDev
Drafting starts from a blank page
And a search of the intranet for whatever the last person wrote about this product.
With SpotDev
A grounding layer, not a blank page
Approved evidence, claims and product facts assembled so your team drafts from something true.
Without SpotDev
Nothing AI produced has a paper trail
You cannot tell what was machine-drafted, who approved it, or on what basis.
With SpotDev
An approval gate you can query
Every AI-assisted asset carries who reviewed it, when, and against which evidence.
What a marketing wave activates
A wave activates two of the five build capabilities. Enablement runs alongside, and approval and audit are engineered into all of them rather than chosen instead of one.
Content generation enablement
The grounding layer, claim controls and review queue your team drafts with. We build the tooling, they write the words.
Decision-brief assembly
The evidence for a marketing decision gathered by a system, so the meeting argues about the decision rather than the numbers.
Campaign operations
Setup, quality checks and dependency tracking, with conflicting or incomplete data caught before a campaign goes out rather than after.
Segmentation integrity
Detection of incomplete, duplicated and contradictory records, and controlled corrections rather than silent overwrites.
Enablement for the marketing team
Your team trained on the tooling we build, so the grounding layer and the approval queue get used rather than worked around.
Why your reporting describes instead of explains
- 01
The native deal source is inherited
HubSpot documents a deal's original traffic source as coming from the associated contact with the earliest activity for that deal.
- 02
So committee dilution follows you
Moving from contact reports to deal reports does not fix it. Whoever touched first wins the credit, which on a committee sale is often nobody who mattered.
- 03
And native attribution is tiered
HubSpot's knowledge base states deal and revenue attribution are Marketing Hub Enterprise only. On Professional it stops at contacts. Check what your subscription covers.
- 04
The fix is a protected property
A separate deal-level source, populated by rules you control at creation, which nothing downstream overwrites.
- 05
Reconciled to money
Tied to invoiced revenue from your finance system, so marketing reporting and the management accounts agree.
- 06
Then the questions change
You stop asking which channel got the most leads and start asking which one produced revenue you actually collected.
What is engineered, and what stays with people
This is where an engineering firm differs from an agency. We build the controls. Your team makes the decisions inside them.
Customer-facing stays gated
Anything a customer will see sits at draft or act-with-approval. It does not publish itself.
Claims gated separately
A factual claim is reviewed apart from the copy around it, because that is where the risk is.
Lists are restricted
No autonomous change to lists, suppressions or consent state. Ever.
Field-level CRM authority
Writes are limited to named properties, not general access because one action needed it.
Definitions are locked
Changing a reporting definition is a restricted action, or your trend data quietly becomes fiction.
Everything is logged
Who reviewed an asset, when, and against which evidence. Queryable, not a report you run later.
What AI for marketing costs
Three ways in, depending on how far you want to go. Prices exclude VAT, and AI subscriptions and usage are billed to you by the vendor. For a whole department of up to 100 people, AI Rollout starts at £18,000.
AI Starter
£3,500
Your marketing team using AI on their own briefs, reporting and campaign work, set up and trained online over four weeks.
- Up to 15 people, fully remote
- Workspace and usage rules set up for you
- Three live online sessions on real marketing work
- Shared starter tools for briefs, reporting summaries and campaign checks
- A review at day 30
Custom Agents
£20,000
One marketing operations workflow built into your CRM and put into controlled use.
- One workflow, such as decision-brief assembly, or segmentation checks
- Integration with your CRM and data sources
- Approval rules written down per action
- Training for the people who use it
- Measured against a baseline
- 60-day post-launch support
AI Accelerator
From £10,000 per month
Marketing as one wave of a twelve-month programme that moves a department into controlled production each quarter.
- Committed a quarter at a time
- Two production agents per wave
- Baseline, measurement and a monthly review
- Model and platform usage billed separately
Implementation expertise that drives adoption
An in-house engineering team that has delivered more than 300 technology projects. Work with engineers who have built this before.
300+HubSpot implementations, firm-wide
In-house engineering team.
Agents and integrations designed, tested and shipped by our own engineers. Nothing is subcontracted.
OpenAI Select Partner, Certified in the Claude Partner Network
SpotDev is an OpenAI Select Partner and is Certified in the Claude Partner Network, alongside HubSpot Diamond status and Custom Integration and Onboarding accreditations.
Model chosen for the job.
Honest evaluation of the right model for the work. We will say when the answer is the AI you already pay for.
AI with security in mind
- Not trained on by default
- Cyber Essentials Plus
- Built on your infrastructure
- In-house engineers
Common questions
Are you going to write our content?
No. We build the system your team writes with: the grounding layer, the claim controls, the review queue and the audit trail. The voice stays yours. If you want an agency, we will say so on the first call.
Do we need Marketing Hub Enterprise?
Not for what we build, and that is rather the point. HubSpot's knowledge base puts native deal and revenue attribution at Enterprise only. A protected deal property is an engineering answer to a licensing problem.
Our CRM data is a mess. Is it too early?
It is exactly the right time, because attribution engineering is largely data engineering. We assess completeness and matching before building anything, and tell you what has to be fixed first.
Can the AI publish or change records on its own?
No. Anything customer-facing is drafted for approval, list and consent changes are restricted outright, and CRM writes are limited to named fields.
What does HubSpot already cover?
More than people assume. Brand voice and its own agents exist and may be enough. Where they are, use them. We build what is missing around them, which is usually the data layer and the audit trail.
What does it cost?
AI Starter gets a team of up to 15 using AI on their own marketing work for £3,500, set up and trained remotely over four weeks. One marketing workflow built into your systems is a Custom Agents project at £20,000. Running marketing as one wave of the AI Accelerator starts at £10,000 per month, committed a quarter at a time. Model and platform usage is billed separately, and enterprise programmes are priced on complexity.
What if a wave does not justify the next one?
Then you stop. The baseline is captured before we build, so the quarterly decision rests on a comparison rather than a mood.
Start with a diagnostic.
A short, fixed-scope assessment of what your data can currently support, and where AI would genuinely change marketing performance rather than increase output. No obligation.
SpotDev is an OpenAI Select Partner and a HubSpot Diamond Solutions Partner, and is Certified in the Claude Partner Network. SpotDev is a separate company from OpenAI and from Anthropic, and partner status does not constitute endorsement of SpotDev by OpenAI or by Anthropic. OpenAI and ChatGPT are trademarks of OpenAI. Claude is a trademark of Anthropic, PBC. HubSpot is a trademark of HubSpot, Inc.